Multi-agent framework for cloud service composition

Abstract

Cloud service composition is required when single cloud service cannot satisfy all requirements of the user. Cloud service composition presents an efficient managerial, on-demand and scalable way to integrate computational resources (hardware, platform, and software). It includes several tasks such as discovery, selection, and deployment of the services. In this paper, an Agent based approach for cloud service negotiation is proposed. Agent based technology has hailed as a new paradigm for implementing software agents. Software agents are autonomous entities which act on behalf of user across open and distributed environments. A Multi Agent System for cloud service composition makes use of a market-oriented approach which would regulate demand and supply chain. This work involves creating service provider agents and consumer agents. The resources offered by the service providers are created as virtual machines in cloud that would be mapped to the consumers. The consumer agent gets the input from the user and finds the best service providers who satisfy the consumer's constraints. Consumer agent further establishes a negotiation between these service providers and purchases the cloud services for market favorable cost. The consumer and service provider agents communicate using Contract Net Protocol, which is a task-sharing protocol in multi-agent system. The workflow architecture of this framework involves interaction and negotiation of the consumer and service provider agents. Also, a mathematical model has been proposed to support market-driven approach in cloud commerce.