Optimizing the cost of DBaaS object placement in hybrid storage systems

Abstract

In a Cloud context, Solid State Drive (SSD) became a must-have technology. This technology is too expensive to replace Hard Disk Drive (HDD), both are used in Hybrid Storage Systems (HSS). When it comes to storing data, placement strategies are employed to find the best storage class to use (SSD or HDD). While for many applications, those strategies need to be I/O performance driven, in a Cloud context, they must be cost driven: minimize the cost of data placement while satisfying Service Level Objectives. This paper presents two Cost based Object Placement Strategies (COPS) for DBaaS objects in HSS: a genetic algorithm based approach (G-COPS) and an ad-hoc heuristic approach (H-COPS) based on incremental optimizations. Both algorithms were tested for small and large instance problems. While G-COPS proved to be closer to the optimal solution in case of small instance problems, H-COPS showed a better scalability as it approached the exact solution even for large instance problems (by 10% on average). Both performed better than state-of-the-art solutions as they reduced the overall cost by more than 40%. In addition, H-COPS showed small execution times which makes it a good candidate for a runtime use. Moreover, H-COPS drastically limits the over-provisioning of resources.