Cost-aware Dynamic Data Center Federation Formation for Short-term Service Requests
Niharika Nayak, Rakesh Tripathi, Govind P. Gupta · 2020 Fourth World Conference on Smart Trends in Systems, Security and Sustainability (WorldS4) · 2020
In current environment, many web services map to the distributed data centers, to respond to variable and increasing workloads. It is not cost efficient for one data center to increase number of servers in order to fulfill all requests coming to it. For this reason, federated data center is a key solution for sharing unused resources and gain some profit. In this paper, we have used demand multiplexing and exploits the spatio-temporal variation of electricity price of distributed data center across locations and also consider the power consumption cost of servers at data center. We develop an Integer Linear Programming Problem (ILPP) for Profit Maximization Problem of federated data centers and proposes Greedy algorithm for computing an optimal federation formation of data centers. The performance of the proposed Greedy federation formation algorithm (G-FFA) compares with two other algorithms which are Optimal Federation Formation Algorithm (OFFA) and Random Federation Formation Algorithm (RFFA). Experiments are performed on real world traffic demands at data center in which the proposed (G-FFA) algorithm gives an optimize solution to the optimization problem.