Cost Optimization Algorithms for Hot and Cool Tiers Cloud Storage Services

Abstract

In this paper, we consider the data placement problem in the new generation tiered cloud storage services offering hot and cool tiers that are characterized by differentiated Quality of Service (i.e, access latency, availability and throughput) and the corresponding storage and access costs. Given a sequence of read and write requests for an object, we propose an optimal off-line dynamic programming based algorithm to determine the optimal placement of an object in the hot or cool tier and the potential transfer of the object between the tiers in order to minimize the monetary cost comprised of storage and access costs. Additionally, we propose two practical online object placement algorithms that assume no knowledge of future data access. The first online cost optimization algorithm uses no replication (NR) and initially places the object in the hot tier then based on read/write access pattern it may decide to move it to the cool tier to optimize the storage service cost. The second algorithm with replication (WR) initially places the object in the cool tier then it replicates it in the hot tier upon receiving read/write requests. Using a real Twitter workload and a 2-tier storage service pricing, the experimental evaluation shows that the proposed algorithms yield significant cost savings compared to storing data in the hot tier all the time.