On Arbitrating the Power-Performance Tradeoff in SaaS Clouds

Abstract

In this paper, we present an analytical framework for characterizing and optimizing the power-performance tradeoff in Software-as-a-Service (SaaS) cloud platforms. Our objectives are two-folded: 1) We maximize the operating revenue when serving heterogeneous SaaS applications with unpredictable user requests. 2) We minimize the power consumption when processing the user requests. To achieve these objectives, we construct a unified profit-maximizing objective to jointly consider revenue and cost in an economic view. An offline solution to maximize the supreme bound of the objective is first developed, to 1) justify the validity of our theoretical model, and 2) establish a benchmark to examine the effectiveness of other control solutions. As a highlight of our contributions, we take advantage of the Lyapunov optimization techniques to design and analyze an optimal yet practical control framework, which makes online decisions on request admission control, routing, and virtual machine (VMs) scheduling. Our control framework can accommodate a variety of design choices and operational requirements in a datacenter. Specifically, buffering facilities can be introduced to alleviate the bursty admitted requests and to improve the robustness of the system, and a power budget can be enforced to improve the datacenter performance (dollar) per watt. Our mathematical analyses and simulations have demonstrated both the optimality (in terms of the cost-effective power-performance tradeoff) and stability (in terms of robustness and adaptivity to time-varying and bursty user requests) achieved by our proposed control framework.

Business-Driven Long-Term Capacity Planning for SaaS Applications

“…In order to deal with this problem, business-driven IT management solutions [1] [26] when planning IT infrastructures. As more high level models, business models enable the merge of operational metrics like power consumption and business metrics [26].…”

Distributed Online Hybrid Cloud Management for Profit-Driven Multimedia Cloud Computing

“…1, in which multimedia service requests are processed in the DCs and the proxies handle the access control and routing of the requests. In this work, we assume that the hybrid cloud operates as a discrete-time system [29] and the service provisioning scheme in it can be changed every time interval . Hence, we can normalize system time with the time slot (TS) and obtain the normalized system time as .…”

“…Since the TS in the system is usually much longer than the duration of the switching transition [29], we can ignore the transition overhead in following analysis. Similarly, the VMs in the public DCs can also be released dynamically for cost saving.…”

“…The cost includes three parts, i.e., the energy cost of private DCs, the payment for renting the resources from public DCs, and the communication cost for data transfers. 1 The energy consumption in a private DC in TS is calculated as where and are the energy consumption per server when it is in working and idle state for a TS, respectively, is the power usage efficiency (PUE) of each DC, which is defined as the ratio of its total power consumption to that of the servers [29]. Then, the energy cost of the private DCs is (12) where is the electricity price for DC .…”

Toward online profit-driven scheduling of inter-DC data-transfers for cloud applications

“…}. Here, the time interval Δt can range from a few seconds to several minutes [11], and to adapt to the network dynamics, it should be sufficiently long for the operator to update the traffic scheduling scheme. Note that in the following analysis, we normalize the TS' with Δt and get TS' as t ∈ {1, 2, .…”