Stochastic Service Provisioning in converged Optical Data Center networks and smart grids

The stochastic nature and power fluctuations of renewable energy sources are two major parameters affecting the stability of the power grid. Traditionally, this issue has been addressed through energy storage elements i.e. batteries, having the disadvantages of increased cost and limited operational lifetime. In this paper, it is argued that a cost effective solution to this problem is to orchestrate the operation of Optical Data Center networks and Power Grids. To this end, a novel formulation based on Stochastic Linear Programming that considers jointly the Stochastic Service Provisioning Problem in the Optical Data Center networks and the Continuation Power Flow problem in the Power Grids, is presented. The proposed scheme takes into account both the time variability and uncertainty of cloud services as well the stochastic nature of renewable energy sources. Our modeling results illustrate interesting trade-offs between the stability of the smart grid, the power consumption of the converged optical datacenter networks as well as the utilization of the infrastructure resources.