AUTONOMOUS TERRITORY Arda Ertan Yildiz, Yongyi Wu
TABLE OF CONTENTS
0. INTRO: DATA AND DATA CENTERS
4
1. TOWARDS HYPERSCALE
12
2. FROM ENIAC TO THE CLOUD
20
3. DUTCH DATA CENTER LANDSCAPE
38
4. A CASE STUDY: EQUINIX AM4
48
5. INTERVIEW: SMARTDC IN VAN NELLE
58
6. REVERSE THE CLOUD
62
7. DATA CENTER SECURITY
68
8. AI AND UNMANNED FUTURE
76
4
Autonomous Territory
MATERIALIZATION OF DATA Google data center Photos by Connie Zhou
“By 2021, more than 90% of large data centers will revise their strategies due to global socioeconomic and environmental trends.� Gartner, 2016
Autonomous Territory
5
INTRO: DATA AND DATA CENTERS
Data volume of global consumer IP traffic (EB per month) Source: Statista
6
Autonomous Territory
GLOBAL INTERNET SCENARIO
“Despite the fact that hardware is constantly getting smaller, faster and more powerful, we are an increasingly data-hungry species, and the demand for processing power, storage space and information in general is growing and constantly threatening to outstrip companies’ abilities to deliver. Courtney, Digital Science Series, EMC A study by International Data Corporation for EMC estimated that 1.8 trillion gigabytes, or around 1.8 zettabytes, of digital information was created in 2011. The amount of data in 2012 was approximately 2.8 ZB and is expected to rise to 40 ZB by the year 2020. We’re using the internet far more than we used to. So says new data from USC Annenberg about the digital lives of Americans. Since 2000, time spent online every week by an average American has risen from 9.4 hours to 23.6. Of that, time spent ogling the internet at home has risen from 3.3 to 17.6 hours a week over the same period. That’s a lot of screen time.
global high speed network which consist of optical fiber cables that connect almost all the data centers all over the world. DATA CENTERS IN THE NETWORK
Currently there are 4417 traceable colocation data centers from 122 countries in the index of Data Center Map website. But obviously these are just a small part of all data centers, especially not including a large number of private data centers. These facilites used roughly 416 TWh in 2016, nearly 40% more than the entire United Kingdom. Different data centers are owned by different enterprises, institutions or specialized cloud providers like Google, Amazon, Microsoft or IBM, running different functions. For instance, if we watch a video on Youtube, the streaming data actually from one of the server in the Youtube data center which probably is in US. The digtial signal travel all the way though different IT equipments in multiple layers, intercontinental submarine cable, your internet service provider’s switch and finally router to your home. Heavy physical equipments are used in every step of network but most of them are even unnoticeable in our daily life.
Behind all of this is a exponentially growing
Autonomous Territory
7
8
Autonomous Territory
ALWAYS ON
Internet companies logo collection Images from Internet
“57 percent of the global population is now connected to the internet, spending an average 6.5 hours online each day.� Global web index 2019
Autonomous Territory
9
10
Place here your project title
GLOBAL NETWORK
Submarine cables and data centers all over the world Image by Arda Ertan Yildiz, Yongyi Wu
Place here your project title
11
User
Access point/Router
Modem
Internet servic
Youtuber
Edge router
12
Autonomous Territory
HOW INTERNET LOOKS LIKE IT equipments in data processing Image by Arda Ertan Yildiz, Yongyi Wu
ce provider switch
Internet exchange point switch
Aggergation router Server Storage area network
ToR switch
Storage server
Core router
Autonomous Territory
13
TOWARDS HYPERSCALE
Apple CEO Tim Cook (center) breaks ground with Nevada Governor Brian Sandoval (right) and Reno Mayor Hillary Schieve (left) on a new data center. https://www.macworld.com/article/3274584/where-does-apple-stores-your-icloud-data.html
14
Autonomous Territory
“‘The cloud’ is a concept, not a place, although collectively, the cloud is made up of places.” Glenn Fleishman, Macworld Cloud computing refers to an aggregation of servers that you don’t have to know much about to gain the benefit. This can be for storage, like iCloud Drive or Amazon Cloud Drive or Google Drive, or for computation and other more specialized purposes. When you use a cloud-based service, your actions and data may be split among many machines and drives—maybe even across continents. So called data centers are these very huge infrastractures. They are geometric, neutral and very different from living enironment of human. Which probably remind us of superstudio’s the continuous monument. The boom of data centers came during the dot-com bubble of 1997–2000. Companies needed fast Internet connectivity and nonstop operation to deploy systems and to establish a presence on the Internet. Installing such equipment was not viable for many smaller companies. Many companies started building very large facilities, called Internet data centers, which provide commercial clients with a range of solutions for systems deployment and operation. New technologies and practices were designed to handle the scale and the operational requirements of such large-scale operations. These practices eventually migrated toward the private data
Autonomous Territory
centers, and were adopted largely because of their practical results. Data centers for cloud computing are called cloud data centers, and the largest among them are called hyperscale data center. GLOSSARY
[HYPERSCALE DATA CENTER] - A Hyperscale (or Enterprise Hyperscale) data center is a facility owned and operated by the company it supports. This includes companies such as AWS, Microsoft, Google, and Apple. They offer robust, scalable applications and storage portfolio of services to individuals or businesses. - Hyperscale computing is necessary for cloud and big data storage. - Has anywhere from 500 Cabinets upwards, and at least 10,000sq ft. in size. - Usually have a minimum of 5,000 servers linked with an ultra-high speed, high fiber count network. - May use external companies on initial fit outs before maintaining internally. - Noticeable difference from Enterprise to Hyperscale is the High Fiber Count utilized across the network. [Cloud computing] - Cloud computing is the the use of various services, such as software development platforms, servers, storage and software, over the internet, often referred to as the “cloud.” - The back-end of the application (especially hardware) is completely managed by a cloud vendor. - A user only pays for services used (memory, processing time and bandwidth, etc.). - Services are scalable
15
16
Autonomous Territory
Autonomous Territory
17
18
Autonomous Territory
THE CONTINUOUS MONUMENT
On the Rocky Coast, project (Perspective), 1969 Image by Superstudio
“On the other hand, during those same years we realized that society could no longer be identified with the rationalmechanical model of the all-production factory, of early capitalism, but that we were well on our way to another model, Data center the supermarket, a place of consumption, a faceless, anonymous container within Data which goods were displayed and moved without requiring Architecture in order to be represented.� CRISTIANO TORALDO DI FRANCIA - founder of Superstudio
Autonomous Territory
19
Colocation
20
Autonomous Territory
FOOTPRINT
Top view of different data centers on the same scale Image by Arda Ertan Yildiz, Yongyi Wu
Hyperscale
Autonomous Territory
21
FROM ENIAC TO CLOUD
Global data center IP traffic, by data center type (EB per year) Source: Statista
22
Autonomous Territory
An increasing amount of computing work is moving from traditional to hyperscale cloud data centres. According to Cisco’s Global Cloud Index, by 2021, 94% of all workloads and compute instances will be processed in cloud data centres. Hyperscale data centers will grow from 338 in number at the end of 2016 to 628 by 2021. They will represent 53 percent of all installed data center servers by 2021. The use of private (rather than public) hyperscale data centres is expected to grow 11% a year from 2016 to 2021. So why and how this shift will happend and why hyperscale is important for us? DATA CENTER TIMELINE
People use specific tools to process data from very long ago, like anicent rome us abacus to calculate taxes. Until last century, mechanical analog computer like differential analyzer start to took a big step to what we called computer nowadays. Finally ENIAC, one of the first modern computer launched in 1946. When we go through supercomputer data center, mainframe data center, client-server data center, colocation data center to the cloud data center wich is also hyperscale, is easy to notice that the way of organize server are changing because its amount and size and we can see the volume of data center is getting bigger, need more area for specific supporting equipments but human are in the data center is getting smaler.
Autonomous Territory
There are some hidden information in this timeline if we look at some cyber characterisc. Data center start from people working inside the computer, then able to control mutiple computer in a local network, then different centrilize their equipment in a colocation then in the end everyone use computing power from a huge computer pool in a global highspeed network. In this process we can see hwo we gradually get away from physical infrastracture when we use more and more data in our daily life.
“If you look back over the history of computing, it started as mainframes or terminals. As PCs or work stations became prevalent, computing moved to the edge, and we had applications that took advantage of edge computing and the CPU and processing power at the edge. Cloud computing brought things back to the center.� Peter Levine, venture capital investor
23
24
Autonomous Territory
BEFORE COMPUTER
Above: Roman tax collector calculating someone’s taxes on an abacus ( 225 AD ) Below: Vannevar Bush with his differential analyzer (1930s) Photos from Internet
“The Encyclopedia Britannica could be reduced to the volume of a matchbox. A library of a million volumes could be compressed into one end of a desk.” VANNEVAR BUSH, As We May Think (1945)
Autonomous Territory
25
26
Autonomous Territory
DATA CENTER EVOLUTION I
The timeline divides data center development into 5 steps by critical computer and Internet techonlogies Image by Arda Ertan Yildiz, Yongyi Wu
Autonomous Territory
27
28
Autonomous Territory
167 sqm Dataspace
Supercomputer E.N.I.A.C
1946
single server 200-digits
Autonomous Territory
29
30
Autonomous Territory
29 sqm Dataspace
Mainframe IBM 1401
1971
single server 16000 characters
Autonomous Territory
31
32
Autonomous Territory
125 sqm Dataspace
Client-Server
Business Private Datacenter
1970-1990s
multiple server Megabyte-Gigabyte
Autonomous Territory
33
34
Autonomous Territory
225 sqm Dataspace
Virtualized
Colocation Datacenter
1990-2010s
800 servers Gigabyte-Terabyte
Autonomous Territory
35
36
Autonomous Territory
1345 sqm Dataspace
Cloud
Cloud Computing Datacenter
2010-Today
6000 server Terabyte-Petabyte
Autonomous Territory
37
38
Autonomous Territory
DATA CENTER EVOLUTION II
Changing relationship between people and computer (data center) Image by Arda Ertan Yildiz, Yongyi Wu
Autonomous Territory
39
DUTCH DATA CENTER LANDSCAPE
AMS-IX is one of the largest Internet exchange point in the world, based in Amsterdam, NL. Photo by HUAWEI
40
Autonomous Territory
DIGTIAL GATEWAY TO EUROPE
In 1988 the Dutch were second in the world to connect to the Internet. The Netherlands quickly developed excellent connections, and the population embraced the Internet. Today, the Netherlands is the best-connected countries in the world, according to the DHL Global Connectedness Index (2016), regarding cross-border flows of trade capital, information and people. The Netherlands is host to the Amsterdam Internet Exchange (AMS-IX), the largest in the world together with Frankfurt-based DE-CIX. Number 4 is IX.br in Brazil and number 4 is LINX in London. Another large Dutch internet exchange is the NL-IX, which climbed to a 6th place globally and has 630 connected networks. AMS-IX stated in their 2018-report that their overall Internet traffic volume increased from 12 EB in 2016 to 13 EB in 2017. The peak traffic at the Exchange grew from 5 to 5.5 Tbps. Last year, 84% of new AMS-IX customers came from abroad, which shows that The Netherlands is an attractive foreign candidate when it comes to digital infrastructure. State of the Dutch data hub report 2018
DUTCH DATA CENTER
Until May 2018, there are 198 multi-tenant data centers in Netherlands with 308,000 sqm data floor totally. And 71% of floor space in located in Amsterdam region. The importance of Amsterdam for the Dutch multitenant data center market continues to grow, like it did over the last twelve months.
Autonomous Territory
This growth does not come without its challenges. Similar to the other European top locations, access to power is becoming a challenge for multi-tenant data centers in the Metro Region Amsterdam (MRA). In a survey among members of the Dutch Data Center Association, Pb7 Research found that more than half of the data centers mention access to power as a key challenge for the next three years, as opposed to none of the data centers from other regions. In line with that challenge, MRA data centers are very concerned about finding suitable locations. State of the Dutch data report 2018
CLOUD IN NETHERLANDS
The strong growth of multi-tenant data centers has, more than anything, been the result of the rise of cloud computing. Cloud computing continues to shape the data center industry in various ways. It has resulted in the arrival of hyperscale data centers, in the Netherlands represented by Microsoft and Google Microsoft and Google have so far built 71.520 m2 of data center floor and are far from done. Interestingly, hyperscale data centers are not attracted to metropolitan areas such as Amsterdam. They require excellent connectivity (for example where an ocean cable comes ashore), access to large quantities of green power and affordable locations. As a result, new data center onevendor hotbeds are emerging in Groningen and the north of North Holland. State of the Dutch data report 2018
41
0
High population density region Low population density region Hyperscale Data center and its capacity
50km
Other Data center and its capacity
GRONINGEN
Hyperscale campuses
NORTH HOLLAND
Hyperscale campuses
42
Autonomous Territory
0
High population density region Low population density region Hyperscale Data center and its capacity
50km
GRONINGEN Regional Hub
AMSTERDAM Regional Hub
AMS-IX
ROTTERDAM Regional Hub
EINDHOVEN Regional Hub
Autonomous Territory
43
0
Internet backbone Other Internet cable Data center and its capacity
50km
44
Autonomous Territory
0
380kV connection/substation 380kV connection/substation project 380kV interconnection 380kV interconnection project Subsea interconnector
50km
Subsea interconnector project 220-150-110 connection/substation Power plant Areas below sea level Data center and its capacity
Autonomous Territory
45
GRONINGEN
Hyperscale campuses
Windpark Eemshaven-west Vopak Solarpark 0
Eemshaven
Vattenfall Solarpark Google
Windpark Delfzijl-noord Solarpark Weefsweer
Solarpark Delfzijl
12.5km
Windpark Delfzijl-zuid
Groningen
46
Autonomous Territory
0
NORTH HOLLAND
Hyperscale campuses
Middenmeer
Windpark Wieringermeer AgriportA7 Greenhouse AgriportA7 Greenhouse
Microsoft
12.5km
Amsterdam
Autonomous Territory
47
48
Autonomous Territory
DATA CENTER AGRIPORT A7 EVOLUTION II
Microsoft hyperscale Changing relationshipdata between center people (left) and and agriport computer (data A7 greenhouse center) (right) Image from by Arda Google ErtanEarth Yildiz, Yongyi Wu
Autonomous Territory
49
A CASE STUDY: EQUINIX AM4
Data center AM4, Benthem Crouwel Architects Photo by Jannes Linders
50
Autonomous Territory
The tower with a height of seventy-two meters has been opened in 2017 on the Science Park, an academic campus in Amsterdam. From the ringroad A10 this ‘cloud box’ is clearly visible. The new building is Equinix’s second data center on the Amsterdam Science Park. The campus processes about 38 percent of all Dutch data traffic. In Europe, the Amsterdam area – together with London and Frankfurt – is one of the most important data center hubs. In 2012, the first data center was opened, AM3 – with horizontal lamellae, and now the tower – AM4 with 24.000m2 of server space – has been added. The buildings are linked by bridges. We designed a canal as the first layer of safety, a more friendly alternative to barbed wire fences. Employees and visitors then go through an identity checkpoint in a light and spacious lobby. To reach the data center, the visitor walks over a bright red bridge which marks the passage to a strictly secured area. After scanning fingerprints, one enters the ‘white space’, where the servers are located. It is a totally different experience from being in a converted warehouse, the type of building mostly used for data centers. The exterior is grand, clean and cool: it resembles a large hard disk, with horizontal slants passing the air from the cooling system. The high-rise building is made of triangular aluminum profiles, which are black on one side and silver on the other, and become more narrow in the top – an optical illusion that makes the tower look slimmer.
When heat is generated, it provides heat for other buildings at Amsterdam Science Park. “On the one hand, you see that we’re increasingly storing data in the cloud; currently, there is a lot of demand of new datacenters. On the other hand, there is continuous miniaturization, which results in less storage space. We have also taken into account that when less server space is needed, the facades can easily be replaced, and the building can be easily transformed into a laboratory, office and/or apartments.” Benthem Crouwel Architects, Temple of the 21st century: Data centers Equinix in Amsterdam
FACTS ARCHITECT: Benthem Crouwel Architects, Generaal Vetterstraat 61, 1059 BT Amsterdam CLIENT: Equinix Netherlands bv LOCATION: Science Park 610, Amsterdam GROSS FLOOR AREA: Bruto floor area: 6.500m2 AM3 ‘ 24.000m2 AM4 START DESIGN AM4: 2014 START CONSTRUCTION AM4: 2015 OPENING AM4: July 2017 OPENING AM3: October 2012 PROJECT TEAM: Jan Benthem, Joost Vos, Saartje van der Made, Guido de Wit, Gwendolyn Huisman, Henk van Rossum, Volker Krenz, Jeroen Jonk, Noortje ter Heege, Marcel Wassenaar, Walter Brouwer STRUCTURAL ENGINEERING/MECHANICAL SERVICES/BUILDING PHYSICS: Royal Haskoning DHV CONTRACTOR: UDD: Unica and Dura Vermeer LIGHTING DESIGN: Bartenbach Lichtlabor PHOTOGRAPHER: Jannes Linders
AM4 is equipped with energy storage systems.
Autonomous Territory
51
52
Autonomous Territory
DATA VS. HUMAN
Data and human both have their own space in city. Different kind of infrastructures serve for them which creates a seperation and defines their own territory. Red lines indicate fiber optic cable network and blue lines indicate human traffic.
Human Data Autonomous Territory
SCIENCE PARK,AMSTERDAM. 53
54
Autonomous Territory
AM4
Above: The entrance Below: Inner space with data servers Photos by Jannes Linders
“The buildings had to fit in with its surroundings. They must be attractive, however not too welcoming. We all want our personal data to be safe, but we feel awkward about large data centers in our cities, because we don’t want fortress-like buildings surrounded by high-security fences in public spaces.” JOOST VOS - partner of Benthem Crouwel Architects
Autonomous Territory
55
56
Autonomous Territory
Logistic Exit
AM3 DC
Entrance AM4 DC
Logistic Entrance Customer Center& Office
Human Power Data
Autonomous Territory
57
58
Autonomous Territory
A NEW DAWN
AM4 for Equnix is on the cover of the 2017/2018 yearbook architecture in the Netherlands Image from Benthemcrouwel.com
“We are designing over 40 years beautiful buildings, famous ones, we never reach the cover of the architectural year book. To our own surprise, with this building without any windows in it, it reach the cover of year book. I think it’s sort of an example what is this kind of new typology doing with our surrondings.” JOOST VOS - partner of Benthem Crouwel Architects
Autonomous Territory
59
60
Autonomous Territory
SMARTDC IN VAN NELLE Richard Boogaard
managing director via SmartDC, Van Nelle Factory, 51.9231° N, 4.4340° E
DP: So we’ve already read something. We belong to chair complex projects where we like complex projects in general. And we are both doing like research design studio. And the title of this tool is called the Datapolis. So the idea is to understand the impact of data and how us as architects can understand better all infrastructure behind the data. Because it’s kind of a totally new field for architecture in general. We’re looking in different scales from homes to offices, factories, data centers, satellites, any kind of architecture needed and exchange, storage of data, infrastructure architecture, but also building and factories and houses. We’re based in Delft. So in fact, also we are based in Netherlands we’re trying to visit a few important sites in the Netherlands. And when we saw that, you are the biggest data center in Rotterdam. And that how we are related to the Van Nelle Factory, it was interesting to try to come and look at it. RB: Well, it’s good to the starting point. You’re here at smartDC and my name is Richard Boogaard. I’m the managing director of Smart DC.I’m also the owner. Just recently sold to Ubisoft last year. I just builded from scratch. And it’s nice to mention Van Nelle factory because nowadays, I think since three-four years it’s a UNESCO World Heritage monument. So it makes it quite interesting for you guys as architects for building itself. If you have the time left later today, you can visit here. Then you have you can see all kinds of things which is back in the days in the beginnings of 1930s where the doors are closed, not by typical door closings, but it’s a ramp which is putting a little bit up. For the main entrance, you have different entrances for men and for women. They’re cross-links so you cannot touch one another. So those kind of things were for basically quite normal in the beginning 30s. This building is built in roughly about in 1970 due to the expansion of the production of coffee-tobacco at the back of us, which is the main building. They had the storage, The long term storage was in the in whitespace of data center and for this part it was(entrance of smartDC) it was Albert Heijn’s.
Autonomous Territory
So if you look at this building, what we’ve done is that, when we started to build smartDC,I wrote a business plan. And it was back in 2000s, end of 2008, beginning of nine, and there was the banking crisis. So we had to come up with a plan so that we could lend some quite some large amounts of money, because data center is a nice business, but you have to invest quite a lot before you can attract customers as well. So what have we done? We built the kind of box in box principles and we started to reorganize this complete facility here. The main reason for us to come to this location is due to the factory of coffee tobacco. It has a quite large infrastructure for power, and the power grid is connected to the 23,000V, which is also the hardware standards and also where the trains are going on. So it’s a little bit higher as the where the rest of the industry is connected to. That makes it less independent for all kinds of maintenance. So over the years we had two outages. So when they come up we have to always have power. Datacenter is local of mobile phones, everything you do on the mobile phone comes somewhere out of a data center and just think of the idea that all data centers are connected in one way or another with fiber. Anywhere in the world, so, you have sea cables running. You have six major lines running from Europe to the US. And also that is why everything is connected. Here in the data center. We quite lucky to have a big connection, which is 12 megawatts. total capacity.
“There’s also another quite important thing for datacenter whoch is connectivity” And when we started here, we had quite some look, because the complete backbone of Amsterdam, Brussels, Paris runs just in this street here. So for us to interconnect all major carriers is quite easy. So basically, that is one of the things, it’s quite important and the rest you can build yourself. So when we started, we had an idea to have built-in modules. Smaller modules are arranging from 30,40 or 50 rack entities just filled up complete facility
61
62
Autonomous Territory
here. And every time when we have new customers, I just go to the investors with the new contracts and then I get another money. It was quite necessary. So this is what you see here is the main entrance where the lorries came into the facilities in here and doors are taken out. I was arguing for roughly about five six years to have all the doors fixed wall. But the UNESCO and the local government said that it is not possible. So, we have a wall looks like a door that is what this is about. So basically that is in general what it is. The main entrance here is redone roughly about in the end of 2017. As last part because what I feel is important is the infrastructure for our customers it’s where we are making our money. And then when the it loads was depending more power was first of all, it was roughly about 500 watts 1000 watts in Iraq, and now they’re pushing five kilo kilo waters 5000s 10,000 okay perfect. So that is quite quite a lot. So they had sold had to come up with an idea just to close the cold cord or because it was already there to have more cold into directs so that could not escape any. Reason why we do it the other way around is that we have the whole car, the whole Ali’s and it’s respect to the record unit and the whole air it makes it more efficient for the crack unit itself. So that is one of the advantages what you find here. I suggest we started the tour over one of the enjoying the store right now. Meeting rooms available to our customers as well. They can look at our control panel. They can take the photos to take the photos from the area here. And in the data center do not take pictures or the name text of likely for instance so for tech, it’s quite right to have pictures of ice or smartest evilness. And because I don’t want to have your customers, displays, whatever. We did it together with my business partner. And he is not so famous Netherlands itself but has a platform to cross roles for the big publishers. So if you ever played a Call of Duty or Division, FIFA in Western Europe they come here to the data center visit the one of these rooms for them. That is why it’s such a powerful location here because due to the gaming and due to the low latency network which is very important for our small and medium enterprises we have. If there is an issue outside and we have quite a lot of carriers and parties. If one of the lines down to two if it’s broken or maintenance or whatever it is just sent to another carrier. For our customers, they don’t even noticed or what lines they go out and we have connection towards Amsterdam, but also to directly to Frankfurt. In Amsterdam you have two biggest worldwide exchanges. And then if you have an issue one on your servers we just rerouted to another. So it will end up in the internet or at any time. This is one of the big advantages of our network.
What you can see here is that this is one of our modules. This one is particularly is built in 2011. Every time we build a new one, we just look if we can make any improvements. So, if you notice that there are small differences.
“Cold and hot corridor temperature difference is roughly about 10 degrees. Technically it’s more efficient and raise the temperature that we had, for instance the normal temperature with the data center spot was ?” DP: Is there some kind of official data center requirements in space organization? RB: You can see we had a local installer they just come up with his design. Look at implementations and say 24 containers and if you have four containers proceeding? to each other, it is fixed by 2.20m . This is why he did it because you can transport it. DP: And if there is an expansion would you consider to increase the storage spaces equipments? or increasing space? RB: I am not responsible for the IT. So, the only thing what we do is have always expansion of this kind of models ? It is filled up now. What you can see here is: this is our power module. We have UPS power for our IT source. The first ups is already 10 years old now. No one sees this this is when but it will fail. But thats why we have extra power source and storages.
DP: So what is the temperature you try to maintain? RB: Roughly about 23-26 degrees. Autonomous Territory
63
REVERSE THE CLOUD
Cloud and edge computing Image by openautomationsoftware.com
64
Autonomous Territory
EDGE COMPUTING
The term “cloud computing” has established itself firmly in the vocabulary of most consumers. “Edge computing” is a concept that can be seen as an extension of now ubiquitous cloud computing and Internet of Things (IoT). In the most fundamental terms, edge computing is the movement of intelligence and computation from centralized data servers in a cloud network to hardware on the fringes of a network. Instead of sensors collecting data at a location and sending it back for a centralized server to process, hardware is available locally that will compute that data and then send the results to the cloud where the information will be immediately available and actionable without further processing. Alright, so now that we know what edge and fog computing are, how are they useful in the real world? The future of automated driving relies on data about traffic, obstacles, and dangers being computed in real-time for quick decision making. In the event of a collision, a second delay is more than enough to change the outcome. Fog computing can also be used to analyze and computer data about local traffic, by collecting information from vehicles, processing it, and then sending it to the overall cloud to share. Chantelle Dubois, Computing on the Fringes: How Edge Computing and Fog Computing Are Changing How We Use the IoT
Autonomous Territory
Among the top services that are driving edge compute, video content delivery, which included 360 video and venues, was first at 92% followed by a three-way tie among autonomous vehicles, augmented reality/ virtual reality and industrial internet of things/ automated factory all at 83%. Gaming was next at 75%, with distributed virtualized mobile core and fixed access in another tie with private LTE at 58%. IHS Markit, Video to Drive Demand for Edge Computing Services
P2P NETWORK AND BLOCKCHAIN
Edge computing is dispersing computing power, but it is still a part of a centralized structure. Something more radical is happening. P2P networks that use blockchain to manage cloud storage based on the sharing of excess drive and network capacity on PCs and in data centers. Those who share capacity get free storage – and can be paid in cryptocurrency. The emerging blockchain-based distributed storage market could challenge traditional cloud storage services, such as Amazon AWS and Dropbox, for a cut of the cloud storage market. The idea of using P2P networks to aggregate computer resources is not new. In the early 2000s, BitTorrent opened as a distributed filesharing service and grew to handle more than half of the internet’s file-sharing bandwidth. Lucas Mearian, Blockchain and cryptocurrency may soon underpin cloud storage
65
66
Autonomous Territory
10SQM OF TROPICS
Above: Edgepod website page Below: Micro data center Photos by Edgepod.com
“In the 5G era, Mobile Edge Computing will be a key technology for next-generation industries including realistic media and autonomous driving.� JONG-KWAN PARKE - senior vice president and head of Network Technology R&D Center of SK Telecom
Autonomous Territory
67
68
Autonomous Territory
P2P NETWORK
Decentralized cloud storage brands Images from Internet
“Storj uses blockchain to track digital ‘farmers’ who, similar to Bitcoin miners, have signed on to allow an application to share excess network and storage capacity on their computers or servers.” LUCAS MEARIAN - senior reporter of Computerworld
Autonomous Territory
69
DATA CENTER SECURITY
Julian Assange, co-founder of Wikileaks, was arrested inside the Ecuadorian embassy Photo by Reuters
70
Autonomous Territory
SECURITY
Data centers are complex and to protect them, security components must be considered separately but at the same time follow one holistic security policy. Security can be divided into physical and software security. Physical security encompasses a wide range of processes and strategies used to prevent outside interference. Software or virtual security prevents cybercriminals from entering the network by bypassing the firewall, cracking passwords, or through other loopholes. A data center building’s most obvious security characteristics are related to design and layout. A data center building is usually built away from major roads in order to establish buffer zones made up of a combination of landscaping and crash-proof barriers.
view of a data center’s security posture. And creating secure zones in the network is one way to layer security into the data center. Before applications and code are deployed, certain tools may be used to scan them for vulnerabilities that can be easily exploited, and then provide metrics and remediation capabilities. CYBER EDU: What is Data Center Security? Data Center Security Defined, Explained, and Explored
POLITICS
When data relates to confidential or illegal content, the data center can become very politically sensitive. Some events involving national interests may evolve into information wars between multiple countries. At this time, data center security will face greater challenges.
Access into a data center facility is fairly limited. Most do not have exterior windows and relatively few entry points. Security guards inside the building monitor for suspicious activity using footage from surveillance cameras installed along the outside perimeter. Visitors may use two-factor authentication to enter the building, including scanning personal identity verification (PIV) cards and entering a personal passcode. Employee badge readers and biometric systems, such as fingerprint readers, iris scanners and facial recognition, may also be used to permit entry. Hacking, malware and spyware are the obvious threats to data stored in a data center. A security information and event management tool (SIEM) offers a real-time
Autonomous Territory
71
Perimeter Wall Password& Retina Scanner
Fingerprint Lock
Motion Sensor
Gated Door
Video Surveillance
X-ray Scanner Security Guard
72
Autonomous Territory
QUESTION OF SECURITY
While many think of data centers and consider the risk of cyberattacks, physical security threats must be evaluated as well. Physical breaches at data centers do happen and can result in costly consequences. Image by Arda Ertan Yildiz, Yongyi Wu
“Applying the five D’s of security – deter, detect, deny, delay and defend – should serve as the framework for every data center’s protection plan” JOHN DISTELZWEIG - Vice President and General Manager of FLIR Security Systems
Autonomous Territory
73
74
Autonomous Territory
UNCLEAR BUNKER
Swedish broadband carrier Bahnhof confirms that some Wikileaks servers are now hosted in its Pionen data center, converted from an underground Cold War-era nuclear bunker in downtown Stockholm. The server farm, carved out of a 100-foot-tall granite hill, has a single entrance, protected by 20-inchthick steel doors.If the power goes out, back-up is provided by a couple of German submarine engines. Photos by Bahnhof
STOCKHOLM, SWEDEN Autonomous Territory
75
76
Autonomous Territory
UNCLEAR BUNKER
Drawings of Pionen data center Images by Bahnhof
Would Pionen’s 8,000 servers survive a nuclear blast?
“I’m not sure about the people, but the machines would survive.” JON KARLUNG - chairman of Bahnhof
Autonomous Territory
77
AI AND UNMANNED DATA CENTER
Google Brings Liquid Cooling to Data Centers to Cool Latest AI Chips Photo by Google I/O 2018
78
Autonomous Territory
ARTIFICIAL INTELLIGENCE
Across the globe, Data science and AI are influencing the design and development of the modern data centers. With the surge in the amount of data everyday, traditional data centers will eventually get slow and result in an inefficient output. Utilizing AI in ingenious ways, data center operators can drive efficiencies up and costs down. A fitting example of this is the tier-two automated control system implemented at Google to cool its data centers autonomously. The system makes all the cooling-plant tweaks on its own, continuously, in real-time- thus saving up to 30% of the plant’s energy annually. Every five minutes thousands of sensors throughout the data center issue and send readings to the cloud. Deep neural networks then work to “predict how different combinations of potential actions will affect future energy consumption. AI has enabled data center operators to add more workloads on the same physical silicon architecture. They can aggregate and analyze data quickly and generate productive outputs, which is specifically beneficial to companies that deal with immense amounts of data like hospitals, genomic systems, airports, and media companies. Nowadays, the use of AI in the data center field is more at the device and system level. With implement of more technologies, AI will greatly change the type of data center.
withdrawal of human technicians from remote data centers, which has enabled data centers to more freely choose the better location. Microsoft CEO Satya Nadella says underwater data centers will play a major role in expanding the firm’s global cloud computing platform. Nadella predicts Project Natick, under which Microsoft deployed a 40-foot data center pod on the seafloor off the coast of Scotland, could be repeated across the world. “Since 50 percent of the world’s population lives close to water bodies, we think this is the way we want to think about future data center regions and expansion,” he told the Microsoft Future Decoded conference in London. Nick Heath , Microsoft CEO Nadella: Underwater data centers are the future
There is more to Microsoft’s sinking a data center the size of a shipping container off Scotland’s Orkney Islands than cheap power, free cooling, and being close to where half the world’s population lives. The big hope is that a truly lights-out data center could end up being cheaper to run and failing less often. Mary Branscombe, Why Microsoft Thinks Underwater Data Centers May Cost Less
NEW SURROUNDINGS
The use of automated systems has led to the
Autonomous Territory
79
Van Nelle Distribition Center NL, 1920-2009
200 People 80
Autonomous Territory
200/5
5 people work in the SmartDC in Van Nelle factory nowsdays while around 200 people worked in the same area when there was a distributuion center Images by Arda Ertan Yildiz, Yongyi Wu
Van Nelle Smart DC NL, Today
5 People Autonomous Territory
81
82
Autonomous Territory
UNMANNED FUTURE
News about AI in data center industry Images from Internet
“By 2020, 30 percent of data centers that fail to apply AI and machine learning effectively in support of enterprise business will cease to be operationally and economically viable.� MILIND GOVEKAR, Gartner
Autonomous Territory
83
10
84
2700
2100
1200
Autonomous Territory
POST-HUMAN LANDSCAPE
Fully automated data centers will develop to the most effective form for itself. Image by Arda Ertan Yildiz, Yongyi Wu
800
4000
000
Autonomous Territory
85
86
Autonomous Territory
OCCASIONAL VISITOR
Many features of data center are not suitable for human, such as high running temperature. If we can be detached from there for most of the time, data center can have less limitation from human being. Images from Internet
“... data center managers can save 4 percent in energy costs for every degree of upward change in the set point.� MARK MONROE - director of sustainable computing of Sun Microsystems
Autonomous Territory
87
88
Autonomous Territory
UNDER THE SEA
Project Natick’s 40-foot long Northern Isles datacenter is loaded with 12 racks containing a total of 864 servers and associated cooling system infrastructure. The datacenter was assembled and tested in France and shipped on a flatbed truck to Scotland where it was attached to a ballast-filled triangular base for deployment on the seabed. Photo by Microsoft
ORKNEY ISLANDS, SCOTLAND Autonomous Territory
89
90
Autonomous Territory
OPPOSITE WORLD
Benefits from advanced computer technology, high-speed networking and artificial intelligence, the next generation data center can be close to the most advantageous external factors and develop to the most effective form for itself. Image by Arda Ertan Yildiz, Yongyi Wu
Autonomous Territory
91
PROJECT SUMMARY “... by the year 2020, about 1.7 megabytes of new information will be created every second for every human being on the planet. As the digital transformation of our society expands, data centers play the critical role of providing the technology backbone that supports our digital lifestyle. In fact, data centers are accommodating these needs by occupying almost two billion square feet of facility floor space across the world. “ Steven Carlini, Schneider-Electric
In this global context, we try to ask why and how data center will develop in the future and how it will influence our city and nature in a more profound way in this research. Due to the huge growth in data demand worldwide, the traditional data center model has shown its limitations. But from the trends in the data center typology evolution and geographical distribution we can see a possibility of a completely independent huge system. Benefits from advanced computer technology, high-speed networking and artificial intelligence, the next generation data center can be close to the most advantageous external factors and develop to the most effective form for itself.
• TUDelft • Complex Projects • Arda Ertan Yildiz, Yongyi Wu • 2018-2019 • 92
Autonomous Territory