Dynamic Profit Optimization of Composite Web Services with SLAs

Abstract

Abstract-In this paper we investigate sequential decision mechanisms for composite web services. After executing each sub-service within a sequential workflow, decisions are made whether to terminate or continue the execution of the workflow. These decisions are based on observed response times, expected rewards, and typical Service Level Agreement parameters such as costs, penalties, and agreed response-time objectives. We propose a model for the sequential decision-making process within which we explore a couple of decision algorithms. We benchmarked these algorithms against the profit made when executing the workflow without decision-making. We show that algorithm based on backward recursion principle of dynamic programming is optimal with respect to profit. Next, we analyse the structure of erroneous decisions for both algorithms and show that significant profit gains can be obtained by sequential decision making.

Cloud Workload Allocation Approaches for Quality of Service Guarantee and Cybersecurity Risk Management

“…It is a core function which effectively and efficiently allocates and releases computing resource to meet the demand of cloud users while providing them with satisfactory QoS [54]. Efficient resource management allows cloud service providers to share cloud resources among cloud services and users with high availability and optimized utilization [20][21][22]. It is, yet, a very complicated task for cloud service providers to provide all the required resources to avoid SLA violations due to the limited capacities of available resources.…”

“…Cloud computing supplies a commonly shared pool of on-demand, scalable, and elastic IT resources and services using a pay-per-use pricing model [18,19]. In the meantime, the constant increase in desire for affordable cloud services with stringent service qualities and high-security levels has posed critical challenges on cloud service providers in finding efficient resource allocation policies capable of optimizing the trade-off among energy consumption, operating expense, quality of service, and cybersecurity risks [20,21]. One standard approach that cloud providers adopt to maximize resource usage is Virtual Machine (VM) multiplexing in which they pack several VMs on the same server.…”

“…One standard approach that cloud providers adopt to maximize resource usage is Virtual Machine (VM) multiplexing in which they pack several VMs on the same server. Allowing multiple applications and VMs, with different resource and cybersecurity requirements, to share the same cloud infrastructure components, notwithstanding, results in performance degradation of cloud services and interdependent cybersecurity IC risks [20][21][22][23]. Cloud providers henceforth need efficient resource allocation policies with multi-objective goals in order to thrive in such an everexpanding and competing cloud computing market [24].…”

Delay-Sensitive Service Request Scheduling for Cloud Computing

“…• pdf : This is a method that applies the request processing time probability density function (PDF) on each server to find out how likely that the server's follow-up processing can meet the request's end-to-end deadline [136]. We extend their method to a global sub deadline assignment method.…”

“…Traditionally, people using M/M/1 queue [102] to represent the request processing procedure [136], as shown in Figure 6.2. Service requests arrive with a rate λ, wait in a queue with an infinite size, and get processed with a rate µ.…”