An energy and deadline aware resource provisioning, scheduling and optimization framework for cloud systems

Abstract

Abstract-Cloud computing has attracted significant attention due to the increasing demand for low-cost, high performance, and energy-efficient computing. In this large-scale, heterogeneous, multi-user environment of a cloud system, profit maximization for the cloud service provider (CSP) is a key objective. In this paper, the problem of global optimization of the cloud system operation (in the sense of lowering operation costs by maximizing energy efficiency, while satisfying user deadlines defined in the Service Level Agreements) is addressed from the perspective of the CSP.The modeling of the workload dictates viable approaches toward cloud operation optimization. Of the two current models: independent batch requests and task graphs with dependencies, we adopt the later. This fine-grained treatment of workloads provides many opportunities for energy and performance optimizations, thus enabling the CSP to meet user deadlines at lower operation costs. However, these optimizations require additional efforts in terms of resource provisioning, virtual machine placement, and task scheduling. Such issues are addressed in a holistic fashion in the proposed framework.In this cloud environment, users can construct their own services and applications based on the available set of virtual machines, but are relieved from the burden of resource provisioning and task scheduling. The CSP will then capitalize on the data parallelisms in each user workload, effectively manage the collective user requests, and apply custom optimizations to create a global energy cost and deadline-aware cloud platform.

A Game Theoretic Framework of SLA-based Resource Allocation for Competitive Cloud Service Providers

“…(3), there is a most desirable utilization level ߶ to optimize the ܲ /߶ value, which is unit power consumption level per allocated resource. The most desirable utilization level is typically around 70% in various references [24] [25].…”

A simulation as a service cloud middleware

“…However, in a scenario where real-time decisions have to be made based on running a large number of multiple, short-duration simulations in parallel, the considerable setup and tear down overhead imposed by VMs, as demonstrated in Section 5.2, is unacceptable. Likewise, a solution based on maintaining a VM pool that is used by many cloud resource management frameworks such as [12,18,24,44] is not suitable either since it can lead to resource wastage and may not be able to cater to sudden increases in service demand. Thus, a lightweight solution is desired.…”

Value and Energy Aware Adaptive Resource Allocation of Soft Real-Time Jobs on Many-Core HPC Data Centers

“…embedded systems [14], can be employed, but they do not perform optimization for value. Further, energy can also be optimized by performing energy-aware resource allocation [23]. The approaches considering DVFS and optimizing both the value and energy consumption are recently reported [5], [11], but they do not perform adaptive resource allocation as job execution status and available system cores are not monitored.…”

“…In case of both the events 1) and 2) or any of them, the algorithm tries to perform resource allocation for the queues job(s) having non-zero values (lines 18-28). However, if event 3) is also detected at the same time, the adaptation is tried first (lines 7-17) followed by the allocation (lines [18][19][20][21][22][23][24][25][26][27][28]. This ensures that the executing jobs are given priority over the queued jobs to be allocated so that value and energy of the executing jobs can be further optimized before doing optimizations for the queued jobs.…”