Metrics and techniques for quantifying performance isolation in cloud environments

Quality of service approaches in cloud computing: A systematic mapping study

“…Table 7 summarises the articles on this topic, and the relevant subtopics are briefly summarised in the following sub-sections. [36], [37], -Validation research - [34], [38], [39], [40], [41], [35], [42], [43], [44], [45] [34], [46], [47] [48], [49] [44]…”

Robust partial-load experiments with Showstopper

“…Typically, the experiment would also rely on available benchmarks as examples of realistic workloads [16][17][18][19][20]. Many benchmarks, however, are designed to execute in isolation, with all system resources at their disposal.…”

“…As an alternative to simulation, many results on workload consolidation are evaluated using experimental measurements [3,4,8,[16][17][18][19][20]30,56,[64][65][66][67][68]. We should first observe that in these results, information about certain platform configuration aspectssuch as how the frequency scaling mechanism was configured -is often not reported.…”

“…15 Additionally, more than half of the studies only used full utilization workloads (8 positive, 6 negative). Only one of the studies [19,68] might have used the full combination of hardware threads, frequency scaling, frequency boosting and partial utilization workloads together. In this context, our work highlights the potential impact when advanced hardware mechanisms are used with partial utilization workloads.…”

Performance modeling and analysis of virtualized multi-tier applications under dynamic workloads

“…Numerous studies thus far have linked resource allocation with interference of VMs in virtualized environments when applications are migrated from traditional servers to virtualized servers (Urgaonkar et al, 2008b;Krebs et al, 2014;Salehi et al, 2014). It is clear that competing applications face degradation during possession of shared resources in virtualized environments.…”