Intelligent virtual machine placement for cost efficiency in geo-distributed cloud systems

Abstract

An important challenge of running large-scale cloud services in a geo-distributed cloud system is to minimize the overall operating cost. The operating cost of such a system includes two major components: electricity cost and wide-area-network (WAN) communication cost. While the WAN communication cost is minimized when all virtual machines (VMs) are placed in one datacenter, the high workload at one location requires extra power for cooling facility and results in worse power usage effectiveness (PUE). In this paper, we develop a model to capture the intrinsic trade-off between electricity and WAN communication costs, and formulate the optimal VM placement problem, which is NP-hard due to its binary and quadratic nature. While exhaustive search is not feasible for large-scale scenarios, heuristics which only minimize one of the two cost terms yield less optimized results. We propose a cost-aware two-phase metaheuristic algorithm, Cut-and-Search, that approximates the best trade-off point between the two cost terms. We evaluate Cutand-Search by simulating it over multiple cloud service patterns. The results show that the operating cost has great potential of improvement via optimal VM placement. Cut-and-Search achieves a highly optimized trade-off point within reasonable computation time, and outperforms random placement by 50%, and the partial-optimizing heuristics by 10-20%.

An on-line algorithm to determine the location of the server in a server migration service

“…The simulation results with the mixed integer programming problem show that the integrated control outperforms the individual control. In [10,11], the algorithms to determine the locations of VMs in geographically distributed data centers have been proposed. The algorithm in [10] successfully reduces the maximum latency between VMs, and the algorithm in [11] successfully reduces both electricity cost of data centers and WAN communication cost.…”

“…Previous work related to SMSes includes live migration of VMs [6], migration of databases (DBs) [7], server migration within a single data center [8], VM migration/placement in geographically distributed data centers [9,10,11], and server location decision in an SMS [4,5].…”

A Virtual Machine Placement Taxonomy

“…Another studied model is a quadratic polynomial function presented in [12] for modeling power consumption according to the workload of each datacenter in a geo-distributed cloud context for electricity cost minimization.…”

Cost-Aware Streaming Workflow Allocation on Geo-Distributed Data Centers

“…One exam ple for VM price d iversity, an m3.large VM w ith 2 virtu al CPUs from Am azon Web Services (AWS) costs USD 0.140 in the US East Coast Region and USD 0.154 in the Eu ropean Union Region, w hile the sam e VM costs USD 0.190 from an Am azon DC in Sou th Am erica, and US 0.196 from their DC in the Asia-Pacific region (Singapore) [13]. As another exam ple of traffic price d iversity, AWS charges inter-DC transfer at USD 0.120-0.200/ GB u nequ ally across geographic regions, USD 0.01/ GB in the sam e region, and ap plies no charge for intra-DC com m u nication [14]. Therefore, it is significant to investigate the p rice d iversity am ong geod istribu ted DCs w hen allocating stream ing w orkflow into geo-d istribu ted DCs w ith the goal of cost m inim ization.…”