DESIGN SOLUTIONS FOR INTELLIGENT SYSTEMS 2013 || 68RD EDITION FALL 2012 EDITION TH
English
繁體中文
简体中文
CAPTURING THE
BIG DATA EXPLOSION The Age of Big Data, Intelligent Systems, and the Internet of Things...PAGE 4
SUPERCOMPUTER TECHNOLOGY REPORTS FOR DUTY High-Performance Embedded Computing with 4th Generation Intel® Core™ Processors...PAGE 9
BREAKTHROUGH PERFORMANCE FOR MEDICAL IMAGING Cutting-Edge Upgrades for the Spectrum of Imaging Applications...PAGE 12
SOFTWARE-DEFINED NETWORKING AND THE TELECOM CLOUD A Phased Approach to Implementation...PAGE 22
Subscribe
Editors’ Note
Big Ideas for Big Data What could you do with 35 Zetabytes of data? According to research firm IDC, connected devices will unleash this much data every year by 2020. This data flow is incredibly huge: a single Zetabyte can hold 36 million years of HD video. As this massive supply of data grows, so do the opportunities to analyze and act on this information. Think of highways that predict congestion and reroute vehicles before the roads clog. Kenton Williston Mark Scantlebury Imagine medical systems that analyze your Editor in Chief Associate Editor records to warn you of potential health issues before they occur. Or picture smart buildings that learn your habits and automatically minimize power consumption. To make this world a reality, our industry must deliver massive improvements in connectivity and compute power. Just supplying the bandwidth to transmit 35 Zetabytes will be a major challenge. Analyzing all this data will require strategic intelligence across the network, from the edge to the data center. In this edition of Embedded Innovator magazine we explore ways you can meet these challenges. We highlight techniques for optimizing telecom and energy infrastructure. We show how to boost signal processing up to 4x for advanced analytics. We examine ways retailers can collect and use data to actively engage consumers. And we illustrate ways factories can leverage Big Data to improve productivity. Throughout the magazine we showcase ways you can accelerate your designs by working with the IntelÂŽ Intelligent Systems Alliance, a global ecosystem of 250+ member companies that collaborate closely with Intel and each other to innovate with the latest technology. Together these companies are creating the solutions that are bringing Big Data to life. We hope you will join us in unleashing the possibilities.
About Us With the rise of high-performance, networked systems, the embedded industry is being transformed like never before. To help you stay ahead of the competition, Embedded Innovator magazine keeps you abreast of the latest trends with fresh design ideas and solutions. It’s full of insights to help you build the intelligent systems rapidly reshaping the world. Published quarterly, Embedded Innovator magazine is joined by a quarterly newsletter. Each newsletter features the same industry expertise as the Embedded Innovator magazine, but with a focus on specific vertical markets. Now you can keep connected with great ideas throughout the year. Visit intel.com/go/embeddedinnovator to subscribe, explore our archives, find additional content and solutions, or send an article to a friend. To learn about advertising opportunities, contact Felix McNulty at felix@intelintelligentsystemsalliance.com or +1 408 666 1270. To contact the publisher, email peggy.mahler@intel.com
Contents Capturing the Big Data Explosion 4
The Age of Big Data, Intelligent Systems, and the Internet of Things By James Robinson, General Manager, Intelligent Systems Group Segments and Broad Market Division, Intel® Corporation Supercomputer Technology Reports for Duty
9
High-Performance Embedded Computing with 4th Generation Intel® Core™ Processors By Eran Strod, System Architect, Curtiss-Wright Controls Defense Solutions Breakthrough Performance for Medical Imaging Cutting-Edge Upgrades for the Spectrum of Imaging Applications
12
By Cheryl Coupé, Roving Reporter for Intel Intelligent Systems Alliance ®
Optimize Video Traffic with Deep Packet Inspection Scalable Policy Enforcement from Gigabits to Terabits
17
By Paul Stevens, Telecom Sector Marketing Director, Advantech Networks & Communications Group Software Defined Networking and the Telecom Cloud 22
A Phased Approach to Implementation By Eric Gregory, Director, Platform Product Management and Chandresh Ruparel, Director, Product Marketing, Radisys Faster Signal Processing with Intel® Advanced Vector Extensions 2.0 Simplify Coding with Vectorizing Compilers and Libraries
28
By Noah Clemons, Technical Consulting Engineer, Embedded Computing & Debuggers, Intel® Corporation, Peter Carlston, Platform Architect, Intel® Intelligent Systems Group, Intel® Corporation and David Murray, Technical Director, N.A. Software Power up Security for the Smart Grid Retrofits and Rebuilds in a Two-Prong Approach
33
By Alexander Damisch, Director of Industrial Solutions, Wind River Transform the Factory with the Internet of Things Performance, Graphics, and Security with 4th Generation Intel® Core™ Processor
38
By Murray Slovick, Roving Reporter for Intel Intelligent Systems Alliance ®
Upgrade Point-of-Sale for the Online Age Transactional Improvements with 4th Generation Intel® Core™ Processors
42
By Kenton Williston, Editor in Chief Smart Digital Signage for Small Businesses Simplify Design and Deployment While Adding Performance By Michael Kelliher, Field Applications Engineer, IEI Copyright © 2013 Intel Corporation. All rights reserved. Intel, the Intel logo, Intel Atom, Intel Core, vPro, Xeon and Xeon Inside are trademarks of Intel Corporation in the U.S. and other countries. *Other names and brands may be claimed as the property of others.
48
Intel
®
Intelligent Systems Framework
Capturing
the
Big Data
Explosion The Age of Big Data, Intelligent Systems, and the Internet of Things By James Robinson, General Manager, Intelligent Systems Group Segments and Broad Market Division, Intel® Corporation
T
he market for Big Data is exploding. Billions of devices connected to the Internet of Things are accelerating a new era of data analytics that promises to revolutionize efficiency and create industry-shifting services. Intel believes that this year marks an inflection point when these Big Data opportunities will move into mainstream use across a broad range of markets including automation, energy distribution, retail, healthcare, and more.
But with all this Big Data comes big challenges in the development and management of the connected devices collecting the data and, in many cases, providing first-stage (edge) analysis. Intel and the 250+ global member companies of Intel® Intelligent Systems Alliance are leading the industry forward, collaborating closely to deliver solutions that provide the necessary: • • • • •
Intelligence to respond to data Interoperability to exchange data Security to protect systems Manageability to reduce total cost of ownership (TCO) Diverse ecosystem to meet the needs of different verticals and applications, as well as work cooperatively with each other to deliver complete solutions
A Convergence of Smart Analytics, Intelligent Systems, and Big Data Billions of devices are making trillions of connections to the Internet of Things, providing unprecedented opportunities for Big Data to improve lives and revamp entire industries. The stakes are huge for designers and developers. An IDC study sponsored by EMC Corporation, “The Digital Universe Decade—Are You Ready?”estimates the amount of digital information created and replicated worldwide will grow to more than 35 trillion Gigabytes of data. Intel estimates that this will lead to $3 trillion USD in new business opportunities (Figure 1). Realizing this opportunity requires the collection and processing of a vast amount of data at every step in the system— from sensor controllers to edge gateways, from cloud to client. In particular, this revolution requires using smart analytics
solutions to distribute analytics between intelligent systems at the edge and Big Data solutions in the data center to quickly, efficiently, and effectively deliver business intelligence, efficiency, and value. An excellent example is the disruption underway in manufacturing. Factory operational technology (OT) and information technology (IT) are combining to create Enterprise Resource Planning (ERP) systems and Manufacturing Execution Systems (MES) that connect the factory floor with the supply chain, enabling automation that extends to ordering, inventory, distribution, financial projection, and more. This level of automation is driving demand for intelligent systems that can overlay legacy control equipment with new classes of industrial IT and services, such as
+$3 Trillion New Business Opportunities Data Created at the Edge Drives New Insights and Business Opportunities and Optimization
+35T GB Data Generated
+15 Billion Smart Devices Driving Data IoT Driving Edge Devices and Data
Data Explosion From Edge Devices Drives Distributed Analytics
Figure 1.
4 | 2013 | 8 th Edition | Embedded Innovator | intel.com/embedded-innovator
Big Data is opening up huge opportunities (sources: “Intelligent Systems—Next Big Opportunity” IDC 2009; “The Digital Universe Decade—Are You Ready?” IDC 2012).
data node and acquisition devices, data hub processing, industry data center servers, and machine-to-machine (M2M) gateways that will enable all this data collection, connectivity, and analysis. Tapping the Opportunity Tapping the full potential of Big Data requires industry co-innovation across three pillars (Figure 2):
Smart Devices
Smart System of Systems
Smart Analytics
• Acquire Data Securely • Enable Seamless Interfaces • Create Value from Data Across • Intelligent systems that are the root of • Local Analytics and Filtering • Ensure Interoperability Smart Systems of Systems Big Data and enable acquiring this data • Root of Big Data Between Edge Systems • Provide Horizontal Building securely and performing local analytics • Secure and Federate Data Blocks for Vertical Between Cloud and Edge for End-end Analytics and filtering Analytics • Distribute Analytics at Edge • Smart systems of intelligent systems Systems and Datacenter that enable seamless interfaces, ensure interoperability between edge systems, and secure and federate data between Figure 2. Big Data requires industry innovation across three pillars. cloud and edge for analytics • Smart analytics that create value from data across the systems of intelligent systems, provide horizontal building blocks for vertical end-toIntel RCM with a robust set of products and services, offering everyend analytics, and distribute analytics at edge systems and in thing from off-the-shelf signage players to software integration to physthe data center ical installation services. This world-class support transforms signage deployment from a time-consuming, piecemeal process into a rapid, To succeed, this new generation of intelligent solutions must turn-key rollout. be highly flexible in order to address long-standing challenges, including industry fragmentation, diverse usage models, legacy Intel and members of the Alliance are continuously working together environments, complex interoperability needs, and requirements to to bring holistic solutions like this to the full range of Big Data applicameet both global and local market needs. The solutions must also be tions. Key focus areas for Intel and the Alliance include: designed to minimize security and reliability risks. The widespread communications required for Big Data will not be accepted without Intelligence to respond to data—Alliance hardware and software top-notch security. And once Big Data enables services that people solutions benefit from Intel’s continuing ability to pack more perfordepend on, it is a big problem if those services go down. mance into less space with greater power efficiency. For example, the Intel® Advanced Vector Extensions (Intel® AVX) 2.0 introduced in the What Intel is Doing 4th generation Intel® Core™ processor family deliver a 2x increase in Intel and the Alliance are working together to overcome these challenges, peak floating-point and fixed-point throughput, significantly speeding delivering solutions that provide the necessary performance, security, up edge analysis for applications such as machine vision. interoperability, manageability, and flexibility to usher in this new age of Big Data. A perfect example of these solutions is the new Intel® Retail Client Manager (Intel® RCM), which is bringing the benefits of Big Data to the digital signage market (Figure 3). Using Intel RCM, operators can broadcast marketing campaigns across distributed signage networks from anywhere. As illustrated in Figure 3, content can be adjusted in near real time based on weather conditions, shopper preferences, inventory levels, and other data gathered from the cloud.
Edge analysis plays a key role as well: Intel RCM can be used in conjunction with Intel® Audience Impression Metric Suite (Intel® AIM Suite), which detects viewer characteristics such as age and gender to play demographic-specific advertisements. Audit trails including view rates and customer interactions are fed back to the cloud, allowing operators to fine-tune campaigns. Intel RCM is currently available through Seneca Data and Avnet, both Associate members of the Alliance. These systems integrators support
Figure 3.
Intel® Retail Client Manager brings the benefits of Big Data to digital signage.
intel.com/embedded-innovator | Embedded Innovator | 8 th Edition | 2013 | 5
» Innovative computing platforms based on 4th generation Intel Core Processors « ®
» » » »
Enhanced performance More power efficiency High graphics and media power Improved security and manageability
™
Profit from our competencies » System- and OS-integration services » Customization and ODM services » Extended lifecycle management » Application and migration support » Excellent global technical support
Introducing the new stable, durable Dell OptiPlex XE2 and shared infrastructure PowerEdge VRTX platforms. COMe
Now available from Dell, the newest Premier Partner mITX cPCI Systems Alliance. Symkloud of the Intel Intelligent ®
Learn more at Dell.com/OEM
and more to come
Get more© information at: reserved. Dell, the DELL logo, the DELL badge are trademarks of Dell Inc. Other trademarks and trade names may be used in this 2013 Dell Inc. All rights document to refer to either the entities claiming the marks and names or their products. Dell disclaims proprietary interest in the marks and names of others. kontron.com/next-gen
The pulse of innovation
Intel® Intelligent Systems Framework
Moving deeper into the Internet of Things architecture, the new Intel® Xeon® processor E5-2600 v2 family offers an unprecedented 20 cores in an embedded two-socket configuration, providing excellent performance for packet processing and other high-performance applications. As the first multi-socket processors to use Intel’s state-of-the-art 22 nm fabrication process, the Intel Xeon processor E5-2600 v2 family delivers a 25 percent greater core count in the same thermal design power (TDP) as the previous generation. These performance upgrades are opening up exciting new opportunities across a range of markets. In the medical imaging market, for example, Alliance members are putting the performance to work in everything from portable ultrasound equipment to advanced molecular imaging—see page 12 for details. And in the mil-aero space, Alliance members like General member Curtiss-Wright are using the new processors to deliver supercomputer performance in mil/ aero platforms (see page 9). Best of all, these performance upgrades are easy to leverage thanks to development tools and libraries from the Alliance. See page 28 for one example from Affiliate member company N.A. Software. Interoperability to exchange data—Among the biggest problems today in realizing the Internet of Things is poor interoperability among individual devices and systems. Designers must spend untold hours analyzing and verifying connections and operation as new devices are added to a network. Intel® architecture-based solutions, offering a wide range of scalability in standards-based form factors, can be used to connect non-interoperable legacy systems together and provide a flexible platform for supporting many different protocols, data formats, and interfaces. For a great overview of how Associate member Radisys is using IA-based blade technology to enable interoperable software defined networking (SDN) and Network Functions Virtualization (NFV), check out their article on page 22. And for additional insights on network infrastructure, Premier member Advantech explains how Intel server-grade processors can accelerate video traffic on page 17. Security to protect systems—The 4th generation Intel Core processor family offers a wealth of security features to help protect the platform and data transfers. Intel® Platform Protection Technology enhances security through the bootup process via a root of trust based in the hardware and extends this safe boot capability to virtual machines. It also protects system BIOS against stealth attacks and unauthorized updates. Intel® Advanced Encryption Standard New Instructions (Intel® AES-NI) enables fast hardware-assisted data encryption to the disk and for transfers without slowing response times. Intel® Virtualization Technology (Intel® VT) protects memory space in hardware and helps prevent attacks from malicious software when used with appropriate software. In addition, Intel is continuing to work with Associate member McAfee, a wholly owned subsidiary of Intel globally recognized for its proactive and proven security solutions, to develop hardware-enhanced software security like McAfee DeepSAFE Technology*. These and other security features are already proving their worth across many market segments. To see how Alliance members are
putting them to work in the connected store, turn to page 42 for “Upgrade Point-of-Sale for the Online Age.” And to learn how Associate member Wind River is helping developers secure the smart grid, see page 33. Manageability to reduce total cost of ownership—Intel® vPro™ Technology enabled when processors are paired with the Intel® Q87 Chipset, delivers expanded management capabilities and improved power management to lower TCO. A new feature, Embedded HostBased Configuration, enables remote device provisioning without any person present at the device. To learn how Intel vPro technology features can minimize signage operating and maintenance costs, read “Smart Digital Signage for Small Businesses” from Associate member IEI on page 48. And for a perspective on the way Alliance members are lowering TCO in industrial automation, see “Transform the Factory with the Internet of Things” on page 38. Diverse ecosystem to meet the needs of different verticals and applications—Intel and the Alliance are the right team to enable this new age of Big Data, intelligent systems, and the Internet of Things. Intel has a long history of creating scalable technologies and supporting ecosystems that drive transformations in computing. Alliance members are forward thinking, and are looking for opportunities not only to streamline solutions, but to innovate to increase efficiencies and productivity in their markets. And as we have shown in this article, they offer cutting edge solutions for a vast array of verticals, including telecom, retail, signage, healthcare, industrial, energy, and mil-aero. Harnessing Big Data and the Internet of Things Intel is focusing considerable research and development, and its portfolio of silicon and technologies, on enabling businesses and consumers to benefit from the Internet of Things and the emerging Big Data economy. And the company is working with its Alliance partners who are creating a wide range of interoperable solutions designed to address connecting, managing, and securing devices and data in a consistent and scalable manner. Intel and the Alliance see 2013 as a year of rapid transition to this new age of Big Data. This transformation is unlike other changes in the industry because it is changing the rules of what intelligent systems can be and how they can drive actionable information at a much larger level. The world of things will never be the same. Explore Alliance products at intelintelligentsystemsalliance.com/solutions-directory
From modular components to marketready systems, Intel and the 250+ global member companies of the Intel® Intelligent Systems Alliance provide the connectivity, manageability, security, and performance developers need to create smart, connected systems. Learn more at: intel.com/intelligentsystems-alliance. Explore Alliance products and services at: intel.com/intelligentsystems-alliance-solutions.
Contact Intel
intel.com/embedded-innovator | Embedded Innovator | 8 th Edition | 2013 | 7
» Innovative computing platforms based on 4th generation Intel Core Processors « ®
» » » »
™
Enhanced performance More power efficiency High graphics and media power Improved security and manageability
COMe
mITX
cPCI
Profit from our competencies » System- and OS-integration services » Customization and ODM services » Extended lifecycle management » Application and migration support » Excellent global technical support
Symkloud and more to come
Get more information at: kontron.com/next-gen
The pulse of innovation
Military & Aerospace
Supercomputer
Technology Reports for Duty High-Performance Embedded Computing with 4th Generation Intel® Core™ Processors By Eran Strod, System Architect, Curtiss-Wright Controls Defense Solutions
H
igh-performance embedded computing (HPEC) brings supercomputer technology to the mil/ aero domain, delivering incredible performance in deployable designs. With the launch of the 4th generation Intel® Core™ processor family, HPEC systems are achieving breakthrough performance, with over 2x to 4x faster signal processing while reducing size, weight, and power, and cost (SWAP-C). In this article we will examine the features of the new processor and show how they can be used for HPEC. We will also demonstrate how a single 16-slot OpenVPX chassis can achieve over 19 trillion floatingpoint operations per second (TFLOPS)–a feat that required 2.4 chassis in the previous generation technology–while offering a 4x improvement in PCI Express* (PCIe*) backplane performance. Finally, we will highlight ways developers can speed and simplify their designs by working with members of the Intel ® Intelligent Systems Alliance, including General member Curtiss-Wright. From modular components to market-ready systems, Intel and the 250+ global member companies of the Alliance provide the performance, connectivity, manageability, and security developers need to create smart, connected systems. The High-Performance Revolution High performance computing (HPC) is solving the most challenging problems in science, engineering, and business. To do so, HPC supercomputers aggregate as many processing elements together as financial, space, and cooling budgets allow. Efficiently operating this many processors taxes interconnects, memory, peripherals, and storage, and the growth in HPC has led to significant advances on each of these fronts. HPEC adapts the distributed computing, I/O, and dataflow models from commercial HPC systems to mil/aero applications such as RADAR, signals intelligence, and electronic warfare. These sensor-processing applications require enormous data bandwidth and processing power. In fact,
the processing demands are practically unlimited; they are expanding as quickly as the industry can deliver more performance. At the same time, sensor-processing applications are typically deployed in scenarios with tightly limited SWAP-C constraints, such as airborne platforms, naval vessels, and ground vehicles. Thus there is intense pressure for HPEC solutions to maximize both performance and efficiency. Double the Performance, Then Double Again The mobile class 4th generation Intel Core processor meets these needs by significantly upgrading performance while maintaining the low-power, low-profile BGA packaging that is essential for rugged, deployed mil/aero embedded applications. The most notable upgrade is Intel ® Advanced Vector Extensions (Intel ® AVX) 2.0 instruction set, which doubles peak floating-point throughput, enabling a quad-core mobile class processor to achieve up to 307 GFLOPS at 2.4 GHz. Among other new features, Intel AVX 2.0 adds a fused multiply-add instruction (FMA3) which can perform a multiply and add in a single cycle. The multiply-add is a key building block of signal processing algorithms, so this improvement is a huge leap for HPEC computing. Many improvements have been made in the microarchitecture to allow applications to realize this greater performance potential: • The memory pipeline can now perform 2 loads and a store on each cycle • L1 cache bandwidth has doubled to 96 bytes/cycle (64 byte read plus 32 byte write) • L2 cache bandwidth has also doubled to 64 bytes/cycle These upgrades along with the internal Last Level Cache, 320 GBps Ring Bus, and DDR3 dual-channel memory (with a peak memory bandwidth of 25 GBps at 1,600 MHz) enable the 4th generation Intel Core processor family to speed up critical RADAR, imaging, and signal processing algorithms significantly. In fact, members of the Alliance have already performed benchmarks to quantify the improvements. For example, Affiliate member N.A. Software’s benchmarks on its VSIPL library show a 2.4x improvement on a real-to-complex out-of-place 2D Fast Fourier Transform (FFT). (For more details on Intel AVX 2.0 see “Faster Signal Processing with Intel® Advanced Vector Extensions 2.0” on page 28.)
intel.com/embedded-innovator | Embedded Innovator | 8 th Edition | 2013 | 9
Military & Aerospace
As if that was not enough, Intel’s mobile class devices now support OpenCL 1.2 programming on the integrated graphics execution units. The specialized architecture’s single-instruction multiple data (SIMD) engines offer another 352 GFLOPS of the raw performance—more than doubling the overall compute potential—while adding only a few watts of power consumption. This on-chip general-purpose graphic computing (GPGPU) engine is an attractive alternative to solutions that add separate graphics co-processors, which can consume 50 W or 75 W. Ready for HPEC Deployment HPEC applications rely on the well-established OpenVPX (VITA 65) system standard to deploy rugged and resilient processing solutions. To pack more processing into a single OpenVPX payload slot, Curtiss-Wright Controls Defense Solutions has introduced the dual-processor CHAMP-AV9, which brings the functionality of two single board computers (SBCs) into a single OpenVPX module (Figure 1). Optimized for HPEC scalability, this digital signal processing (DSP) engine features two Intel® Core™ i7-4700EQ processors with a total of 614 CPU GFLOPS and 704 GPGPU GFLOPS. Using CHAMP-AV9
Figure 1.
in a dual-star 16-slot OpenVPX backplane, a system can achieve 19 TFLOPS of processing in a 19" wide 6U chassis. This industry-leading performance more than doubles the capabilities of previous Curtiss-Wright DSP solutions while maintaining the same thermal envelope. The CHAMP-AV9 is supported with operating systems and development tools including Wind River* VxWorks*, Linux*, and Gedae (Wind River is an Associate member of the Alliance). Curtiss-Wright provides signal processing libraries and data movement middleware for message passing and bulk data transfers. The Curtiss-Wright Continuum Insights suite of multi-processor development tools provides system management, development, and event viewing features to lower the cost of developing multi-core, multi-processor software. The CHAMP-AV9 can also be supported by third-party tools from Alliance members, such as the N.A. Software VSIPL library mentioned earlier. This robust support significantly speeds development and provides a seamless migration path for both existing CHAMP-AV DSP applications and large, distributed node HPEC systems.
The CHAMP-AV9 offers a total of 614 CPU GFLOPS and 704 GPGPU GFLOPS.
4th Generation Intel® Core™ i7 Processor
4th Generation Intel Core i7 Processor
x16
x16
PCI Express* 3.0 Switch x8
x8
x8
x8
ADC DAC ADC DAC Figure 2.
PCIe architecture lends itself well to multi-node topologies.
10 | 2013 | 8 th Edition | Embedded Innovator | intel.com/embedded-innovator
Feeding the Beast Processing power demands equally fast I/O interfaces. The 4th generation Intel Core processor includes a PCIe 3.0 interface with three controllers that can be configured as 1 x 16, 2 x 8, or 1 x 8 + 2 x 4 (on the CHAMP-AV9 board this interface remains in 1 x 16 mode for the fastest possible throughput). PCIe 3.0 improves on the previous generation PCIe 2.0 in two ways. First it increases the bit rate per lane from 5 Gigatransfers (GT) to 8 GT per second. Second, it moves from 8b/10b encoding to the more efficient 128b/130b encoding scheme. With fewer bits devoted to encoding, the throughput rate more than doubles to 8 Gbps per lane. With 16 lanes, the processor can send and receive 16 GBps of data simultaneously. The PCIe architecture lends itself well to multi-node topologies that interface with end-point devices such as analog-to-digital converters (ADC) or digital-to-analog converters (DAC). Figure 2 shows two 4th generation Intel Core processors each controlling an ADC and DAC via a Gen 3 PCIe switch. Note that the PCIe non-transparent bridging capability of the PCIe switch enables multiple CPUs to be connected to the same switch without having a root device conflict. As shown in Figure 3, the CHAMP-AV9 supports 32 lanes of PCIe 3.0 (configurable as 8 x 4 lanes, 4 x 8 lanes and 2 x 16 lanes) to connect to a battery of I/O peripherals including standard XMC carriers, and heterogeneous processing elements. Additionally, CHAMP-AV9 supports four 40 Gigabit Ethernet (GbE) or InfiniBand FDR10 interfaces for a total OpenVPX backplane interface of 112 Gbps or 14 GBps full duplex non-blocking fabric bandwidth per payload module. At a chassis level, this totals 196 GBps of non-blocking fabric bandwidth. The CHAMP-AV9 also offers 32 GBps of expansion plane performance.
These performance levels are significant upgrades over the previous-generation CHAMP-AV8, which used PCIe 2.0. Specifically, the CHAMP-AV9 delivers 2x the fabric bandwidth and 4x the PCIe interconnect bandwidth over the previous generation. As illustrated in Figure 3, the 4th generation Intel Core processors provide additional I/O interfaces accessible via the Peripheral Controller Hub (PCH), including:
Legend 4x10 40 Gbps DDR3 SDRAM 4 GB DDR3 SDRAM 4 GB
x8 PCI Express* (PCIe*)
DDR3 SDRAM 4 GB DDR3 SDRAM 4 GB
96-Lane
PCI Express* 3.0 Switch
4th Generation Intel® Core™ i7 Processor
Recovery Flash
Platform Boot Flash 16 MB SPI_A Control Hub (PCH) SATA (NAND) Flash 16 GB SATA TPM SPI_A
4th Generation Intel Core i7 Processor
Ethernet
Ethernet
Ethernet
Ethernet
Recovery Flash
Platform Boot Flash Control Hub SPI_B 16 MB S (PCH) SATA (NAND) SATA
Build Option
Build Option
x16 PCIe
Flash 16 GB TPM
SPI_B
4x EIA-232 • Display—The processors support 2x EIA-422 high-resolution image rendering (EIA-485) 40 Gb 40 Gb interfaces such as embedded 2x EIA-422 IPMI (EIA-485) DisplayPort (eDP). DisplayPort 16x GPIO allows the chaining of devices so that many devices can be run off 2x 1000 1000 P1 2x8, 2x16 1000 1000 2x GPIO VPX Backplane Serial Serial SATA USB USB SATA Display Port Base-KX Base-T 4x4 P2 & P5 Base-T Base-KX Display Port 16-bits a single port. In HPEC computing, Connector display interfaces like these are Front Panel 1000 1000 Display Port USB SATA 232 Serial SATA USB Display Port Base-T Base-T Connector tremendously useful to foster situation awareness • Storage—Onboard Serial ATA Figure 3. The CHAMP-AV9 offers ample I/O. (SATA) supports the robust storage capability essential to archiving data for later review • Peripherals—USB connects to many electronic devices including Breakthrough HPEC Performance keyboards, pointing devices, and other adapters With 2x to 4x the signal processing performance, nearly 2x the fabric bandwidth and 4x the PCIe interconnect bandwidth over the preMil-Grade Security vious generation, CHAMP-AV9 offers a breakthrough in performance HPEC systems by their very nature require security, as the war of inforfor HPEC applications. This advance is due to the unprecedented mation has become a major threat both to global commercial indusperformance and peripheral integration of the 4th generation Intel tries and to the military. Intel® Advanced Encryption Standard New Core processor, which includes the Intel AVX 2.0 instruction set and Instructions (Intel® AES-NI) refer to the set of instructions available OpenCL-programmable on-chip graphics engine. Whether you are with the 4th generation Intel Core processor family. The instructions creating a new design or migrating an existing sensor-processing enable rapid and secure data encryption and decryption based on the application, the Intel platform offers a compelling combination Advanced Encryption Standard (AES). of performance and efficiency that meets the needs of SWAP-C constrained deployments. The instructions were designed to implement some of the complex and performance-intensive steps of the AES algorithm in hardware, For more information on the Curtiss-Wright CHAMP-AV9, accelerating execution by up to 10x over software-only implementasee intel.com/SD-champ-av9 tions. Originally introduced in 2010, Intel AES-NI has been upgraded in the Intel 4th generation Intel Core processor family with six new instructions that perform several compute intensive parts of the AES For more on performance and security in mil/aero, see algorithm. These instructions can execute using significantly less intel.com/embedded-milaero clock cycles than a software solution. Four of the new instructions are for accelerating the encryption/decryption of a round, and two new instructions are for round key generation. Curtiss-Wright Controls Defense Contact Curtiss-Wright Solutions (intel.com/ea-cwcds) is a Also relevant to security, each processor in CHAMP-AV9 is equipped General member of the Intel® Intelligent Systems Alliance and a with 16 MB of flash memory used to store the system BIOS and leader in rugged electronic modules and systems for defense power-on diagnostics firmware. In the event of accidental corruption applications. Its broad engineering capabilities combine systems, of the primary boot flash, both processors may boot from a secondary software, electrical, and mechanical design expertise with recovery flash in order to regain operation of the card and initiate comprehensive program management and a broad range of life-cycle recovery of the primary flash. support services.
intel.com/embedded-innovator | Embedded Innovator | 8 th Edition | 2013 | 11
Healthcare
Breakthrough Performance for
Medical Imaging
Cutting-Edge Upgrades for the Spectrum of Imaging Applications By Cheryl Coupé, Roving Reporter for Intel® Intelligent Systems Alliance
Figure 1.
Patient Infotainment
Medical Gateways
Histopathology Ventillators
Anesthesiology
Lab Instruments, Analyzers, Blood Diagnostics
Robotic Surgical Systems
XRAY
Digital Pathology
IVUS POC Handheld
Intel® Atom™ Processors
Low Power Value ~40 W Low Power Entry 10-20 W Entry Performance 8-10 W Ultra Low Power <5 W
Fitness Equipment
Intel® Atom™ Processors
PET
Fanless 15-25 W
ECGs
Thin Blade/ Mezzanine 35-50 W Mainstream Performance 30-65 W Value Performance 30-65 W Low Power Performance 30-45 W
MR
OCT
Blades 50-100 W
Point of Care Terminals Imaging Stations
We will also highlight off-the-shelf medical solutions from members of the Intel ® Intelligent Systems Alliance. From modular components to market-ready healthcare systems, Intel and the 250+ global member companies of the Alliance provide the performance, connectivity, manageability, and security developers need to create smart, connected systems. Close collaboration with Intel and each other enables Alliance members to innovate with the latest technologies, helping developers deliver firstin-market healthcare solutions to improve patient care and electronic recordkeeping.
Full Performance >100 W
Patient Monitoring
In this article we will show how developers and OEMS can take advantage of the full spectrum of opportunities using the latest Intel® architecture (IA) processors. We will examine new imaging features, such as the 2x faster image processing in the 4th generation Intel® Core™ processors and an unprecedented 20 cores available in Intel® Xeon® processor E5-2600 v2.
Ultrasound
A
fter a lull in which economic crises and healthcare policy changes slowed medical imaging equipment sales, the picture is definitely brightening. Frost & Sullivan expects the global market to hit $30 billion USD by 2017, with much of the growth coming from emerging markets (source: 2013 Global Medical Imaging Equipment Market Outlook Report). All segments offer opportunities. Low-end modalities like ultrasound are migrating into new form factors and applications, while high-end modalities like MRI are increasingly adopted in emerging markets.
Intel® architecture processors can address a wide range of medical applications.
12 | 2013 | 8 th Edition | Embedded Innovator | intel.com/embedded-innovator
The Scalability Advantage IA processors offer the performance, energy efficiency, and features for a wide range of modalities, from advanced CT, MRI, and molecular imaging scanners, to mammography, X-ray, and ultrasound. As shown in Figure 1, the new Haswell microarchitecture, as implemented in the 4th generation Intel Core and Intel® Xeon™ processor E3 v3, delivers the superior image-processing performance and enhanced security for next-generation devices such as portable ultrasounds and dental radiography. At the other end of the spectrum, the Intel Xeon processor E5-2600 v2 offers exceptionally high performance along with energy efficiency for lower total cost of ownership for the most demanding imaging applications. This scalability offers several important advantages. First, medical OEMs can scale their products to different price and performance points, and can quickly diversify into new markets. Second, the consistent architecture and toolsets speed and simplify development, letting OEMs focus on their application expertise instead of learning new hardware. Finally, the entire IA roadmap is now marching forward with a “tick-tock” cadence that is consistently delivering new levels of performance and power efficiency. This gives developers a clear upgrade path and helps them maintain software designs across product generations. On the development side, the IA platform is supported by a rich set of tools, libraries, and development kits that provide portable, optimized signal ® ® processing Lauterbach-Atom_7x4,875in.pdf from Intel® Atom™ processors to Intel 1 28.06.2013 08:33:13Xeon processors. ® One key example is the Intel System Studio, which is discussed on
C
M
Y
CM
MY
JTAG Debugger for Intel® Atom TM Processor
page 28. Other tools include Intel® Math Kernel Library (Intel® MKL), the Intel® Integrated Performance Primitives (Intel® IPP), and the Intel® Signal Processing Development Kit (Intel® SPDK). In addition, developers have access to a rich ecosystem of third-party tools and operating systems (OSs) from Alliance members such as Associate members Microsoft and Wind River, and Affiliate member Green Hills Software. 8x Floats 4x Doubles Intel Advanced Vector Extensions (Intel® AVX) ®
16x Bytes 8x 16-bit Shorts 4x 32-bit Integers 2x 64-bit Integers 1x 128-bit Integers 8x Floats w/FMA3 4x Doubles w/FMA3 32x Bytes
Intel® AVX 2.0
16x 16-bit Shorts 8x 32-bit Integers 4x 64-bit Integers 2x 128-bit Integers
Figure 2.
Intel® Advanced Vector Extensions (Intel® AVX) 2.0 doubles peak floating-point and fixed-point throughput.
Intel® AtomTM Processor
► Multicore debugging (SMP/AMP) including Hyper-Threading
CY
CMY
K
► Comprehensive debugging of UEFI based BIOS ► OS-Awareness, Support for Linux, Windows CE and others
www.lauterbach.com intel.com/embedded-innovator | Embedded Innovator | 8 th Edition | 2013 | 13
4 th Generation Intel ® Core TM Processors Powerful‧Flexible‧Scalable‧Innovative WADE-8015 Intel ® Core TM i5/i7 processors Intel ® Q87 Chipset Targets at Digital Signage, POS, Industrial, Gaming, Automation and Embedded Control
CAR-4020 Intel ® Xeon ® processor E3 Intel ® C226 Chipset Targets at Network Security and Network Management
ROBO-8112VG2AR Intel ® Core TM i3 processor & Intel ® Xeon ® processor E3-1200 v3 series Intel ® C226 Chipset Targets at Medical, Factory Automation, Digital Signage, Storage, Image Processing and Military
CAR-3040 Intel ® Core TM i3/i5/i7 processors Intel ® Q87 Chipset Targets at Network Security and Network Management
Healthcare
Performance Drives Imaging Innovation Parallel processing is key to medical imaging, and the 4th generation Intel Core processors and Intel Xeon processor E3 v3 deliver parallelism in spades. In addition to offering four cores with up to 15 percent faster performance than previous generations, these processors benefit from the Intel® Advanced Vector Extensions (Intel® AVX) 2.0 illustrated in Figure 2 (Page 13). This upgraded vector-processing technology introduces a fused multiply-add (FMA3) that effectively doubles the peak floating point throughput in comparison to the previous generation. Multiply-add workloads are a critical component of image processing, so this upgrade will significantly benefit medical imaging applications. Intel AVX 2.0 also expands most integer Intel AVX instructions from 128 bits to 256 bits, doubling fixed-point performance as well. And with gather support, Intel AVX now enables vector elements to be loaded from noncontiguous memory locations to simplify code vectorization. The Haswell microarchitecture also offers an upgraded graphics engine with a 60 percent 3D graphics improvement over previous generations. Enhanced high-resolution display capabilities support up to 4K resolution on three independent displays. What’s more, the new “collage display” mode can present multiple displays to the OS as a unified resolution. These features are all invaluable for display of high-resolution medical images, and they reduce the need for expensive and power-hungry discrete graphics cards. Security and Manageability In addition to performance upgrades, the Haswell microarchitecture includes significantly upgraded security features to help protect patient data and enhance reliability. New and enhanced security measures include Intel® Platform Protection Technology with BIOS Guard, which cryptographically verifies the BIOS to keep malware out. Intel® Platform Trust Technology and Intel® Boot Guard (available on the forthcoming U-Series processors) support Microsoft* Windows* 8 secure and measured boot, and works with Trusted Platform Module (TPM) 2.0 to protect against boot-level malware. Finally, Intel® Data Protection Technology with Intel® Advanced Encryption Standard New Instructions (Intel® AES-NI) enables rapid and secure data encryption and decryption to protect patient records without burdening the processor.
Figure 3.
ADLINK’s Express-HL and Express-HL2 COM Express modules combine high performance with compact design.
On the manageability side, Intel® vPro™ Technology with Intel® Active Management Technology (Intel® AMT) enables remote updates and repairs, even if the equipment is powered off, ensuring that vital medical equipment can be easily monitored, maintained, and repaired. And, of course, the IA hardware features are enhanced by complementary McAfee solutions that ensure platforms stay protected and up-to-date according to a health organization’s IT policies (McAfee is an Associate member of the Alliance). Solutions for Medical Imaging Off-the-shelf module and board solutions from the Intel Intelligent Systems Alliance enhance the Haswell micro-architecture’s benefits, giving developers standards-based platforms that can jump-start development and speed time-to-market. For example, Associate member ADLINK’s Express-HL and Express-HL2 COM Express modules combine high performance with a compact design for applications like cart-based ultrasound (see Figure 3). COM Express modules support semi-custom designs where high-speed, complex circuitry is contained on the module, freeing developers to focus on custom I/O implementation on the carrier board. ADLINK’s modules offer two variants on this theme: The Express-HL uses a COM Express Type 6 interface with cutting-edge interfaces like DDI, PCIe Gen3, SATA 3.0, Gigabit Ethernet (GbE), and USB 3.0. In contrast, the Express-HL2 features the legacy COM Express Type 2 pinout with interfaces like LVDS, Analog CRT, PCI bus, and IDE. In another approach, the AIMB-274 Mini-ITX motherboard from Premier member Advantech comes with ample built-in I/O that is easy to expand, upgrade, and maintain (see Figure 4). It is suitable for both full-size equipment as well as space-constrained portable or labbased systems. Notable features include dual GbE, SATA 3.0, USB 3.0, three-display support, PCIe x16 (Gen 3), and two mini PCIe ports to support mSATA drive and various mini PCIe modules. More Cores for Advanced Imaging For high-end equipment such as CTs and MRIs, time-to-image is critical, but the market is also demanding higher-resolution images along with energy efficiency to help meet hospital “green” initiatives and reduce total cost of ownership. Here, the Intel Xeon processor E5-2600 v2 family fits the bill.
Figure 4.
The Advantech AIMB-274 has ample built-in I/O.
intel.com/embedded-innovator | Embedded Innovator | 8 th Edition | 2013 | 15
Healthcare
Intel® Xeon® Processor E5-2600 v2 Series (up to 10 cores) 4 memory channels up to DDR3 1,866 MHz per processor (maximum 3 DIMMS per processor)
x4 DMI 2.0
Two (2) Intel® QuickPath Interconnect Links
x8 PCI Express 2.0
4 memory channels up to DDR3 1,866 MHz per processor
40 Lanes PCI Express Gen 3.0 Per Processor
14 USB 2.0 ports 2 SATA Gen 3.0 Ports (6 Gbps) 4 SATA Gen 2.0 Ports (3 Gbps)
Intel® Xeon® Processor E5-2600 v2 Series (up to 10 cores)
(maximum 3 DIMMS per processor)
4 Integrated SAS Ports (3 Gbps)* Intel® C604 or Intel® C602-J Chipset
Integrated Gigabit Ethernet Controller (MAC) Compatible with Intel® 82579LM Gigabit Ethernet PHY
(Figure 6). Each SuperBlade module can house two dualsocket nodes per TwinBlade, for a total of up to four Intel Xeon processors E5-2600 v2 per blade. Up to 10 TwinBlades (20 nodes) can be installed in a 7U enclosure, providing more than 65 percent density savings compared to 20 1U servers. Power, cooling, and networking devices are integrated in the rear of the chassis to optimize space even more.
The blades also offer high performance, dense I/O, Figure 5. The Intel® Xeon® processor E5-2600 v2 offers up to 20 cores in two sockets. including InfiniBand or 10G mezzanine HCA cards, optional PCIe 3.0 expansion card, This processor series is the first to deliver up to 20 cores in an support for up to 6 SATA3 or SAS2 hot-plug hard drives, as well as GbE embedded two-socket offering, as illustrated in Figure 5. These parallel and FDR/QDR IB switches and redundant ports on a per node basis. This processing capabilities provide significant performance improvements high level of integration enables excellent power efficiency. Supermicro for imaging applications compared to previous generations. Imaging estimates that its blade architecture provides 30 to 50 percent energy performance is further boosted with Intel® AVX Float 16, an enhanced savings over three-year deployment when compared to 1U servers and version of the original Intel AVX instruction set with support for 16-bit external switches. half-precision floating point data. To facilitate high-bandwidth imaging, the higher core counts are supported by up to 25 MB of on-chip Focusing on the Opportunity cache (up from 20 MB), four channels of 1,866 MHz DDR3 (up from Medical imaging is a booming market, helping save lives and 1,666 MHz) and 40 lanes of PCIe per socket. enhance patient outcomes with increasingly advanced diagnostics. Every aspect of the market is ripe with opportunity, from portable As the first multi-socket processors to use Intel’s state-of-the-art ultrasound to next-gen MRI. Medical OEMs can make the most of 22 nm fabrication process, the Intel Xeon processor E5-2600 v2 the opportunity with IA processors, which offer solutions for the full product family continues Intel’s focus on reducing total cost of spectrum of applications. And by working with Alliance members, ownership by improving energy-efficient performance. Specifically, medical OEMs can quickly take advantage of the platform’s high the new processors boast a 25 percent greater core count in the same performance, efficiency, and security features. thermal design power (TDP) as the previous generation. For information about healthcare solutions from the Alliance members offer a number of servers that employ the new Alliance, see intel.com/SD-medical processors for high-end imaging applications. For example, Affiliate member Supermicro offers the SuperBlade* series of high-density blade servers for medical applications such as 3D rendering For more on delivering quality healthcare with secure, connected devices, see intel.com/embedded-medical
From modular components to marketready healthcare systems, Intel and the 250+ global member companies of the Intel® Intelligent Systems Alliance (intel.com/intelligentsystems-alliance) provide the connectivity, manageability, security, and performance developers need to create smart, connected systems. Advantech (intel.com/ea-advantech) is a Premier member of the Alliance; ADLINK (intel.com/ea-adlink), McAfee (intel.com/ea-mcafee), Microsoft (intel.com/ea-microsoft), and Wind River (intel.com/ea-windriver) are Associate members; Green Hills Software (intel.com/ea-ghs) and Supermicro (intel.com/ea-supermicro) are Affiliate members.
Contact Intel
Figure 6.
A Supermicro TwinBlade incorporates up to four Intel® Xeon® processors E5-2600 v2.
16 | 2013 | 8 th Edition | Embedded Innovator | intel.com/embedded-innovator
Deep Packet Inspection
Optimize Video Traffic with
Deep Packet Inspection Scalable Policy Enforcement from Gigabits to Terabits By Paul Stevens, Telecom Sector Marketing Director, Advantech Networks & Communications Group
M
obile operators cannot build capacity fast enough to keep up with the explosive growth in video traffic. The resulting bandwidth constraints are making it difficult for operators to maintain quality of service (QoS) and quality of experience (QoE) for video traffic. To meet the challenge, operators must move to a more dynamic, content-driven Policy and Charging Enforcement Function (PCEF) approach based on deep packet inspection (DPI).
This article describes how mobile operators can implement the latest QoE and video optimization techniques on Intel® architecture, allowing them to deploy new and higher quality video services faster, improve the subscriber experience, and reduce the cost of content delivery. We show how the Intel architecture can implement DPI without specialized network processors, enabling the use of off-the-shelf hardware for rapid deployment and scalability.
The key to removing this bottleneck is recognizing that video, voice, gaming, and browsing require different network characteristics. Applications
We also present a proof-of-concept that brings together off-the-shelf components to implement a QoE solution. This proof-ofconcept demonstrates the value of working with members of the Intel® Intelligent Systems Alliance like Premier member Advantech. From modular components to market-ready communications platforms, the 250+ global member companies of the Alliance collaborate closely with Intel and each other to innovate with the latest technologies, helping developers deliver first-in-market communications infrastructure and reduce cost. The Consolidation Opportunity As shown in Figure 1, the PCEF implements operator policies initiated by the Policy and Charging Rules Function (PCRF)—such as ensuring adequate per-use bandwidth—in
the 3G and LTE packet core. As mobile bandwidth grows, today’s PCEF may lack the DPI and packet processing capabilities needed to keep up with the surging traffic rates of 100 Gbps or more.
Portal
IP
AAA (Radius)
Policy Server
Billing & Charging OCS
PCRF Gx
Gy Core
SGSN
Figure 1.
GGSN
SCE PCEF
Internet
The Policy and Charging Enforcement Function (PCEF) is responsible for managing traffic based on operator policies. intel.com/embedded-innovator | Embedded Innovator | 8 th Edition | 2013 | 17
We Put a Whole New Twist on Configurability No matter the workload, your ideal system is now just a twist away
At Advantech, we've created a whole new twist on how to optimally configure and accelerate the throughput of your next generation networking platform. With our Customized COTS (C2OTS) framework, there are hundreds of high performance variations. The choice and flexibility is all yours.
Fabric Mezzanine Modules (FMMs) Plug in a quad 40G FMM for 160G of Fabric I/O per ATCA blade Add more acceleration using FMMs based on Intel® QuickAssist Technology Design your own FPGA-based FMM - we'll share the spec
Network Mezzanine Cards (NMC) Deploy NMCs across all your appliances to scale from 4 x gigabit Ethernet to 16 x 40GbE
Optimal Configurations. Accelerated Performance. Configurability with a whole new twist!
FWA-3220 1U rackmount appliance Intel® Platform for Communications Infrastructure processors 2 NMC slots
www.advantech.com/nc
FWA-6510 2U rackmount appliance Intel® Xeon® processor E5-2600 Up to 8 NMC slots
Netarium™-14 14-slot ATCA System Over 200 Intel® Xeon® Processor cores 1.28 Tbps switching capacity
Advantech Headquarters
No. 1, Alley 20, Lane 26, Rueiguang Road, Neihu District, Taipei 114, Taiwan, R. O. C. Tel: 886-2-27927818 Email: NCG@advantech.com
Deep Packet Inspection
For example, users are highly sensitive to video freezing, but tolerant of momentary delays while web browsing. By implementing policies that are application-specific, mobile operators can increase QoE without necessarily using more network capacity. This entails a careful analysis of the different traffic types and content. Video quality, in particular, is dependent on frame rate, delay, jitter, size, and codec. By extracting these parameters from the packet stream and analyzing video quality, mobile operators can implement policies to ensure appropriate QoE. An effective solution must include DPI to identify content for analysis, video decode to harvest the content, and video analysis to score (assign priority rating) video traffic. Thus, the PCEF must be built on a high-performance, scalable packet processing platform with both I/O optimization and application acceleration features. The Intel® Platform for Communications Infrastructure shown in Figure 2 provides an excellent match for these requirements. This platform incorporates three key elements for PCEF applications: • Packet Processing—the Intel® Xeon® processor E5-2600 v2 family at the heart of the solution offers up to 20 cores and 80 lanes of PCI Express* 3.0 (PCIe* 3.0) in a dual-socket configuration, enabling up to 160 Gbps of packet processing without specialized NPUs • Video Optimization—the processors support Intel® Advanced Vector Extensions (Intel® AVX), which boosts vector processing performance by up to 2x compared to the previous generation, making the processor well-suited to video analysis • DPI—the Intel® Communications Chipset 89xx Series incorporates Intel® QuickAssist Technology, which accelerates cryptography and compression tasks The platform software includes the Intel® Data Plane Developer Kit (Intel® DPDK), a set of libraries and drivers for multi-core packet processing offload, and the Intel QuickAssist Technology framework and application programming interfaces (APIs). Together these software packages enable the seamless integration of hardwareaccelerated offload that can migrate to future architectures without changing the source code. This is an important benefit for rapidly-evolving PCEF applications, as it allows the Intel platform to quickly adapt to new requirements.
Another crucial advantage of the Intel platform is its ability to consolidate multiple workloads onto a single architecture. By eliminating specialized NPUs, the Intel platform offers a unified architecture for multiple workloads, including video optimization, content filtering and adaptation, targeted advertising, analytics, and monetization of new services. Consolidating all of these workloads onto a single platform reduces complexity, simplifies platform management, and drives greater cost efficiencies. In addition, operators are finding that moving all workloads onto Intel architecture facilitates migration into a virtualized cloud or carrier cloud, enabling the transition to softwaredefined networking (SDN). Proof-of-Concept Demonstration Advantech has worked with Qosmos (an Affiliate member of the Alliance) and Argon Design to develop a QoS/QoE video scoring proof-of-concept based on the Intel platform. As shown in Figure 3, the proof-of-concept
Intel® Xeon® Processor E5-2600 v2 Series (up to 10 Cores)
4 Memory Channels of up to DDR3 1,866 MHz Per Processor (Maximum 3 DIMMs Per Channel)
PCI Express* x4, x8, or x16
Intel® QuickAssist Technology 4 Integrated 10/100/1000 GbE MACs
2 Intel® QuickPath Interconnect Links
Intel® Xeon® Processor E5-2600 v2 Series (up to 10 Cores)
40 Lanes PCI Express* Gen 3.0 Per Processor
DMI
4 Memory Channels of up to DDR3 1,866 MHz Per Processor (Maximum 3 DIMMs Per Channel)
4 PCI Express Gen 1.0 Intel® Communications Chipset 89xx Series
2 SATA Gen 2 Ports; Port Disable 6 Hi-Speed USB 2.0 Ports
Intel® Management Engine Ignition Firmware and BIOS Support
Figure 2.
The Intel® Platform for Communications Infrastructure is well suited to video traffic management. Advantech FWA-6500 Video Server
Up to 8x 10 GbE Pipes
Advantech FWA-6510 Running Argon Design Qualmon Employing Deep Packet Inspection on Qosmos ixEngine
Figure 3.
Qualmon Monitor
The video analysis proof-of-concept uses the Advantech FWA-6510 and Qosmos ixEngine. intel.com/embedded-innovator | Embedded Innovator | 8 th Edition | 2013 | 19
Deep Packet Inspection
is able to recognize video streams, extract details of the streams from the container format, and extract a representative video thumbnail.
for up to 1,300 protocols. For video, the parameters extracted include frame rate, delay, jitter, size, and codec.
The solution can recognize the media player in use and replicate its buffering model in order to simulate the player and display the buffer margin as a time graph, as shown in Figure 4. This enables the QoS/ QoE application to predict when a video player will freeze due to data starvation. The application also intercepts and decodes video samples, and then performs image processing to assess the visual quality and provide a numeric quality measure.
The metadata extracted by the Qosmos ixEngine can be used in a policy control platform to implement an application-specific policy. Additional content extracted from the packet stream can be used to further analyze application-specific performance.
The proof-of-concept is built on the Qosmos ixEngine*, which leverages the Intel DPDK to extract metadata and content from packets flowing through the network (Figure 5). The Qosmos ixEngine can be used with any processor supported by Intel DPDK, including the Intel Xeon processor E5-2600 v2 family. This flexibility is a key benefit of using Intel® architecture for network processing. It also illustrates the benefits of working with members of the Alliance, whose hardware and software offerings can be readily combined to create highly complete solutions. The Qosmos ixEngine supports real-time Layer 7 IP flow analysis and identification of protocols and applications based on flow passing and statistical analysis. It can extract more than 6,000 types of information
Scaling to Terabits per Second The proof-of-concept solution previously described was built using the Advantech FWA-6510 Network Application Platform with up to 160 Gbps of throughput. This same software can scale seamlessly beyond 1 Tbps using these same processors on Advanced TCA* (ATCA*) and extended ATCA (eATCA) architectures. As shown in Figure 6, eATCA systems combine standard off-the-shelf ATCA blades with extended rear transition modules (eRTM) that provide almost four times more real estate than traditional rear transition modules (RTMS) for PCIe-based I/O and acceleration hardware. Specifically, the eRTMs support up to 350 W thermal design performance (TDP) to facilitate blade extensions such as inline acceleration, specialized eRTMS with switch-based I/O, or 10 Gigabit Ethernet (GbE) and 40 GbE ports using Advantech’s Network Mezzanine Cards (NMCs). Blade configurations are similarly flexible. Blade capacity ranges from 2 to 14 slots per system, supporting ATCA blades, including Intel architecture blades and switches in both dual-star and dual dual-star topologies, with up to 350W TDP per slot. For top-of-the-range networking gear, 14-slot systems can support up to four hub blades, each with dedicated eRTMs offering over 2 Tbps backplane switching capacity. Hub blade eRTMs can provide full system I/O capabilities along with fan-out switches, traffic managers, load balancing or even customized functions.
Figure 4.
Screenshot of the Argon Design Video Quality Monitor.
Application (Video Analysis) Flow Manager
Metadata Content Qosmos ixEngine* DPI Engine
Control Plane/Mgt
Packets Intel® Data Plane Developer Kit (Intel® DPDK) Core 1 Figure 5.
Core 2
…
Core n
Core
Qosmos ixEngine* performs DPI and metadata analysis.
20 | 2013 | 8 th Edition | Embedded Innovator | intel.com/embedded-innovator
The eATCA architecture is a good fit for PCEF applications for several reasons. The ability to scale eATCA systems to over 1 Tbps in a single system meets the performance needs of today’s high-end network computing and packet processing applications, and offers the capacity to manage millions of subscribers in policy enforcement installations requiring multiple 10 GbE and 40 GbE ports. By using commercialoff-the-shelf ATCA blades with enhanced I/O in systems with the switching capabilities of ATCA, equipment providers can benefit from a wider choice of suppliers on a futureproofed architecture. At the same time, customers can gain distinct product differentiation by adding their own hardware acceleration on an eRTM. Keep the Video Streaming Keeping up with the explosion in video traffic requires a new, dynamic approach to traffic management. Design flexibility is also important, as it gives equipment manufacturers and
NMCs for I/O,Offload & Acceleration on eRTM
eRTM
Fabric Mezzanine Modules (FMMs) for I/0, Fabric Connectivity, Offload & Acceleration on ATCA Blade
Standard Commercial ATCA Blade
2-slot eATCA Chassis
Figure 6.
The Intel Platform for Communications Infrastructure is an excellent match for these requirements, providing a flexible, high-performance solution for policy enforcement. The software support from the Intel DPDK and Qosmos ixEngine enable rapid development and workload consolidation. Together, the hardware and software provide an excellent platform for the PCEF functions, giving TEMs valuable new options for equipment design and service providers important new options for service expansion.
An eATCA system combines standard ATCA blades with eRTMs that offer four times the real estate of an RTM for Network Mezzanine Cards (NMCs) and other I/O.
For more on the Advantech FWA-6510, see intel.com/SD-fwa-6510; for more on Advantech eATCA platforms, see intel.com/SD-advantech-atca
operators the capability to respond to new network conditions as they evolve. The ability to scale designs to meet new deployment scenarios in both carrier and data center environments also helps mitigate risk. Thus, platforms designed for PCEF must be: 1. Built on off-the-shelf hardware for minimum design time and maximum flexibility 2. Highly scalable to allow operators to match network loading, adapt to Gigabit or Terabit throughput to meet future growth demands 3. Highly configurable to allow for offload and acceleration technologies to be deployed in response to increases in bandwidth demand or to address new paradigms at lower cost, avoiding forklift equipment replacement
For more on flexible, scalable, standards-based communications visit intel.com/embedded-comm
Advantech (intel.com/ea-advantech) is a Premier member of the Intel® Intelligent Systems Alliance and provides business-critical hardware to the world’s leading telecom and networking equipment manufacturers. Our standard and customized products based on Intel architecture are embedded in OEM equipment that the world’s network infrastructure depends on.
Contact Advantech
Instant visibility over 1,300 protocols, 6,000+ application metadata
Qosmos embedded dPI and metadata engIne for complete traffic visibility
Ready-to-use C libraries with continuously updated protocol signatures Optimized on Intel multicore processor architecture
Control Plane
Data Plane Application
Packets
Control Plane / mgt
ixEngine DPI Engine OS & Packet Processing (Linux / Intel DPDK / WIND RIVER / 6WIND)
Core 1
Core 2
Core 3
…
Core n
Core
Supports Intel® Data Plane Development Kit (Intel® DPDK) and n X 10 Gbps throughputs Qosmos Inc. 440 N Wolfe Rd, sunnyvale, CA 94085, UsA Tel: +1 (408) 329-4548 Email: us-sales@qosmos.com
www.qosmos.com
intel.com/embedded-innovator | Embedded Innovator | 8 th Edition | 2013 | 21
Software-Defined Networking
Software-Defined
Networking
and the
Telecom Cloud A Phased Approach to Implementation By Eric Gregory, Director, Platform Product Management and Chandresh Ruparel, Director, Product Marketing, Radisys
W
ith the impressive success The Challenge: Bring the Cloud to the Network of cloud computing in the Cloud technology offers numerous benefits including economies of scale, enterprise, telecom opercost-effectiveness, scalability, efficient hardware utilization, and CAPEX ators are looking to reap and OPEX reduction. These benefits are all key objectives for the telecom its benefits in their central industry, so the appeal of cloud technologies is clear. However, deploying offices and network functhe telecom cloud presents a different set of challenges due to the industions. However, the telecom industry presents try’s demanding requirements for availability, scalability, reliability, and many tough challenges for cloud technology, complex networking (see Figure 1). such as the prevalence of heterogeneous hardware architectures Enterprise Cloud Telecom Cloud and the need to maintain existing infrastructure. Less Strict 3NINES Reliability Requirements
Strict 5NINES Reliability Requirements
Some Latency
Low Latency
Homogeneous Transport (Ethernet)
Heterogeneous Transport (Optical, Ethernet, Wireless)
Single Control Protocol (OpenFlow)
Multiple Control Protocols (OpenFlow, SNMP)
Controlled Data Center Operating Environment
Regulatory Requirements (NEBS)
Smaller Number of Warehouse-Sized Data Centers
Larger Number of Smaller, Distributed Data Centers
Software-defined networking (SDN) and Network Functions Virtualization (NFV) provide operators with a solution. With their complementary abilities to abstract telecom applications from the underlying hardware, SDN and NFV enable operators to move their networks onto cloud-based platforms in a staged manner, minimizing integration risk while leveraging existing infrastructure. This article will describe how Intel® architecture (IA)-based blade technology enables operators to implement SDN and NFV. We will examine an edge router use case and show how a staged approach based on Radisys ATCA blades can be extended to most network elements, bringing the benefits of the cloud to a wide range of telecom applications.
Figure 1.
The enterprise cloud and telecom cloud have different requirements.
22 | 2013 | 8 th Edition | Embedded Innovator | intel.com/embedded-innovator
In the IT world, 3-nines reliability is generally considered good enough and some latency is tolerable because it doesn’t typically impact the overall user experience. The telco industry, on the other hand, requires strict 5-nines reliability, very low latency, and relatively complex telecom-cloud infrastructure. While the enterprise cloud has one transport technology, Ethernet, the telecom cloud must support optical and wireless as well as Ethernet. Similarly, the enterprise cloud has a single control protocol, while the telecom cloud must support multiple control protocols. Scalability is also an issue—the telecom cloud must support a large number of smaller, distributed data centers as opposed to an enterprise’s smaller number of warehouse-sized data centers. Finally, the telecom cloud must also meet NEBS regulatory requirements.
OpenFlow
Control Plane Control Plane Control Plane
Control Plane
Data Plane
Data Plane
Data Plane
Traditional: Tight Coupling of Control and Data Plane
Overlay: Existing Network Element Exposes an SDN Interface to Enable Centralized Control
Hybrid: New Network Element Starts Off with Separated In-Skin Distributed Control and Data Plane. Control Plane Moved Physically at a Future Date
Figure 2.
Separation of the control and data plane in a software-defined networking (SDN) is enabled by use of the OpenFlow communications protocol.
Perhaps most importantly, today’s network infrastructure uses different architectures for routers, RAN gear, backhaul equipment, broadband access gear, and other components, making it difficult to create cloud architectures that can scale across applications. In addition, operators want to leverage the CAPEX already made in their existing network infrastructure to make better use of what they have.
The Solution: SDN and NFV SDN and NFV, which separate and abstract telecom networking applications from their hardware, can address these challenges and open up new opportunities. SDN enables the separation of the control plane and the data plane, as shown in Figure 2, typically using OpenFlow as the transport protocol between the two planes. For existing equipment, this separation is achieved by exposing an SDN interface on existing equipment. New equipment can be deployed with the control and data planes already separated. Later, the control plane can be physically moved to a central location. SDN offers two key benefits. First, it enables centralized control of the entire network, transforming the way applications are deployed and managed. For example, new applications can be deployed without upgrading the data plane devices. Second, SDN enables control resources to be physically grouped together on centralized cloud resources. By deploying control plane applications on virtual machines Goal instead of dedicated hardware, SDN enables more flexible, efficient hardware utilization, lowering CAPEX and OPEX. Hardware SDN is not to be confused with NFV, which are complementary technologies that use cloud technology for different purposes (Figure 3). NFV virtualizes specific functions such as firewall, VPN, or intrusion detection, allowing these functions to run in virtual machines on high-volume-based servers. By decoupling applications
Strategy
Example Applications Figure 3.
from the underlying hardware, NFV makes it possible to power down hardware when network traffic is low, leading to considerable energy savings. NFV also makes it easier to scale hardware, as many different applications can use the same physical resources. Figure 4 (Page 24) shows the wireless functions that are well suited for SDN and NFV. Note that some functions, like policy control, are suitable for both approaches. While both SDN and NFV bring cloud computing to the network, we will primarily focus on SDN for the remainder of this article. Intel® Architecture-Based Blade Technology The ultimate goal of SDN and NFV is to simplify network deployments and upgrades by using homogenous platforms at a hardware level, and moving the complexity of networking and applications to software. IA-based blades are uniquely able to meet this goal thanks to their ability to address the control plane, data plane, and the media plane. Intel technology is particularly useful in SDN because it offers a broad range of control-plane application support. SDN
NFV
Centralize network control and programmability to enable faster innovation
Relocate network functions from dedicated appliances to generic servers to improve CAPEX and OPEX
Leverage existing infrastructure
Introduce new hardware
Evolved packet core, policy control
Routers, firewalls, gateways, VPN, intrusion detection
Software-defined networking (SDN) and Network Functions Virtualization (NFV) serve different purposes for telecom equipment manufacturers and operators respectively and must be carefully considered. intel.com/embedded-innovator | Embedded Innovator | 8 th Edition | 2013 | 23
Real people. Real topics. Real answers. SOLVE PROBLEMS
GET ADVICE
The IntelÂŽ Embedded Community is a place where you can discuss embedded design solutions with the people who know. Share and solve problems, get the advice you need, and exchange design ideas with others in your field. And thatâ&#x20AC;&#x2122;s just the beginning. intel.com/embedded/community
SHARE IDEAS
Software-Defined Networking
The benefits of IA are demonstrated in the Intel® Platform for Communications Infrastructure. This platform combines the Intel® Xeon® processor E5-2600 v2 product family with the Intel® Communications Chipset 89xx Series to deliver features including:
Home eNodeB
Mobility Management Entity
Policy & Charging Routing Function
IMS App Server
SDN
• Up to 20 cores in a dual-socket eNodeB Media Resource NFV LTE Security Serving Packet Policy & Charging configuration for up to 160 Gbps of Function Gateway Gateway Gateway Enforcement Function Radio packet processing performance Access IP Multimedia • Four DDR3-1866 memory controllers, Network Subsystem Policy Control Evolved Packet Core two 8 GTps Intel® QuickPath Interconnect (Intel® QPI) interfaces, Applications in an LTE network suitable for software-defined networking (SDN) and Figure 4. and 25 MB L3 cache for significantly Network Functions Virtualization (NFV). higher performance than previous generations • Up to 40 PCI Express 3.0 lanes supporting I/O bandwidth member companies, the Alliance offers solutions ranging from opersignificantly in excess of 100 Gbps. ating systems to market-ready communications platforms, addressing • Support for dual 40 GbE network interfaces and additional the full range spectrum of networking and communications applicaPCI Express lanes to support network interfaces on the front tions. Close collaboration with Intel and each other enables Alliance panel, and storage or further interface options on the rear members to innovate with the latest technologies, helping developers transition module (RTM) deliver first-in-market communications infrastructure and reduce cost • Intel® Advanced Vector Extensions (Intel® AVX) for improved (Radisys is an Associate member of the Alliance). floating point and Linux RAID performance Edge Router Use Case The Intel Communications Chipset 89xx Series also provides hardware To minimize risk and leverage existing infrastructure, operators acceleration of cryptographic and compression functions by incorposhould consider a phased approach to implementing SDN, and an rating Intel® QuickAssist Technology. These accelerators are a flexible edge router is an excellent place to start. In the example shown in and cost-effective alternative to the use of a separate network processor. Figure 5 (Page 26), the first step is to absorb the edge router function into the wireless gateway. This enables telecom equipment manuThe platform’s hardware features are supported by the Intel ® Data facturers to offer a more complete solution and allows operators to Plane Development Kit (Intel® DPDK). The Intel DPDK greatly speeds focus on building a transport network. up the packet pipeline with features like buffer management, queue management, flow classification, and NIC poll mode drivers The second (and parallel) step is to physically separate the control and that work without asynchronous, interrupt-based signaling. These data planes within this same hardware unit, and centralize all conenhancements substantially improve performance on NFV-specific trol functions for that platform on an IA-based blade. Since transport initiatives such as Open vSwitch. efficiencies are difficult to duplicate in a network environment, this approach allows operators to logically separate control and data while Intel offers a variety of other platforms that are valuable for SDN and NFV. keeping them on the same platform. Virtualization is an important For example, Radisys and Intel recently collaborated on a media plane part of this separation. The control blade can employ separate virtual demonstration using multiple Intel® Core™ i7-3612QE processors on a machines for the various control applications, and the data plane blade single ATCA blade. This blade delivered five to six times the HD (1080p) can support edge routing and load balancing functionality in different transcoding performance of popular digital signal processor options. virtual machines. Overall, the applicability of IA across the control, data, and media planes eases application development in a larger ecosystem and makes the provisioning of compute resources under SDN and NFV architectures easier. Applications can be consolidated across homogenous platforms that offer resources on demand. These resources can address a broad range of applications spanning access and core infrastructure, such as eNodeBs and LTE-EPC, deep packet inspection (DPI) and policy management, IP-Multimedia Subsystems, and content delivery networks. The capabilities of IA are enhanced by the rich software offerings available from the Intel® Intelligent Systems Alliance. With its 250+ global
As the industry improves transport efficiencies, operators will be able to migrate the control plane out of the box and use the network as the transport layer. This is the third phase of SDN deployment—moving the control plane outside the platform to a virtualized environment. This phased approach is more practical in the short run since separating at the network level is more difficult. By first separating the functions within the same platform, the industry can make a smoother transition to the telecom cloud. This approach can then be extended to other network elements that contain a control plane and data plane, enabling a more cohesive management approach across the network and reducing operator OPEX.
intel.com/embedded-innovator | Embedded Innovator | 8 th Edition | 2013 | 25
Software-Defined Networking
Radisys T-Series Platform Addressing the need for efficient, commercial off-the-shelf (COTS) network elements that decrease delivery costs and improve performance, the Radisys T-Series platform offers a comprehensive answer to the challenges of moving to the telecom cloud (see Figure 6). It supports cloud deployments by delivering economies of scale, cost effectiveness, scalability, and lower CAPEX and OPEX. This telecom-grade system, built to NEBS specifications, is pre-integrated with a large selection of ATCA blades equipped with cutting-edge Intel® Xeon® processors to provide a robust, highperformance solution capable of 5-nines reliability. Highdensity and high-throughput I/O deliver the connectivity needed to keep up with increasing access speeds. Radisys’ ATCA solutions deliver the benefits of standards-based architecture, accelerating time-to-market and revenue with a pre-integrated platform. Designed to provide a migration path to next-generation silicon and I/O technology, the T-Series platform also provides the flexibility and scalability to meet the needs of today’s and tomorrow’s wireless networks. The platform is customizable to support specialized applications, including DPI, security, and Evolved Packet Core (EPC) functions. Integrated software eases development and provides a new R&D model—infrastructure as a service (IaaS)— enabling TEMs to focus solely on building their application and transform fixed R&D expenses into variable expenses.
Step 1: Absorb Edge Router into Wireless Gateway
Edge Router Wireless Gateway Step 2: Split Control and Data Planes Control Plane
Data Plane Edge Routing and Load Balancing Step 3: Break Out Control Plane to SDN Application Layer Open Flow
Data Plane Edge Routing and Load Balancing
Figure 5.
Starting Down the Path to SDN Using the latest Intel Xeon processors, Radisys ATCA solutions give operators the agility to quickly and cost-effectively begin to implement SDN by separating the data planes and control planes within the same platform. This implementation approach lowers risk and enables better distribution of resources by incorporating switching and load balancing functionality into the system. Taking this step sets the stage for a seamless transition of the data and control planes into separate platforms when desired.
The three steps in implementing software-defined networking (SDN) through a staged approach that separates the data and control planes in an edge router application.
However, transitioning to a pure SDN architecture will take time when considering the life cycles and investment in legacy equipment in the field. The Radisys approach enables operators to start with small, cost-effective steps on their way to a complete and successful transition to telecom cloud technology. Operators can use this approach to fulfill their current technology needs, while paving the way for next-generation developments in their move to the telecom cloud.
Telecom Cloud Control Plane • Telecom Grade Platform • 10th Gen Intel® Architecture Blades
Telecom Cloud Data Plane • Packet Processing • Load Balancing/DPI Software
SDN Controller • Implemented Within the Same Physical Platform as a Control Plane App
SDN Forwarding Plane • Dual NPU Blade • Edge Router Software
Figure 6.
Radisys T40
Control Plane
Radisys T-Series provides a high-performing, cost effective platform for use as telecom cloud building blocks.
26 | 2013 | 8 th Edition | Embedded Innovator | intel.com/embedded-innovator
For more information on the Radisys T-Series ATCA platform, see intel.com/SD-t40
For more on flexible, scalable, standards-based communications visit intel.com/embedded-comm
Radisys (intel.com/ ea-radisys) is an Associate member of the Intel® Intelligent Systems Alliance and a leading provider of wireless infrastructure solutions for telecom, aerospace, and defense. Radisys’ ATCA, Media Resource Function (MRF) and COM Express platforms and its Trillium software, services, and market expertise help customers bring solutions to market faster with lower investment and risk.
Contact Radisys
Signal & Image Processing
Faster Signal with Intel® Advanced Vector Simplify Coding with Vectorizing Compilers and Libraries By Noah Clemons, Technical Consulting Engineer, Embedded Computing & Debuggers, Intel® Corporation Peter Carlston, Platform Architect, Intelligent Systems Group, Intel® Corporation David Murray, Technical Director, N.A. Software
O
ver the past few years, Intel has significantly improved the vector-processing performance of its processors, making them increasingly popular targets for signal and image processing. The Intel® Advanced Vector Extensions (Intel® AVX) 2.0 introduced in the Haswell microarchitecture take these capabilities to a new level, delivering a 2x increase in peak floating-point throughput for an impressive 307 billion floating point operations per second (GFLOPS) at 2.4 GHz in a quad-core 4th generation Intel® Core™ processor. Fixed-point arithmetic also sees a 2x boost in peak throughput, and both fixedand floating-point algorithms benefit from new vector gather, scatter, and permute operations. This article discusses the value of implementing signal and image processing on the Haswell microarchitecture and highlights applications that can benefit from Intel AVX 2.0. We quantify the improvements developers can expect through performance tests provided by N.A. Software, an Affiliate member of the Intel® Intelligent Systems Alliance, and show how libraries from Alliance members can help them achieve these gains. In addition, we explain how to use the tools in Intel® System Studio to code for Intel AVX 2.0 and shorten development time. SIMD Enhancements in Intel® Architecture (IA) Signal and image processing software written for Intel® architecture (IA) processors can use single instruction multiple data (SIMD) instructions to process data in parallel—a technique known as vector processing. With this technique, multiple data values are loaded into SIMD registers to perform operations on all data elements at once (see Figure 1).
Performing signal and image processing solely on an IA processor— as opposed to requiring a companion digital signal processor (DSP)—has numerous benefits. For example, eliminating the DSP enables developers to: • Run their solutions on a single processor, shrinking system size and bill-of-material costs • Reduce programming efforts through a single code base and best-in-class tools • Minimize the risk of data-stream timing issues and other runtime problems associated with the communications between the IA processor and DSP Starting with the advent of Intel AVX in 2011, Intel has significantly improved vector-processing performance in each generation of its processors, enabling a growing range of signal and image processing applications to migrate to IA (Figure 2). This year, Intel AVX 2.0 takes vector performance to a[3] a[2] a[1] a[0] a new level with improvements + + + + that include: b[3] b[2] b[1] b[0]
a + b
a[7] + b[7]
a[6] a[5] + + b[6] b[5]
a[4] + b[4]
c
c[7]
c[6]
c[4] c[3]
Figure 1.
c[5]
c[2]
c[1]
c[0]
SIMD vectorization loads multiple data values into SIMD registers to perform operations on all data elements at once.
28 | 2013 | 8 th Edition | Embedded Innovator | intel.com/embedded-innovator
• F used multiply-add (FMA) instructions for 2x higher peak throughput—up to 307 GFLOPS at 2.4 GHz in a quad-core 4th generation Intel Core processor
Processing Extensions 2.0
• Extension of most integer instructions to 256 bits for 2x higher peak integer throughput • New vector gather, shift, and cross-lane permute functions that enable more vectorization and more efficient loads and stores Many improvements have been made in the Haswell microarchitecture to allow applications to realize this greater performance potential: • The memory pipeline can now perform 2 loads and a store on each cycle • L1 cache bandwidth has doubled to 96 bytes/cycle (64 byte read plus 32 byte write) • L2 cache bandwidth has also doubled to 64 bytes/cycle These upgrades, along with the internal Last Level Cache, 320 GBps Ring Bus, and DDR3 dual-channel memory (with a peak memory bandwidth of 25 GBps at 1,600 MHz) help keep the processor fed for maximum performance. The applications likely to benefit from Intel AVX2 include those that are CPU-bound, and also those that spend a significant time in vectorizable loops with:
Performance Benchmarks N.A. Software (NAS) develops and licenses advanced radar algorithms and low-level DSP libraries including the Vector, Signal, and Image Processing Library (VSIPL). This open application programming interface (API) provides portable computational middleware for image and signal processing functions as defined by the VSIPL Forum. VSIPL supports
As noted above, NAS is an Affiliate member of the Intel Intelligent Systems Alliance. VSIPL implementations are available from other Alliance members including Associate member GE Intelligent Platforms and General member Curtiss-Wright* Controls Defense Solutions. These companies are just some of the 250+ global members of the Alliance. From modular components to marketready systems, Intel and members of the Alliance collaborate closely to provide the performance, connectivity, manageability, and security developers need to create smart, connected systems.
Future Extensions Intel Microarchitecture Code Name Sandy Bridge (32nm Tock)a ®
Performance/Core
• Iteration count ≥ vector width (i.e., ≥ 8 integers, 8 floats, or 4 doubles) • Integer arithmetic and bit manipulation (e.g., video processing) • Floating-point operations that can make use of FMAs (e.g., linear algebra) • Non-contiguous memory access (i.e., those that can use the new gather and permute instructions)
multithreading and is typically used on large multi-core and shared memory systems, providing scalable performance for large problems. NAS has produced a highly optimized Intel AVX 2.0 VSIPL library that is especially well optimized for complex vector multiply operations, sine/cosine (when the data is not range reduced), and split complex FFTs. The NAS library is standalone code that does not rely on any third party software, enabling the library to be recompiled for any operating system quickly and easily to gain the most out of the Intel AVX 2.0 instruction set.
Haswell Microarchitecture (22 nm Tock)
Die Shrink Code Name Ivy Bridge (22 nm Tick) Sandy Bridge Microarchitecture (32 nm Tock)
Intel® AVX 2.0: Fused Multiply-Add (2x Peak FLOPS) 256-bit Integer Vectors (2x Peak Throughput) Gather/Shift/Permute
Intel® AVX (Float 16): Half-Float Support Random Numbers
Intel® Advanced Vector Extensions (Intel® AVX): 256-bit Floating-Point Vectors (2x Peak Floating Point Operations Per Second (FLOPS))
Since 1999: 128-bit Vectors
1999
Figure 2.
…
2011
2012
2013
…
Intel has significantly upgraded vector-processing performance in recent processor generations. intel.com/embedded-innovator | Embedded Innovator | 8 th Edition | 2013 | 29
Intel® System Studio
Accelerate time to market. New tools for embedded and mobile system developers. Intel® System Studio provides deep system-level insights into power, performance, and reliability. > Speed development and testing > Enhance code stability > Boost power efficiency and performance
LEARN MORE AT: http://intel.ly/system-studio
©2013, Intel Corporation. All rights reserved. Intel, the Intel logo, and Xeon Phi are trademarks of Intel Corporation in the U.S. and/or other countries. *Other names and brands may be claimed as the property of others.
Signal & Image Processing
NAS recently used VSIPL to benchmark Intel AVX 2.0. As shown in Figure 3, Intel AVX 2.0 produces speedups reaching over 2x depending on the function. Note that NAS also benchmarked the processors with the Intel® Math Kernel Library (Intel® MKL). NAS found that Intel MKL was much better at 2D FFTs with non-square matrix data, but the NAS library was more optimized for other operations and lengths. Figure 4 provides more detail on the 1D FFT routine using split complex data.
VSIPL Function
Parameters [columns*rows]
Intel® AVX 2.0 vs. Intel® AVX
1D FFT
256 – 512K
1.45 – 1.77 X
Multiple 1D FFT
256*256 – 2K*2K
1.43 – 1.6 X
2D FFT, non-square matrices
256*256 – 128K*20
1.14 - 1.22 X
2D FFT, smaller square matrices
64*64 – 2K*2K
1.39 – 1.61 X
Complex Matrix Transpose
256*256 – 2K*2K
1.09 - 1.48 X
NAS also benchmarked Intel AVX 2.0 with their Synthetic Aperture Radar and Moving Target Indication (SARMTI) advanced radar processing algorithm. SARMTI is able to extract high resolution data with the positions of all slow- and fast-moving objects directly from the SAR image itself so a separate MTI radar is not needed. Four different sets of image sizes and numbers/locations of moving objects were studied. The results showed Intel AVX 2.0 performed faster than Intel AVX across all scenarios, with speedups of 1.26x to 1.52x.i
Vector Multiply
1.28 – 1.57 X
Vector Sine
1.90 – 2.32 X
Vector Cosine Vector Square Root
Figure 3.
• Intel® C++ Compiler 13.0 for on-the-fly generation of Intel AVX 2.0 code • Hand-tuned vectorized and threaded implementations of commonly used embedded signal/image and math-intensive processing functions in found in two performance libraries: Intel® Integrated Performance Primitives (Intel® IPP) and Intel® Math Kernel Library (Intel® MKL) The auto-vectorization capabilities of the Intel C++ Compiler provide an excellent way to generate Intel AVX2 code. This compiler and its libraries can:
1.90 – 2.05 X 1.21 – 1.29 X
Vector Scatter
1.24 – 1.34 X
Vector Gather
1.21 – 1.56 X
N.A. Software benchmarked different versions of Intel® Advanced Vector Extensions 2.0 to show the performance gain.i
NAS vsip_ccfftip_f with Split Complex Data MFLOPS = 5 N Log2(N)/(Time for One FFT in Microseconds)
45 40 35 30 GFLOPS Rate
Coding for Intel® Advanced Vector Extensions SIMD code has a well-earned reputation for being timeconsuming to write and rarely portable. To overcome these obstacles, Intel and members of the Alliance have created compilers and libraries that streamline programming and enable straightforward porting across IA processors. In addition to the VSIPL libraries mentioned earlier, Intel offers its Intel® System Studio, which incorporates:
256 – 128K
25 20 15 10 5 0
256
1K Intel® AVX 2.0
Figure 4.
4K 16K Data Length Intel® AVX (Float 16)
256K
512K Intel® AVX
Intel® Advanced Vector Extensions 2.0 shows performance gains across a range of vector lengths.i Intel System Studio provides access to the Intel C++ Compiler and other tools through industry common integrated development environments (IDEs) such as the Eclipse* IDE. These tools include a guided autoparallelism selection tool and a unique environment file editor that enables integration of the Intel C++ Compiler into a GNU cross-build toolchain from within the IDE.
• Generate faster code through speed optimizations • Enable shorter execution times for low-power code • Support GNU* cross-build, integration into Eclipse* CDT, and Yocto Project* Application Development Toolkit The Intel C++ Compiler supports cross-compilation and integration with Poky-Linux* or OpenEmbedded* compatible GNU* cross-toolchains as used for Wind River Linux*, Yocto Project* and many other custom GNU* cross-build toolchain (Wind River is an Associate member of the Alliance). It comes with predefined compiler environment files that make cross-development a simple matter of applying a compiler switch.
Vectorization works the same as for past Intel® Streaming SIMD Extensions (Intel® SSE) and Intel AVX instruction sets. To include older code paths, developers can set a compiler flag. For example, the developer can set flags to generate code for both Intel AVX 2.0 and Intel AVX, and have the code target the appropriate version automatically at runtime.
intel.com/embedded-innovator | Embedded Innovator | 8 th Edition | 2013 | 31
Signal & Image Processing
Performance Libraries within Intel® System Studio Two libraries in Intel System Studio enable highly optimized data and signal processing on IA: • Intel IPP is an extensive library of highly optimized software building blocks for the most demanding signal, data, and multimedia applications • Intel MKL provides highly optimized threaded math routines The libraries are designed for portability and can run across different versions of the IA architecture. They detect the host processor at runtime and deploy the correct optimized code. Particularly advantageous, these hand tunings will not require a redesign of critical functions with successive hardware platform upgrades. Intel IPP offers an extensive C library of highly optimized building blocks for a wide variety of domains and data types (see Figure 5). Intel IPP is supplied as a sequential library for higher efficiency for smaller data sets, latencyconstraint applications, or better control via applicationlevel threading. Intel also makes dynamic link libraries available that provide internal multithreading; to support multithreading, Intel IPP is fully thread-safe. C language call-convention and linkage (undecorated functions names) allow for calling Intel IPP functions from almost all programming languages and compilers.
Signal Processing (1D) • Transforms (e.g., Wavelet) • Convolution/Correlation • Filtering (e.g., IIR, FIR) • Statistics
Image Processing (2D) • Transforms (e.g., Rotation) • (Non-)lin. Filter (e.g., Noise) • FFT, DFT, DCT • Statistics
Color Conversion • Color Space Conversion • Pattern (e.g., Bayer) • Brightness/Contrast • Resampling
Vector/Matrix • Logical, Shift, Conversion • Trigonometric Functions • Decomposition, Eigenvalues • Transpose
Integrity /Compression/ Cryptography • Error Correction, Reed-Solomon • Compression (Entropy, Dict.) • MD5, T(DES), RSA, DSA • Random Number Generators
• • •More Domains • Video, Picture Coding • Audio (e.g., Speech Coding) • String Processing • Utilities
Figure 5.
Intel® Integrated Performance Primitives cover many domains.
Linear Algebra • BLAS, Sparse BLAS • LAPACK Solvers • Sparse Solvers (DSS, PARADISO) • Iterative Solver (RCI) • ScaLAPACK, PBLAS Random Number Generators • Congruential • Wichmann-Hill • Mersenne Twister • Sobol • Neiderreiter • Non-deterministic
Figure 6.
In addition to the libraries, Intel IPP provides supporting code such as an example driver, along with an application requesting this driver’s service. Such an example may help in developing code that runs in kernel mode (ring 0). To support such code, Intel IPP provides libraries that are not position-independent (“nonpic”). For floating point data types, Intel MKL is available as a single- or multi-threaded implementation. Linear algebra is one of the primary domains of Intel MKL (see Figure 6), and these functions see a particularly large benefit from the fused multiply add introduced in Intel AVX 2.0. The Fast Fourier Transforms (FFTs) in Intel MKL support sophisticated descriptors, giving forward/backward functions a slightly simpler interface compared to Intel IPP. Along with threading, Intel MKL includes sophisticated optimizations. Massive workloads up to the 64-bit integer index space are supported (“ilp64”). Intel MKL focuses on high throughput; however, applicationlevel threading can be served by sequential implementations similar to Intel IPP. Note that calling multi-threaded Intel MKL functions from multiple threads is thread-safe. Take Advanced Signal and Image Processing to the Next Level As the results cited from NAS show, Intel AVX 2.0 delivers excellent vector-processing performance gains. To realize these gains, developers can generate Intel AVX 2.0 code on-the-fly with the Intel C++ Compiler and can call on hand-tuned libraries using Intel IPP, Intel MKL, or VSIPL.
Fast Fourier Transforms • Multidimensional • FFTW Interfaces • Cluster FFT • Trig. Transforms • Poisson Solver • Convolution via VSL Summary Statistics • Kurtosis • Variation Coefficient • Quantiles • Order Statistics • Min/Max • Variance-covariance
Vector Math • Trigonometric • Hyperbolic • Exponential, Logarithmic • Power/Root
Data Fitting • Spline-based • Interpolation • Cell Search
Intel® Math Kernel Library domains.
As developers of signal and image processing equipment look for ways to eliminate DSPs and run their entire solution on the latest IA processors, these tools and libraries will continue to grow in importance—but fortunately not in complexity. These best-in-class tools streamline SIMD code development, making one-processor solutions that simplify programming with a single code base all the more attractive. For more on Intel System Studio, see intel.com/SD-systemstudio; for more on NA Software VISPL, see intel.com/SD-svsipl
To learn more about advanced signal processing, see intel.com/embedded-analytics
From modular components to marketready systems, Intel and the 250+ global member companies of the Intel® Intelligent Systems Alliance (intel. com/intelligentsystems-alliance) provide the connectivity, manageability, security, and performance developers need to create smart, connected systems. GE Intelligent Platforms (intel.com/ea-ge) and Wind River (intel.com/ea-windriver) are Associate members of the Alliance, N.A. Software Ltd. (intel.com/ea-nasoftware) is an Affiliate member, and Curtiss-Wright Controls Defense Solutions (intel.com/ea-cwcds) is a General member.
Contact Intel
32 | 2013 | 8 th Edition | Embedded Innovator | intel.com/embedded-innovator
Smart Grid
Power up Security for the
Smart Grid
Retrofits and Rebuilds in a Two-Prong Approach By Alexander Damisch, Director of Industrial Solutions, Wind River
R
ecent reports of cyber warfare have highlighted the vulnerability of the power grid and prompted calls for action. Yet the risks are set to grow exponentially with the rollout of smart grids, which will incorporate Internet of Things technology to transform power grids into cloud-based networks. These distributed systems will create millions of new targets for attacks, and vastly greater dependence on data integrity. To avoid disastrous outcomes, security must be built into the smart grid from the ground up.
In this article, we will show how developers can meet this goal with an incremental process that starts by retrofitting existing infrastructure while creating a path to a full rollout of the smart grid. We will illustrate how Wind River software and Intel hardware can work in concert to meet the critical challenges smoothly and efficiently, allowing utilities to incorporate smart grid technology with minimal disruption.
Challenges For designers responsible for laying the foundation of the smart grid, the security issue requires a holistic view and comprehensive approach encompassing the hardware, operating system, and software requirements. They must also ensure that the software solutions are upgradeable because the nature of threats will evolve over time in ways that cannot be anticipated. The biggest challenge is the need to maintain the function of the installed base while upgrading the system. In many regions, the infrastructure is badly dated. For example, 70 percent of the U.S. infrastructure is more than 30 years old. This legacy infrastructure is typically designed for fairly static distribution patterns based on predictable thermal power plants. Transforming the system into a dynamic network that can react to
Increased Complexity, Increased Vulnerability In the smart grid, operators will expect to have complete transparency and visibility to monitor, analyze, and control energy systems. They will need to know where and how energy is being created and consumed. They will need to communicate with and control the various systems deployed in the cloud-based network in real time to ensure efficiency in the flow. Achieving this level of automation, Traditional: Service Provider Driven Service Driven Device Driven Emerging: Telco Cloud Driven communication, and connectivity will require unprecedented integration and sophistication (Figure 1). Operators will Smart Service be challenged to minimize complexity Connected Capabilities while managing the cost of development, Application & Devices All-IP Core M2M implementation, and maintenance. Services An equally demanding and arguably more critical challenge, however, will be reducing security vulnerability. Highly interconnected systems and cloud-based networks create more points of potential intrusionâ&#x20AC;&#x201D;with potentially disastrous results. Hacking and malware could wreak havoc on a scale ranging from disconnecting individual meters to immobilizing entire cities or regions.
Client Application
Operations Data Center Figure 1.
Network
Network Cloud Domain
Area Network
Internet of Things Domain
The smart grid will employ Internet of Things technology.
intel.com/embedded-innovator | Embedded Innovator | 8 th Edition | 2013 | 33
Smart Grid
fluctuating demand and intermittent energy resources like renewable energy is a major leap, to say the least. While these problems have been known for some time, the solutions have been slow to arrive, and the proposed solutions in many cases are insufficient. For example, simply distributing smart meters is not enough to make a grid intelligent. Connecting everything without proper security is extremely dangerous—possibly even creating a national security risk. On the other hand, building a complete defense network is cost prohibitive and potentially inflexible. What’s needed is a way to protect key points in the network, such as security perimeters, flexibly and efficiently. The solution is to upgrade the installed base with an evolutionary approach that involves a two-pronged approach. First, we retrofit secure, managed communications into the installed base to enable new services. This “brown field” or “bolt on” approach can use existing mainstream technology and carries the least risk and cost. It also has the potential to reduce operational costs (OPEX). At the same time, we can build in the flexibility to accommodate future versions of the smart grid, enabling a smooth transition. As an example of this approach, legacy devices can be connected to the network through gateways that create a secure perimeter. The bolt on strategy is only part of the solution, and may even add to the capital expenditure (CAPEX) of transforming the system, as new infrastructure has to be added. Therefore, we must also pursue a “green field” redesign of the system. Once core parts of the infrastructure are redesigned, all new devices can migrate from function-oriented to service-oriented architectures for greater flexibility and efficiency. For example, gateway, protection, management, and communication functions can be consolidated into a single device to lower the CAPEX and OPEX of the overall system. This process takes longer and requires bigger change, so it is useful to deploy brown-field retrofits while we pursue the longer range green-field vision.
Module (TPM) technology available on Intel® architecture-based platforms. This technology uses cryptographic keys to ensure only the vendor authorized that the firmware, operating system (OS), and applications are launched. This prevents low-level attacks known as “jail breaking” or “rooting” in the smart phone world. • Resource control protects the system at runtime by preventing processes and applications from accessing memory, processing time and file resources they have not been provisioned for. Again, well established silicon extensions help to restrict access to resources at runtime. • Secure package management enables remote updates and management. This is a critical feature for reacting to zero day exploits, as it enables timely reaction to new incidents. TPM can also be used to securely store keys used in this process. Of course, it is not enough to secure the endpoints. A secure infrastructure is also needed, especially if one wants to enable a move from provider-based to cloud-based infrastructure. Cloud-based networks can greatly improve system flexibility while lowering CAPEX and OPEX, but they also introduce a new set of security risks. To deal with these risks, Wind River created the Intelligent Network Platform (INP), a software platform for sophisticated network equipment that can analyze, prioritize, and secure the flood of traffic for current and future networks (Figure 3). It can scale from a single gateway deep in the energy network up to the network management level. Specifically, Wind River INP includes the following capabilities for a secure infrastructure: • The Wind River Content Inspection Engine analyzes data streams at line speed, enabling low-latency software pattern matching to scan for security attacks and anomalies in protocols, and block them from entering critical domains. For example, this can be used to block a denial of service against end devices that lack the capability to handle such threats. It can also block a stop command from an external, unauthorized domain to avoid shut down of a critical device.
Workbench Tools, Services, and Support
Brown-Field Retrofit As noted above, gateways are a key method for connecting existing edge equipment, and for creating a secure perimeter around legacy devices. Wind River Intelligent Device Platform 1.0 These gateways require two key attributes: the ability to connect and translate Core (M2M Features) ISVs Secure Remote Management between different domains, and the secuWeb Interface Secure Package OMA DM OSGi rity to prevent intrusions. The Wind River Management Management Integration Intelligent Device Platform (IDP) was creTR-069 JVM ated with both goals in mind (Figure 2). It Network Profiles Resource Control is a complete software development envi(grsecurity) ronment with pre-integrated components Connectivity that address the connectivity, security, 2G/3G/4G Wi-Fi AP Ethernet Bluetooth Zigbee Integrity and manageability required by gateways Measurement and other IoT devices. Wind River Linux 4.3 Trusted A key feature of Wind River IDP is Secure Boot BSP: Intel Atom Remote Management (SRM), which solves three core challenges: • It provides a trusted boot mechanism by leveraging the Trusted Platform
Figure 2.
The Wind River Intelligent Device Platform provides a complete package for secure inter-domain connection.
34 | 2013 | 8 th Edition | Embedded Innovator | intel.com/embedded-innovator
Customer Customer Customer Data Plane App Customer Data Plane App Customer Data Plane App Data Plane App Data Plane App
Native Linux Apps Global Support Migration & Design Services
• The Wind River Application Acceleration Engine optimizes the speed of layer 3 and 4 protocols, which is instrumental to handling the skyrocketing number of connected devices. Leveraging the Intel® Data Plane Development Kit (Intel® DPDK), this engine ensures the highest levels of performance. When deployed on the Intel® Platform for Communications Infrastructure, for example, the engine can achieve up to 200 Gbps throughput. • The platform is optimized for use with Wind River Carrier Grade Linux, which helps to build a secure platform with 99.999 percent uptime. With a Yocto project foundation, the platform is based on a rich ecosystem-enabled Linux stream for longterm stability.
Management Plane
Flow Application Content Future Analysis Acceleration Inspection Engine(s) Linux* Engine Engine Engine w/Intel® DPDK Core Affinity Core Affinity Core Affinity Core Affinity Linux Kernel Space Core
Core
Core
Simics Workbench Development Lifecycle Tools
Linux User Space Core
Core
Core
Core
Core
Core
Core
Intel® Platform for Communications Infrastructure
Figure 3.
The Wind River Intelligent Network Platform simplifies communications infrastructure security.
Green-Field Redesign Wind River’s INP and IDP are also both useful for green field redesigns of devices and systems. IDP in particular can scale down into the low-footprint real time space, enabling new devices that are ready for IoT out of the box, with all credentials on board. When redesigning it is key to make sure that even the endpoint device is a self-defending secure system, rather than expecting that it will always
be within a protected perimeter. Therefore, even the RTOS is required to provide the key aspects of security such as hardware root of trust, runtime protection and hardening, remote attestation, and the ability to update for security flaws. Developers can meet these goals with products like Wind River VxWorks CERT, a commercial off-the-shelf (COTS) platform for delivering safety critical applications, part of the Wind River portfolio of trusted systems (Figure 4 on page 37).
Hardware-Assisted Firmware Debugger Just the facts about Arium: 1. Intel support since 1992 2. First UEFI-aware debug features 3. First Linux OS-aware debug features 4. Easy deployment (target auto discovery) 5. Code execution trace 6. All processor modes fully supported 7. Legacy processor support standard 8. Full multi-thread/processor/device control 9 Architectural Event Trace (AET) available 10 Arium is the most widely used JTAG-based IA debugger in the world! For more information: www.arium.com 714-731-1661 intel.com/embedded-innovator | Embedded Innovator | 8 th Edition | 2013 | 35
the performer!
congatec presents its fastest COM Express module, the conga-TS87 ▪ 4th Generation Intel® Core™ processor-based platform ▪ 3x DisplayPort 1.2, up to 4k resolution (QFHD=3840x2160@60Hz) ▪ Intel® Advanced Vector Extension (Intel® AVX) 2.0 for improved floating point computing ▪ Up to 13% higher computing performance* ▪ Up to 32% higher 3D graphics performance* conga-TS87
* SPECint_base2006, 3DMARK Vantage
Find more details at:
www.congatec.us
congatec Inc. | 6262 Ferris Square San Diego | CA 92121 USA | Phone: 858-457-2600 | sales-us@congatec.com © 2013 congatec AG. All rights reserved. conga and congatec are registered trademarks of congatec AG. Intel, Intel Core are trademarks of Intel Corporation in the U.S. and other countries. COM Express is a registered trademark of PICMG.
Smart Grid
The Wind River product portfolio of trusted systems also includes Wind River Hypervisor, a virtualization platform that enables multiple OSs on the same hardware. Wind River Hypervisor takes advantage of Intel Virtualization Technology (Intel® VT), a set of hardware accelerators that greatly improve the performance and security of virtualized solutions. Together these technologies can deliver important advantages in CAPEX, OPEX, and security:
Development Suite Wind River Workbench Wind River Compiler GNU Compiler Wind River Workbench On-Chip Debugging*
Workbench Debugger Host Shell VxWorks Simulator
• Consolidation—Instead of a 1:1 relationship between function and device, Software Partners different functions can now be combined Java OpenGL Graphics on one physical device. This helps reduce hardware cost without a software Certification Services Simulation Platforms redesign. For example, a communications stack built on Linux and a control software Design Tools Ada built on an RTOS can be combined on one platform without porting either stack or reworking either OS. This allows system Operating System and Middleware designers and architects to migrate their device-centric world into a more softwareIEC 61508 SIL3 DO-178B Level A defined, service-oriented architecture, Certification Evidence* Certification Evidence* without having to compromise on flexibility or performance. What’s more, this Safety-Certified Network Stack (TCP/UDP/IPv4)* functional consolidation is an excellent way to optimize the use of multi-core Intel Cert Highly Reliable File System* architecture processors. • Separation—Wind River Hypervisor and VxWorks Cert 6.6 Intel VT ensure that runtimes are kept separated in time and space. Different *Optional virtual machines cannot steal each other’s CPU cycles or access each other’s data Figure 4. Wind River VxWorks CERT is built to deliver safety-critical applications. unless they have been designed to do so. This strict separation is a key aspect to build safe and secure systems. • Certification—The trusted capability built into Intel VT can support does not have to. Wind River’s expertise and technologies already certification of systems with runtimes of different criticalities. available will enable energy providers to secure their systems today Because services of different criticalities are strictly separated, while providing the flexibility to adapt to as-yet unknown threats that designers can also keep the life cycle separated when it comes may be looming on the horizon. to updating or certifying individual runtimes. This provides tremendous risk control, as well as CAPEX and OPEX reductions. For more information on solutions from Wind River, see intel.com/SD-windriver-products Intelligently Planning an Intelligent Grid A more intelligent, more efficient energy distribution system is in the long-term best interests of the planet and everyone who lives on it. To learn more about connected, secure energy solutions, Leveraging advanced automation technologies will benefit energy see intel.com/embedded-energy producers, distributors, and consumers alike, potentially reducing the costs of delivery while increasing control at every level. The key to delivering the benefits successfully is designing security at every step. Wind River (intel.com/ea-windriver) is Contact Wind River By pursuing a gradual, phased deployment of secure systems, develan Associate member of the Intel® opers can foil those who would tamper with the system. Intelligent Systems Alliance and a global leader in device software optimization (DSO). Wind River platforms are pre-integrated, fully Those risks are nonetheless real, and it is incumbent upon standardized enterprise-wide development solutions. They reduce energy providers to protect the grid at every point of vulnerability. effort, cost, and risk, and optimize quality and reliability at all phases Performance cannot come at the price of security—and fortunately of development from concept to deployed product. intel.com/embedded-innovator | Embedded Innovator | 8 th Edition | 2013 | 37
Industrial
Transform the Factory
with the Internet of Things Performance, Graphics, and Security with 4th Generation Intel® Core™ Processor By Murray Slovick, Roving Reporter for Intel® Intelligent Systems Alliance
T
hroughout history, new technologies have transformed the manufacturing industry. From the invention of steam engines to the introduction of computerized controls, these technologies have led to enormous leaps in productivity and quality. Today we are at another turning point. The introduction of intelligent systems and Internet of Things technology are enabling unprecedented data sharing and analysis, turning previously disconnected manufacturing systems into an efficient, highly responsive whole.
This article will show how Intel is enabling the transformation with its 4th generation Intel® Core™ processor family based on the Haswell microarchitecture. Boasting up to 15 percent higher performance, 2x faster image and signal processing, up to 60 percent faster graphics, newly flexible high-speed I/O, and enhanced security, this platform offers major upgrades over its predecessors. As we will see, the new processors deliver scalable, robust performance for applications ranging from machine vision to machine diagnosis, from building controls to operator panels, and even top floor administrative servers (Figure 1). In addition, the platform’s new security and manageability features make it well-suited for the highly-connected, Internet of Things-enabled factories of the future. Along the way we’ll present examples where members of the Intel® Intelligent Systems Alliance are taking advantage of the new processors. From modular components to market-ready industrial systems, Intel and the 250+ global member companies of the Alliance provide the performance, connectivity, energy efficiency, manageability, and security developers need to
Figure 1.
create smart, connected systems. Close collaboration with Intel and each other enables Alliance members to innovate with the latest technologies, helping developers deliver first-in-market factory automation solutions to improve manufacturing efficiency.
The 4th generation Intel® Core™ processor family improves numerous operations on the factory floor.
38 | 2013 | 8 th Edition | Embedded Innovator | intel.com/embedded-innovator
Improved Performance Feature Benefits Performance is critical in Internet of Things-enabled 2x peak floating-point throughput for machine factories, as it enables greater analysis of product Fully pipelined Fused Multiply vision applications like automatic inspection, quality, equipment performance, and other factors. Add (FMA3) on two ports motion control, and other signal processing Overall performance is up 15 percent in the new processors, while signal and image processing get 2x peak fixed-point throughput 256-bit SIMD integer operations for video & image processing and an additional 2x boost with Intel® Advanced Vector general signal processing Extensions (Intel® AVX) 2.0. As shown in Figure 2, Intel AVX 2.0 introduces a fully pipelined Fused Gather, scatter, and permute operations Easier vectorization for many applications Multiply Add (FMA3) that provides twice the floating Compression, Encryption, and point performance of the previous generation for Bit manipulation instructions general-purpose code the multiply-add workloads common in image processing. Intel AVX 2.0 also extends integer Intel® Advanced Vector Extensions (Intel® AVX) 2.0 enhancements for vector processing from 128 bits to 256 bits, also Figure 2. image and signal processing. doubling throughput for many common workloads. For more details on this new instruction set, see “Faster Signal Processing with Intel® Advanced Vector Extensions 2.0” Serial ATA for an onboard 2.5” SATA hard disk or industrial grade NAND on page 28. Flash SSD device—all in a single-slot (4HP) computer. What’s more, the 4th generation Intel Core processors are more energy efficient than their predecessors, enabling smaller form factors for constrained spaces and easier heat dissipation. The high performance and low power also make it possible to consolidate multiple automation workloads (such as supervisory control and image processing) on a single computing device—a topic we will revisit shortly. Visual Displays To keep up with the increasing sophistication of factory equipment, human machine interfaces (HMIs) need to deliver sophisticated 2D and 3D graphics, video, and other media types that clearly communicate a machine’s status and intended operation. Advanced visuals are also important in central control rooms, where management needs to understand increasingly complex and distributed Internet of Thingsenabled systems at a glance. 4th generation Intel Core processors offer several features that improve its HMI capabilities, starting with an up to 60 percent improvement in 2D/3D graphics. The processors also expand their video decode accelerators, adding support for MJPEG and SVC and for 4K x 2K ultra-high-definition resolution. In addition to supporting up to three independent displays, the processors now offer a Collage Display mode, which presents multiple displays to the operating system (OS) as a single screen—a feature that is particularly useful for multi-screen HMIs. Together these features reduce the need for expensive and power hungry discrete graphics cards. Several Alliance members offer off-the-shelf solutions that help designers take advantage of the improved performance. For example, Premier member Kontron offers seven platforms based on the 4th generation Intel Core processors, such as the Kontron CP6005-SA shown in Figure 3. This 6U CompactPCI board offers high computing and graphics performance along with a complement of high bandwidth I/O that includes five Gigabit Ethernet (GbE) channels for external and internal PICMG* 2.16 compliant Ethernet traffic. The board also supports two additional 10 GbE ports combined with PCI Express* (PCIe) 3.0 x4 on the backplane, all based on PICMG* 2.20 for applications with high bandwidth requirements. The features are rounded out with USB and
Consolidation The performance upgrades will have a positive impact on another factory floor challenge: efficient hardware utilization. Industrial facilities are confronted by the proliferation of servers and controllers that each run a single OS and perform a single function, such as HMI on Linux* or motion control on a real-time operating system (RTOS). Virtualization makes it possible to consolidate these workloads onto a single physical platform, lowering overall solution cost and footprint. Virtualization has many other benefits. By reducing the number of physical systems, virtualization reduces the number and variety of parts and spares required for repairs. For that matter, virtualization can reduce downtime by enabling hot-swappable software spares that can smoothly take over if the main instance fails. To combine OSs, a virtual machine monitor (also known as a hypervisor) creates a virtual machine (or guest) for each OS. The key challenge for virtualization of industrial applications is ensuring that real-time guests maintain deterministic performance. Intel® Virtualization Technology (Intel® VT) is available on 4th generation
Figure 3.
Kontron’s CP6005-SA family uses 4th generation Intel® Core™ i7 processors.
intel.com/embedded-innovator | Embedded Innovator | 8 th Edition | 2013 | 39
Industrial
Intel Core processors and addresses this concern with a variety of hardware features that include: • Memory protection to prevent applications from interfering with one another • Hypervisor accelerators that ensure deterministic, low-latency responses for real-time guests • The ability to assign I/O to guest OSs for unimpeded, highperformance access
costly interruptions to factory operation. The 4th generation Intel Core processor adds a number of features to its security portfolio, including McAfee’s Deep Defender technology, which resides between the memory and OS to perform real-time memory and CPU monitoring without impacting overall system performance (McAfee is an Associate member of the Alliance). As shown in Figure 6, additional security elements include: • Intel® Platform Protection Technology with BIOS Guard provides authentication and protection against BIOS recovery attacks. BIOS updates are cryptographically verified to ensure malware stays out of the BIOS • Intel® Platform Trust Technology and Intel® Boot Guard are designed to work with Microsoft* Windows* 8. Available on the forthcoming U-Series processors, Intel Platform Trust Technology supports secure and measured boot, and Intel Boot Guard protects against boot block-level malware
Several Alliance members provide RTOS and hypervisor products that support the 4th generation Intel Core processors. For example, TenAsys* eVM* for Windows* is a real-time hypervisor that uses Intel VT to enable RTOSs and other guest OSs to run along with Microsoft* Windows*. TenAsys also offers the INtime* RTOS family, which can run as a standalone RTOS or alongside Microsoft Windows as shown in Figure 4. Both products enable users to partition a multi-core platform to run mixed workloads, making better use of the processor’s advanced features to provide highly integrated solutions TenAsys* INtime* TenAsys* INtime* Distributed RTOS (Microsoft is an Associate member for Windows* of the Alliance, and TenAsys is an Associate member). Robot 2 Robot 3 Robot 1 INtime* Ni INtime I/O Flexibility The chipsets for the 4th generation GOBSnet ti INtime Camera 3 INtime Intel Core processors introduce GOBSnet INtime Intel® Flex I/O for configurable I/O, GOBSnet INtime permitting designers to assign up to INtime four to six SATA 6.0 Gbps ports, six Windows wss* to eight PCI Express Gen 2.0 ports, and four to six USB 3.0 ports. Select processors also support PCIe 3.0 configurations including 1 x 16, Camera 1 2 x 8, or 1 x 8 + 2 x 4. Portwell takes Microsoft* Windows*-based HMI Camera 2 advantage of this rich I/O with its ROBO-8112VG2AR, a full-size PICMG Figure 4. TenAsys’ INtime effectively partitions a multicore platform to perform numerous tasks. 1.3 form factor System Host Board (SHB). As shown in Figure 5, the board supports the full range of PCIe 3.0 Three SATA III Ports Four USB 3.0 Ports outputs for flexible expansion, as well Dual 240-Pin DDR3 DIMM Sockets HDMI Port One RS232 Port as three SATA 6.0 Gbps ports and Four USB 2.0 Ports four USB 3.0 ports. Other I/O includes Parallel Port FDD Port dual GbE, two serial ports (one RS232 and one RS232/422/485 selectable), Dual USB 3.0 Ports one FDD port, PS/2 interface for keyboard/mouse and one parallel Dual GbE Ports port for traditional factory automation applications (Portwell is a Premier DVI-I Port member of the Alliance). (DVI-D+VGA) Security and Reliability A factory is only as strong as its weakest link, so every Internet of Things client in the factory needs protection from viruses, malware, and hacking to prevent
LGA1150 Socket +12 V Power Connector
Figure 5.
Intel® C226 Chipset
PS/2 KB and Mouse One RS232/422/485 Selectable Ports
The Portwell ROBO-8112VG2AR features multiple ports.
40 | 2013 | 8 th Edition | Embedded Innovator | intel.com/embedded-innovator
• Intel® Data Protection Technology with Intel® Advanced Encryption Standard New Instructions (Intel® AES-NI) enables rapid and secure data encryption and decryption Virtualization can also improve factory security and uptime because VMs can be segregated into different protection domains so a problem with one I/O device in one domain is isolated from affecting the other domains. To keep machines running in large, distributed manufacturing systems, operators must be able to remotely monitor and manage their equipment. Intel® Active Management Technology (Intel® AMT)
supports this requirement with out-of-band management that allows remote access to industrial computers even when the OS or firmware is damaged or the unit is powered down. Alliance members provide a number of hardware and software products that provide additional remote management tools. For example, Associate member NEXCOM’s ICES 670 COM Express Type 6 Basic module supports Xcare* 3.0, a utility that allows users to track parameters such as fan speed, temperature, voltage, watchdog timer, I/O, and disk status (Figure 7). These parameters can be fed back to a central operating room, alerting operators when an industrial device is in danger of failing so they can perform preventative maintenance.
McAfee* Supported on All Intel® Architecture Platforms
McAfee Embedded Control
Malware Prevention (White Listing)
Intel® Advanced Encryption Standard New Instructions (Intel® AES-NI)
McAfee Endpoint Encryption
Enhanced Encryption Speed
Intel® Virtualization Technology for Intel® 64 and IA-32 Architectures
McAfee Deep Defender v1.0
Greater Protection
McAfee ePO + Deep Command
Greater Device Manageability
Intel® Advanced Management Technology (Intel® AMT) Out-of-Band Management New
Additional Platform Security Enhancements
Figure 6.
Intel® Platform Protection Technology with BIOS Guard
Intel® Boot Guard
Intel® Platform Trust Technology
The Next Industrial Revolution The 4th generation Intel Core processor family will enable the transformation of the modern factory with Internet of Things technology—not only because of its higher performance but also because it permits industrial equipment manufacturers to significantly reduce design complexity and cost through consolidation of multiple applications and operations. All told, managers of industrial computing solutions will find that the new processor’s performance, signal processing capability, and I/O flexibility are true game changers, enhancing productivity as well as augmenting the security and manageability of industrial PCs, controllers, SCADA systems and HMI stations. For information about industrial solutions from the Alliance, see intel.com/SD-industrial
Intel® platforms support numerous McAfee protection elements.
Cloud Server Contact Intel
NEXCOM Xcare* Utility Remote Management Embedded API
Hardware Status
Remote KVM
System Restore
Remote Configuration
3rd Party Applications
NEXCOM ICES 670 COM Express Type 6 Basic Module with 4th Generation Intel® Core™ Processor
Figure 7.
The NEXCOM ICES 670 supports the NEXCOM Xcare* Utility.
From modular components to market-ready industrial systems, Intel and the 250+ global member companies of the Intel® Intelligent Systems Alliance (intel.com/intelligentsystems-alliance) provide the connectivity, manageability, security, and performance developers need to create smart, connected systems. Kontron (intel.com/ea-kontron) and Portwell (intel.com/ea-portwell) are Premier members of the Alliance; McAfee (intel.com/ea-mcafee), Microsoft (intel.com/ea-microsoft), and NEXCOM (intel.com/ea-nexcom) are Associate members; and TenAsys (intel.com/ ea-tenasys) is an Affiliate member.
intel.com/embedded-innovator | Embedded Innovator | 8 th Edition | 2013 | 41
Retail
Upgrade
Point-of-Sale for the Online
Age
Transactional Improvements with 4th Generation Intel® Core™ Processors By Kenton Williston, Editor in Chief
I
ntelligent point-of-sale (POS) systems with Internet connectivity are changing the retail landscape, enabling physical stores to catch up technologically with online retailers. This new generation of POS systems uses Internet of Things technology to help stores build customer relationships, better target promotions, and make informed business decisions with up-to-the-minute data.
At the Crossroads of the Connected Store Retailers have never been more dependent on technology. Transactional retail systems—including stationary and mobile POS systems as well as kiosks—are being used to complete sales, manage customer loyalty programs, provide personalized incentives, display targeted advertising, and provide real-time inventory information (see Figure 1). These activities increasingly rely on Internet of Things and Big Data to present shoppers with timely data and generate advanced business intelligence.
The 4th generation Intel® Core™ processor family is driving this The growing dependence on technology and the always-competitive transformation forward, delivering the performance, graphics, secunature of retailing makes it more important than ever to select solutions rity, and manageability developers need to create advanced POS systems. In this article, we show how these features fill Retail of Yesterday Enabling Technologies Retail of Today the needs of today’s connected store. Media & Graphics for Print Promotions Digital Promotions Rich Interactive Experiences We also examine example solutions ® from members of the Intel Intelligent Generic Advertising Targeted Advertising Compute Power for Systems Alliance. From modular comAnonymous Viewer Analytics ponents to market-ready retail systems, Loyalty Card Programs Personalized Incentives Intel and the 250+ global member comSecurity & Manageability for panies of the Intel® Intelligent Systems Truck Roll Support Remote Administration Protected, Reliable Devices Alliance provide the technology developers need to create smart, connected Checkout Lines At Register Checkout Choices Low Power For Thinner, Lighter, Sleeker Devices systems. Close collaboration with Intel and each other enables Alliance members to innovate with the latest technologies, helping developers deliver The use of intelligent devices in physical stores is transforming retail transactional Figure 1. first-in-market retail solutions to technology. increase sales and efficiency. 42 | 2013 | 8 th Edition | Embedded Innovator | intel.com/embedded-innovator
Scale up NAS Storage up to Unparallel Expandability for Future Business Growth REXP-1600U-RP
TS-EC1679U-RP
■ Powerful
unified storage for SMB and entry-level enterprise
■ Full-ranged
business-class models
The QNAP TS-EC1679U-RP, a powerful NAS unit with Intel® Xeon® CPU, 10GbE-ready network, and 4GB ECC memory, offers unified storage for both NAS and iSCSI applications. The flexible and economical REXP-1600U-RP/ REXP-1200U-RP* expansion enclosures seamlessly scale up the storage capacity up to 400 TB. ■ Proficient
business applications
Design with business in mind, the large storage capacity of TS-EC1679U-RP is ideal for server virtualization storage, client data backup and disaster recovery, file sharing, and IP surveillance storage. ■ Easy
TS-EC879U-RP
installation, secure, and reliable
With plug-and-use design and intuitive UI, managing and expanding storage capacity are made easy. ECC memory, dual power supply, and LAN fail-over ensure secure and reliable system operations.
*Expected to be released in Q2, 2013, subject to change.
Copyright © 2013 QNAP Systems, Inc. All rights reserved.
www.qnap.com
TS-EC1279U-RP
tieto.com
Tieto - Enabling Mobile Life With over 18.000 experts, Tieto is a leading provider of IT, telecommunications and product engineering services. Our product engineering solutions comprise design, development and maintenance of software for our customersâ&#x20AC;&#x2122; products. Our strong R&D offering covers telecom networks, mobile devices as well as automotive and industrial R&D areas. Tieto differentiates from competitors through superior customer centricity, unrivalled quality of deliverables and a proven track record of meeting our customersâ&#x20AC;&#x2122; Time to Market and Total Cost of Ownership needs. The Networks R&D business focus is to accelerate your business growth, let your company benefit from our capacity to improve lead-times and quality at the same time as you benefit from Tieto Global Delivery Model that helps in cutting management and administrative costs. We can assist in aligning product and service development effectively and support you in every step of your product life cycle, providing one stable business driven interface. With over 3.500
experts globally, we have extensive experience and competence covering the full range of Radio Access Networks, Wireline Access, Core and Service Networks, Operations and Business Support systems, Signaling, Platform and R&D Environment Services. Our Connected Devices R&D business has successfully supported Tier-1 OEM and Chipset vendor customers in delivering products to the markets globally for over 15 years, sustaining their growth whilst at the same time reducing operational costs. With over 1.500 experts globally, our offerings gives our customers the competitive edge: horizontally for all dominant mobile platforms and application technologies, and vertically across the whole mobile devices technology stack, from HW base porting to applications and mobile services. We ensure reliability through common processes and operational maturity, productivity through extensive experience in developing Smartphone platforms and products, and efficiency through our Global Delivery Model.
Retail
that are both highly responsive and cost effective. In addition, expanded connectivity increases the potential points of failure and the sophistication of security threats, making remote maintenance and sophisticated security capabilities critical. The 4th generation Intel Core processor family addresses these needs with an array of features including high performance, advanced media and graphics capabilities, and hardware-based security and manageability technologies (see Figure 2). Smart, Interactive Retails Systems As retailers seek to gather, analyze, and act on an ever-growing amount of data, compute performance has become increasingly important. For example, displaying targeted ads during checkout can prompt shoppers to make additional purchases, but determining which ads to display is a complex task. Factors to consider include the customer’s purchase history; their current shopping list; the retailer’s inventory levels and promotional strategy; and external factors like weather conditions or upcoming holidays. Kiosks can take customer analysis a step further by employing optical sensors to perform anonymous viewer analytics, collecting data such as age and gender. Quickly processing all of this data can mean the difference between a satisfied shopper and a missed sale. With an up to 15 percent increase in compute performance, the 4th generation Intel Core processor can deliver faster customer analytics and more rapidly display targeted advertising. In addition, Intel® Advanced Vector Extensions (Intel® AVX) 2.0 increases peak floating-point throughput by 2x, significantly accelerating the image
processing required for anonymous viewer analytics. (See ”Faster Signal Processing with Intel® Advanced Vector Extensions 2.0” on page 28 for details.) These upgrades add headroom for performance-intensive analytics while at the same time maintaining the overall responsiveness and experience of the device. Of course, an ad must catch a customer’s eye and hold their attention long enough to encourage an action, such as an additional purchase or return visit to redeem a coupon. To deliver attention-getting graphics, the 4th generation Intel Core processor family offers up to 60 percent graphics improvements for smooth Flash playback and rich interactive 2D/3D graphics. Video playback also gets a boost, with flexible hardware decoding for multiple video streams, now at up to 4K ultra-HD resolution. The integrated graphics support Intel® Flexible Display Interface (Intel® FDI), DirectX 11.1, OpenGL 4, and OpenCL 1.2, as well as the ability to connect up to three independent displays. This highquality visual playback enables compelling customer displays with videos, graphics, and interactive content, as well as a graphically rich screen for sales staff. Enhanced POS Security and Manageability Data compromises can incur thousands of dollars in lost revenue, not to mention the damage they can do to a store’s brand reputation. Preventing these compromises gets more difficult as attackers use firmware to gain access to a device’s operating system and applications, in addition to creating viruses and malware that can disable a POS system or provide thieves with access to sensitive data.
Media & Graphics Improved Video Playback Quality with Media Enhancement and Better Interactive Content with Up To 60% Graphics Improvement Over Previous Generation
Digital Kiosks
Compute Increased Performance of Retail Applications with Up To 15% Improved CPU Performance Over Predecessors
Secure
Digital Signs
New Features for Securing Retail Devices Including Boot & BIOS/FW Protection
Manage More Simplified Retail Device Provisioning and Graceful Shutdown Enhancements
POSs
Power & Form Factor More Innovative Form Factors with Up To 25% Lower Design Power (TDP) Than Predecessor (for U-Series processors)
Figure 2.
Vending Machines
The 4th generation Intel® Core™ processor family delivers greater capabilities for retail devices. See intel.com/performance/ for details. intel.com/embedded-innovator | Embedded Innovator | 8 th Edition | 2013 | 45
Retail
The 4th generation Intel Core processor family includes a comprehensive range of hardware-based platform and data security capabilities (see Figure 3). These include: • Enhanced Intel® Advanced Encryption Standard New Instructions (Intel® AES-NI) for hardware-accelerated data encryption and decryption that won’t impact system responsiveness • Intel® Platform Protection Technology with BIOS Guard for hardware-assisted authentication and protection against BIOS recovery attacks • Intel® Platform Trust Technology for credential storage and key management with Microsoft* Windows* 8 products (Microsoft is an Associate member of the Alliance) • Intel® Boot Guard for an authenticated code module (ACM)-based secure boot that verifies a known and trusted BIOS is booting the platform
Virus/ Malware
System Compromise or Theft
Content Graffiti
Data Theft
Identity Theft
Figure 3.
Provisioning / Updating
Intel security technologies address the full range of threats to retail transactional devices and provide important remote management features.
These technologies are complemented by comprehensive software security solutions from McAfee that protect systems against malware attacks—including providing security protection below OS where today’s most insidious attacks take place. For example, McAfee Deep Defender takes advantage of Intel® Virtualization Technology (Intel® VT) to deliver virtualization-based “below the OS” malware protection without impacting overall system performance (McAfee is an Associate member of the Alliance). In remote manageability, the 4th generation Intel Core processor family helps retailers control IT costs on POS deployments ranging from tens to thousands of systems. New embedded host-based configuration simplifies Intel® vPro™ Technology activation of remote, unattended systems, unleashing the benefits of Intel® Active Management Technology (Intel® AMT) for retail devices at a lower cost, without
Figure 4.
Kontron KTQ87/mITX Mini-ITX motherboard processors offers a full range of I/O and connectivity options.
compromising Intel AMT features or platform security. Embedded hostbased configuration eliminates the need to use complex back-end systems for setup and stationing a person at a device for provisioning. Also new are graceful shutdown enhancements that help IT staff deal with a hung OS. These enhancements helps increase POS system uptime by allowing staff to remotely analyze a problem and then initiate a graceful shutdown, rebooting the device and restoring it to normal operation.
Figure 5.
The Portwell PCOM-B630VG offers modern I/O via a COM Express Type 6 module.
Speeding Time-to-Market Developers interested in offering leading edge POS and kiosk systems based on the 4th generation Intel Core processor family have many ways to get started with a wide range of solutions from Alliance members. One good example comes from Kontron, a Premier member of the Alliance. The Kontron KTQ87/mITX Mini-ITX board (Figure 4) enables graphics-intensive POS applications in a small form factor and small
46 | 2013 | 8 th Edition | Embedded Innovator | intel.com/embedded-innovator
ECC DIMM slots and six SATA 6 Gbps ports with RAID 0, 1, 5, 10, watchdog timer, as well as SMART FAN which controls the fan speed to keep the temperature within a specified range. Support for Intel AMT and Intel® Trusted Platform Module (Intel® TPM) 1.2 help secure the system and enable remote management tools. Axiomtek offers various industrial housing enclosures and can assist with integration requirements.
USB 3.0 DDR3L DIMM
SATA-600 COM 1~5
A Audio LPT Ethernet VGA USB 3.0
LGA1150 A11 Figure 6.
USB B 2.0 2 0 DVI DV D
Internal K/B+M/S I
The Axiomtek SHB130 is designed for high reliability.
thermal footprint. Kontron has integrated an extensive lineup of interfaces, including PCI Express* (PCIe*) 3.0 and mini PCIe slots, USB 3.0, and support for up to three independent 4K displays, making it easy to build out a system and connect retail peripherals. With its rich feature set and long-term availability, the KTQ87/mITX is an excellent fit for retail kiosks and POS. Developers looking for even smaller form factors—or for more customized designs—can choose from an array of COM Express modules. Portwell, a Premier member of the Alliance, offers one such module in its PCOM-B630VG (Figure 5). This COM Express Type 6 module features a cutting-edge expansion interface with I/O including 6 GTps SATA, USB 3.0, 3x DisplayPort, and a PCIe 3.0 interface that can be configured as 1 x 16, 2 x 8, or 1 x 8 + 2 x 4. By combining the module with a custom carrier board, developers can create highly optimized systems with exactly the features they need.
Venture Corporation, an Associate member of the Alliance, offers the compact eIPC380 industrial PC with a ruggedized and low profile enclosure to protect the system in tough operational environments (see Figure 7). Developed with reference to the Intel® Intelligent Systems Framework (Intel® ISF) guideline, the eIPC380 is an excellent solution for bringing connectivity, manageability and security to the retail market segment. The eIPC380 supports Windows 8, 7, and Windows Embedded Standard. Bringing Online Business Smarts to the Physical Store We are at an important inflection point for the retail industry: brickand-mortar retailers have a crucial opportunity to deploy the same Internet capabilities that online merchants use. By taking advantage of Internet of Things technology, retailers can build stronger customer relationships and improve business operations—and the 4th generation Intel Core processor family makes it possible with important enhancements in performance, graphics, security, and manageability. Now is the time to get started, and the Alliance is ready to help developers bring these new technologies to market.
For applications that emphasize uptime—such as unstaffed kiosks—the SHB130 PICMG 1.3 card shown in Figure 6 is a good choice. This full-size CPU card from Axiomtek, an Associate member of the Alliance, features high-reliability features like two unbuffered
For information about retail solutions from the Alliance, see intel.com/SD-retail
To learn more about bringing intelligence and manageability to retail, see intel.com/embedded-retail
From modular components to market-ready retail systems, Intel and the 250+ global member companies of the Intel® Intelligent Systems Alliance (intel.com/ intelligentsystems-alliance) provide the connectivity, manageability, security, and performance developers need to create smart, connected systems. Kontron (intel.com/ea-kontron) and Portwell (intel.com/ ea-portwell) are Premier members of the Alliance; Axiomtek (intel.com/ea-axiomtek), McAfee (intel.com/ ea-mcafee), Microsoft (intel.com/ea-microsoft), and Venture (intel.com/ea-venture) are Associate members of the Alliance.
Contact Intel
Figure 7.
Venture elPC380 industrial PC comes in a ruggedized, low-profile enclosure for use in tough operational environments.
intel.com/embedded-innovator | Embedded Innovator | 8 th Edition | 2013 | 47
Digital Signage
Smart
Digital Signage
for
Small Businesses Simplify Design and Deployment While Adding Performance By Michael Kelliher, Field Applications Engineer, IEI
L
arge retailers may have been the first to embrace digital signage, but today the small business sector is adding significantly to the industry’s growth. Out of this year’s $1.4 billion USD digital signage market in the U.S., for example, nearly half is projected to come from the small businesses (source: Digital Signage Today, January 2013). However, these resource-limited retailers need signage solutions that do not disrupt their business practices and provide nearly immediate return on investment (ROI).
In this article, we show how the 4th generation Intel® Core™ processor codecs via the Intel® Media SDK. Meanwhile, the Intel® Clear Video family can meet these needs, for small and large retailers alike. We HD Technology (Intel® CVT HD) gains a separate Video Quality Engine examine ways to drive ROI with upgraded graphics that capture with new features that improve quality while lowering system power. shoppers’ attention, as well as up to 2x the FLOPS to better analyze viewer behavior. We 2D/3D also look at off-the-shelf signage solutions that Media Display Graphics simplify media player design and deployment. Intel® Intel® Clear Intel® HD Intel® Media Video HD Quick Sync Finally, we examine new security and InTru™ 3D Ultra HD SDK Graphics Technology Video manageability features that maximize uptime and simplify maintenance. High-Performance Graphics The 4th generation Intel Core processor family delivers a host of media, graphics, and display upgrades that enable smooth video and Flash performance as well as rich interactive 2D/3D experiences without the cost or power draw of a separate graphics card (Figure 1). The media upgrades are particularly notable for signage applications, as they improve video decode performance and visual quality for outstanding multi-stream HD media playback. Specifically, Intel® Quick Sync Video adds acceleration for MJPEG and H.264/MPEG-4 AVC Scalable Video Coding (SVC), as well as support for custom
Transfer promotional content super fast to customers’ devices and phones
Figure 1.
Use hardware acceleration to dynamically improve video quality through frame rate conversion, image stabilization and more
Design custom HW accelerated players on industry leading hardware without the complexity of low level APIs, for simultaneous playback of multiple videos (up to UHD)
Mix crystal clear video with immersive 3D and responsive 2D graphics to bring products to life on-screen
Add another dimension to retail video devices with stereoscopic 3D
Increased resolution support from standard full HD to quad full HD for crisper, sharper, and more vibrant images
The 4th generation Intel® Core™ processor family provides a full range of technologies designed to enable immersive graphics experiences and outstanding graphics performance.
48 | 2013 | 8 th Edition | Embedded Innovator | intel.com/embedded-innovator
The 3D graphics engine, known as Intel® HD Graphics, also gets a major boost. With wider data paths, sampler improvements, double clocked geometry, more execution units, and support of the latest 3D APIs (DX11.1 and OGL4.3), the updated engine provides up to 60 percent faster graphics than the previous generation (Figure 2).
Increasing Graphics Performance
75x Increase
80x 70x 60x
4th Generation Intel® Core® Processors
50x
Display capabilities are also significantly improved, with native support for resolutions up to 4K x 2K. The 4th generation Intel Core processors can power three independent displays, and gains the ability to operate the displays in collage mode, which presents the three displays to the operating system (OS) as a single, large display—without a need for dongles or other added hardware. These features enable cost-effective large-scale displays.
40x 30x
3rd Generation Intel® Core® Processors
20x 10x Baseline
2nd Generation Intel® Core® Processors 2006
2008
2009
2010
2011
2012
2013
Intel’s trajectory of increasing graphics performance continues with a 75x increase from 2006 to 2013 and a 30 percent increase from 3rd generation to 4th generation in Intel® Core™ processors. Source: Intel, 3DMark06. For more information see www.intel.com/performance.
Figure 2. Advertisers can also boost ROI by using anonymous video analytics to optimize their campaigns. Analytics applications like Intel ® Audience Impression Metric Suite (Intel® AIM Suite) offer two key optimization techniques: First, they can detect viewer characteristics like age and gender, enabling advertisements to deliver customized messages. Second, they can track the number of viewers and time spent viewing an ad, allowing advertisers to measure ROI and replace underperforming ads. Analytics applications require significant processing power, and can greatly benefit from the new Intel® Advanced Vector Extensions (Intel® AVX) 2.0, which support 2x the FLOPS for faster image and signal processing.
Off-the-Shelf Signage Solutions Developers can take advantage of the performance upgrades with offthe-shelf digital signage players like the iS-3600 and iS-3620 from IEI subsidiary iNDS (Figure 3). The players offer a rich set of standard features in either a stand-alone form factor or an Open Pluggable Specification (OPS) module. The stand-alone iS-3600 incorporates HDMI, DVI, and SDVO outputs that can drive up to three displays in clone, independent, and collage modes. The player supports a wide range of formats—including RMVB, MOV, MPEG1, MPEG2, MPEG4, and WMV7/8/9 s—and comes with an end-to-end software
Figure 3.
2007
package for content layout/placement, playlist scheduling, and player management. Expansion options include SATA or mSATA storage as well as full- and half-size mPCIe slots for wireless LAN, 3G module, or video capture for anonymous video analytics. I/O includes Gigabit Ethernet, audio out, USB 3.0, and RS-232. The iNDS iS-3620 offers all the standard features of the iS-3600 but in the OPS form factor. OPS provides a modular connector and form factor for media players, enabling easy integration with OPS-compliant displays through a slot in the back of the display. This modularity simplifies installation, maintenance, and upgrades of either player or display, leading to better lifecycle management. If a player fails, for example, it can be quickly swapped out. For developers looking to design their own players, IEI offers the Mini-ITX KINO-AQ870. This board can handle up to 16 GB of dual channel DDR3. It has triple display support and dual Gigabit Ethernet connections. I/O includes RS-232/485, USB 2.0/3.0, SATA, HD audio, DIO, SMBus, and TPM, making this board an excellent base for customdesigned digital signage solutions.
The iS-3600 (left) and iS-3620 (right) offer standard features in stand-alone and Open Pluggable Specification (OPS) form factors, respectively. intel.com/embedded-innovator | Embedded Innovator | 8 th Edition | 2013 | 49
Digital Signage
Maximum Security, Minimum Maintenance Although retailers are eager to install highperformance displays, they need to keep costs down. Thus, it is critical for signage systems to include top-notch security and manageability to maximize uptime and simplify repairs. On the security front, the 4th generation Intel Core processors provide enhancements to provide greater “below the OS” malware protection— protecting systems where anti-virus software can’t. These include:
Intel® Active Management Technology (Intel® AMT)-Enabled Platforms
iEZman Management Console
Power Management + Remote Access + Group Management + Patch Management + Alert and Events + Scan for Devices + Alarm Clock
Digital Signage Network
• Intel Platform Protection Technology with BIOS Guard, which blocks BIOS recovery IEI Easy Manager (iEZman) works with Intel® Active Management Technology Figure 4. attacks to enhance remote access management of digital signage installations. • Intel® Platform Trust Technology and Intel® Boot Guard, which work with Microsoft* Windows* to defend against boot-level malware A companion utility, IEI One Touch Recovery, allows a retailer or signage (Microsoft is an Associate member of the Alliance) vendor to quickly backup and recover the OS, decreasing downtime • Intel® Data Protection Technology with Intel® Advanced Encryption from system failures. The utility performs backups and recoveries from Standard New Instructions (Intel® AES-NI), which supports rapid a hidden partition not accessible from the OS. As with the other tools, data encryption and decryption—a key concern for signage and all tasks can also be performed from a remote management console. kiosk systems designed to capture customer information or hold proprietary information A third tool, IEI Intelligent System Management Module, uses onboard sensors to monitor and manage CPU and system voltages, CPU and To help manage operating costs, the 4th generation Intel Core system temperature, and fan speed. The retailer or signage vendor processor family provides enhanced manageability with Intel® Active can use the module to set thresholds on critical parameters and get Management Technology (Intel® AMT). This technology provides “outalerts if these are exceeded, allowing them to take corrective actions. of-band” access, meaning technicians can access and repair signage players remotely, even if the OS or BIOS is damaged or the player Helping Retailers Take the Next Step is powered down. This technology is particularly useful for small The combination of smart digital signage solutions based on retailers, as it allows them to rely on the signage vendor’s centralized 4th generation Intel Core processors and well-designed management IT staff for cost-effective maintenance. tools provide a nearly foolproof package for both small and large retailers looking to get into digital signage or upgrade their system to A valuable update in Intel AMT 9.0 is the addition of Embedded Hostnew levels of impact, performance, uptime, and cost-of-ownership Based Configuration. This new capability simplifies remote device savings. There’s never been a better time or way to reach out and provisioning by enabling centralized IT staff to remotely provision engage customers, providing in-store experiences that surpass unattended signage. No longer must a store employee be present for online shopping experiences. Today’s latest digital signage solutions a device to be provisioned. Intel AMT 9.0 also adds graceful shutdown, level the playing field and give in-store customers brand experiences which improves uptime and reduces in-store visits by allowing IT staff that make them not only come to the store to “shop,” but motivate to remotely analyze a hung OS, initiate a graceful shutdown, reboot the them to buy. device, and restore it to normal operation. For more information on solutions from IEI, see To help with management, recovery, and diagnostics, IEI includes a intel.com/SD-iei-products robust package of software tools with all its players and board-level products. These powerful tools enable monitoring and control of digital To learn more about bringing intelligence signage through an easy-to-use Windows-based graphical user interface and manageabilty to digital signage, see (GUI). IEI Easy Manager (iEZman) is a compact server-management tool intel.com/embedded-retail built on top of Intel AMT (Figure 4). It provides local and remote control. Users can manage and secure networked computing resources, plus IEI Technology Corp. (intel.com/ea-iei) is an Contact IEI remotely inventory assets even with the power off. iEZman facilitates Associate member of the Intel® Intelligent the use of Intel AMT’s out-of-band management to remotely upgrade or Systems Alliance, and a leading provider of industrial computing repair systems and push software updates. It can also redirect keyboard, platforms and OEM/ODM services. IEI has an innovative R&D team, video, mouse, and storage control to IT management. Even the power effective management system, dedicated quality assurance, and over status of a device can be controlled remotely to ensure power is shut 400 products passing through more than 100 distributors, providing down during off hours to save on energy costs. customers worldwide with fast time-to-market. ®
50 | 2013 | 8 th Edition | Embedded Innovator | intel.com/embedded-innovator
Featured Solutions
Seneca
Creating Solutions
800.227.3432 marketing@senecadata.com www.senecadata.com
To Simplify Technology.
Seneca is a leading US-based custom computer and appliance
business, education, healthcare, digital signage, digital security and
manufacturer and value-added distributor with an international
surveillance, and digital broadcast customers around the world.
reach. With over 30 years of experience, Seneca has taken our
Â
expertise and passion for engineering to provide our partners
Our engineering expertise is evident in our NexlinkÂŽ brand of
with a complete technology lifecycle of services. With a partner
custom-built computers, xVaultâ&#x201E;˘ storage platforms, digital signage
ecosystem of global leaders in technology, logistics, and financial
HD media players, and digital broadcast solutions.
services, Seneca creates solutions that simplify technology for
intel.com/embedded-innovator | Embedded Innovator | 8 th Edition | 2013 | 51
Featured Solutions: Intel Haswell Microarchitecture
TenAsys Corporation
Hard real-time embedded virtualization solutions supporting mixed workloads on multicore PC platforms
+1 503 748-4720 sales@tenasys.com www.tenasys.com
TenAsys® leadership in the embedded virtualization initiative for PC platforms delivers software solutions reduce cost and simplify system consolidation. Our product portfolio provides the most comprehensive set of platform solutions available anywhere: • INtime for Windows, our native Real-Time Operating System (RTOS) that runs alongside Windows, • INtime® Distributed RTOS, the scalable RTOS version that runs applications across one or several processor cores and/or platforms, and
TenAsys delivers revolutionary engineered solutions for evolutionary applications. Our products are cost effective and easy to use combining world class development tools for optimal performance and integration. Solutions are field proven in industrial control, material handling, robotics, test & measurement, military, medical, and laboratory systems for over 30 years. TenAsys headquarters are based in Beaverton, OR, with supporting sales offices serving Europe from Germany, and Asia from Japan.
• eVM® for Windows, featuring HaRTH™ technology, combines our hard real-time hypervisor enabling native embedded and 3rd-party RTOS-based applications alongside Windows.
52 | 2013 | 8 th Edition | Embedded Innovator | intel.com/embedded-innovator
intel.com/embedded-innovator | Embedded Innovator | 8 th Edition | 2013 | 53
Intel® Intelligent Systems Member Directory
C
om
m
s
st ria l
et ai l
R
du
In
ca re M ilA er o /S ec ur En ity er gy
ea lth
ut om
pp
Market Segments
Portwell www.portwell.com.tw
Aaeon-Onyx www.aaeon.com
ADI Engineering, Inc. www.adiengineering.com
ADLINK www.adlinktech.com
Amax Engineering Corp. www.amax.com
Arrow - North America www.arrow.com
Avalue - BCM www.avalue.tw.com
AXIOMTEK Co. Ltd. www.axiomtek.com
CIARA Technologies www.ciaratech.com
congatec AG www.congatec.com
Dedicated Computing www.dedicatedcomputing.com
DFI-ITOX www.dfi-itox.com
Digi International www.digi.com
DTx Inc. www.dtx.com
Eurotech www.eurotech.com
Evoc Intelligent Technology Co., Ltd. www.evoc.com
GE Intelligent Platforms www.ge-ip.com
Gemalto M2M GmbH www.cinterion.com
GraniteMEDSystems www.granitemed.com
Hewlett-Packard www.hp.com
IBASE Technology Inc. www.ibase.com.tw
IEI Technology Corp. www.ieiworld.com
Emerson Network Power www.emersonnetworkpower.com/embeddedcomputing
Lanner Electronics Inc. www.lannerinc.com
McAfee www.mcafee.com
Norco-Habey www.norco-group.com
Oracle www.oracle.com/us/products/servers-storage/servers/index.html
QNAP Systems, Inc. www.qnap.com
Radisys www.radisys.com
SBS Ltd. Science & Technology Co. www.sbs.com.cn
Seneca www.senecadata.com
Shenzhen Topstar Technology Co., Ltd. www.cszte.com
Telit Wireless Solutions www.telit.com
Xyratex International Inc. www.xyratex.com
6WIND SA www.6wind.com
Adeneo Embedded www.adeneo-embedded.com
ads-tec GmbH www.ads-tec.de
Anovo Co., Ltd. www.anovotech.com
American Megatrends, Inc. www.ami.com
Altera Corporation www.altera.com
ALT Software Inc. www.altsoftware.com
Wind River Systems www.windriver.com
UNICOM Engineering (fka NEI) www.nei.com
Premio Inc. www.premioinc.com/ems/about/?pg=embedded
Patriot Technologies, Inc. www.patriot-tech.com
Nexcom www.nexcom.com
Network Allies www.networkallies.com
Aewin Technologies Co. Ltd www.aewin.com.tw
NCS Technologies, Inc. www.ncst.com
Venture Corporation Limited www.venture.com.sg
Microsoft Corporation www.windowsembedded.com
M&A Technology, Inc. www.macomp.com
National Instruments www.ni.com
Intequus (a division of Equus Computer Systems) www.intequus.com
Affiliate Members
ot iv e
In
In
ls
lic at io
n
ls
To o
To o ftw
ar e
Sy st em
ar e
g
ar dw
pe ra tin
H
Premier Members
Expertise Areas
Company Name
Associate Members
A
te gr at io te n O gr D at M i on /E M O S EM
Kontron www.kontron.com
Sy st em
A
Dell OEM www.dell.com/oem
So
H
IS V
Advantech www.advantech.com
O
B
en ta ry
&
im pl
ar e
C
om
H
ar dw
Si
oa rd s
lic on
From modular components to market-ready systems, Intel and the 250+ global member companies of the Intel® Intelligent Systems Alliance provide the performance, connectivity, manageability, and security developers need to create smart, connected systems. Close collaboration with Intel and each other enables Alliance members to innovate with the latest technologies, helping developers deliver first-in-market solutions. Learn more at: intel.com/go/intelligentsystems-alliance. Explore their products and services at: intel.com/go/intelligentsystems-alliance-solutions.
Archermind Technology (Nanjing) Co. Ltd www.archermind.com
Arium www.arium.com
EEPD www.eepd.com
EMBEDTEC Science & Technology Co., Ltd www.embed-tec.com
Ennoconn www.ennoconn.com
Exar Corporation www.exar.com
Green Hills Software, Inc. www.ghs.com
HCL Technologies Ltd www.hcltech.com
Hectronic AB www.hectronic.se
Insyde Software Corp. www.insydesw.com
Jabil Ltd www.jabil.com
Lauterbach GmbH www.lauterbach.com
Macraigor Systems LLC www.macraigor.com
PFU Ltd. / PFU Systems, Inc. www.pfusystems.com
PROMISE Technology Inc. www.promise.com
Protech Systems www.protech.com.tw
Qosmos www.qosmos.com
Red Flag Software Co., Ltd. www.redflag-linux.com
ROHM Co., Ltd. / Lapis Semiconductor Co., Ltd. www.rohm.com/index.html
SECO www.seco.it
TimeSys Corporation www.timesys.com
Toshiba Personal Computer System Corporation www.toshiba-tops.co.jp
Trenton Systems www.trentonsystems.com
u-blox AG www.u-blox.com
Tieto Signaling Solutions www.tieto.com
s
Shenzhen Prafly Technology Co. Ltd www.prafly.com
TenAsys www.tenasys.com
m
Signal Integrity Software, Inc. www.sisoft.com
om
Sensory Networks Inc. www.sensorynetworks.com
C
Real-Time Systems GmbH www.real-time-systems.com
st ria l
QNX Software Systems, Ltd. www.qnx.com
et ai l
R
Phoenix Technologies, Ltd. www.phoenix.com
du
Netronome www.netronome.com
In
NEC Corporation, Application and Appliance Division www.nec.com
ca re M ilA er o /S ec ur En ity er gy
ot iv e
ea lth H
ut om
NASoftware Ltd www.nasoftware.co.uk
Universal Scientific Industrial Co. Ltd www.usi.com.tw
Wipro Technologies www.wipro.com
Wistron Corporation www.wistron.com Xilinx, Inc www.xilinx.com
Nanjing Byosoft Co., Ltd www.byosoft.com.cn
WIN Enterprises www.win-ent.com
Mobica Limited www.mobica.com
Thecus Technology Corporation www.thecus.com
MindTree Ltd www.mindtree.com
Mentor Graphics www.mentor.com
Super Micro Computer Inc. www.supermicro.com
Klocwork www.klocwork.com
Shenzhen Seavo (China) Technology Co., Ltd. www.seavo.com
ILS Technology www.ilstechnology.com
Ricoh www.ricoh.com
LynuxWorks, Inc. inside@lnxw.com
DUX Inc. www.dux.jp
MSC Vertriebs GmbH www.msc-ge.com
A
lic at io
pp A
CriticalBlue www.criticalblue.com/prism/intel
Micro Star International us.msi.com
n Sy In te st gr em at In io te n O g ra D M tio /E n M O S EM
ls
ls
To o
To o ftw
ar e
Sy st em g
ar dw
So
DT Research,Inc. www.dtresearch.com
Bsquare www.bsquare.com
BroadWeb Corporation www.broadweb.com
Affiliate Members
H
pe ra tin O
IS V
MEN Mikro Elektronik GmbH www.men.de
Market Segments
Aricent www.aricent.com
Concurrent Technologies www.gocct.com
ar e
lic on Si
oa rd s
en ta ry
B &
im pl
ar e
om C
ar dw H
Arbor Technology www.arbor.com.tw
Expertise Areas
Company Name
* Intel® Intelligent Systems Alliance memberships as of 7/22/2013. Premier, Associate, and Affiliate listing only. For the latest listing of membership visit: www.intel.com/go/intelligentsystems-alliance-roster.
Your One Stop for Intelligent Solutions The NEW Intel® Intelligent Systems Alliance’s Solutions Directory Here’s your trusted resource for connecting with solutions and partners offering over 2,500 leading-edge embedded hardware, software, tools, and services. Use it now to help you get products to market faster. intelintelligentsystemsalliance.com/solutions-directory