Distributed Stochastic Geographical Load Balancing over Cloud Networks
Abstract: Contemporary cloud networks are being challenged by the rapid increase of user demands and growing concerns about global warming, due to their substantial energy consumption. This requires future data centers to be both energy efficient and sustainable, which ich calls for leveraging cutting cutting-edge edge features and the flexibility provided by the modern smart grids. To fulfill those goals, this paper puts forward a systematic approach to designing energy energy-aware traffic-efficient efficient geographical load balancing schemes for data-center center networks that are not only optimal, but also computationally efficient and amenable to distributed implementation. Under this comprehensive approach, workload and power balancing schemes are designed jointly across the network, both delay delay-tolerant rant and interactive workloads are accommodated, novel smart smart-grid grid features such as energy storage units are incorporated to cope with renewables, and incentive pricing mechanisms are adopted in the design. To further account for the spatiospatio temporal variation on of demands, energy prices and renewables, the task is formulated as a two-timescale timescale stochastic optimization. Leveraging dual stochastic approximation and the fast iterative shrinkage shrinkage-thresholding thresholding algorithm (FISTA), the proposed optimization is decompose decomposed across time slots (first-stage) stage) and data centers (second-stage). stage). While the resultant online algorithm is strictly feasible and provably optimal under a Markovian assumption for the underlying random processes, extensive numerical tests further demonstrat demonstrate e that it also works well