Abstract
Many cloud center operators are turning to leverage on-site renewable energy to reduce power cost for sustainable development consideration. But how to effectively coordinate the intermittent renewable energy with the workload remains to be a great challenge. This paper investigates the problem of power cost minimization under the constraints of different SLAs of delay-tolerant workload and non-delay workload for green data center in smart grid. To handle the randomness of workload, electricity price and renewable energy availability, we formulate a constrained stochastic problem with the consideration of the affection of zero price in smart grid. Then we propose a low-complexity Online workload Scheduling algorithm with Delay Guarantee (OSDG) which makes online scheduling decision based on the system current state and guarantees a bounded worst scheduling delay for delay-tolerant workload. The rigorous theoretical analysis demonstrates that our algorithm achieves a [0 (1/v), O (V)] cost-delay tradeoff. Extensive simulations based on real-world trace are done to evaluate the performance of our algorithm in reality. The results show that OSDG achieves about 4.9% improvement compared with the baseline algorithms.