Coordination-theoretic framework for sharing e-business resources
Market-based models are frequently used in the resource allocation on the computational grid. However, as the size of the grid grows, it becomes more difficult for customers to negotiate directly with all the grid resource providers. Middle agents are introduced to mediate between the providers and the customers so as to facilitate the resource allocation process. The most frequently deployed middle agents are the matchmakers, the brokers, and the market-maker. The matchmaking agent finds possible candidate providers who can satisfy the requirements of the customer, after which the customer directly negotiates with the candidates. The broker agents are mediating the negotiation with the providers in real time. The market-maker acquires resources and resource reservations in large quantities, and resells it to the customers. The purpose of this study was to establish a negotiation framework that enables optimal allocation of resources in a grid computing environment, such that: clients are allowed to have an on - demand access to pool of resources; collaboration is enhanced among resource providers; cost saving and efficiency are ensured in resource allocation. The objectives to realize this purpose were: first, to design an appropriate negotiation model that could be adopted to achieve optimal resource allocation; second, to determine an effective search strategy that could be employed in order to reach a Pareto efficient negotiation solution; third, to adopt a negotiation strategy or tactics that negotiators could use to arrive at optimal resource allocation. In order to achieve the goals and objectives set for the study, the following methodologies were used: (i) a critical survey of the existing economic approach and models for negotiating grid resources was conducted; (ii) The knowledge gained from the literature surveyed was used to construct a novel model called Co-operative Modeler for mediating grid resource sharing negotiation. We used Mathematical notations: first, to construct a theoretical model for allocation of resources to clients* task; second, to present a novel Combinatorial Multi-Objective Optimization model (CoMbO). by modeling negotiation offers of agents as multi-objective optimi2ation problem; and third, to present Genetic and Bayesian Learning Algorithms for implementing the model presented; (iii) an implementation prototype of the Co- operative Modeler was developed by: first, implementing the Co-operative Modeler to mimic the real world negotiation situation; second, using time-dependent negotiation tactics to evaluate the negotiation behavior of the Co-operative agents. The Co-operative Modeler has been shown to guarantee: (i) Scalability of number of users, i.e. multiple users can access a virtualized pool of resources in order to obtain the best possible response time overall by maximizing utilization of the computing resources;(ii) Enhanced collaboration.- that is promoting collaboration so that grid resources can be shared and utilized collectively, efficiently, and effectively to solve computer-intensive problems; (iii) Improved business agility - that is decreasing time to process data and deliver quicker results; and (iv) Cost saving. - i.e. leveraging and exploiting unutilized or under utilized power of all computing resources within a grid environment.
A thesis submitted in fulfillment of the requirements for the degree of Doctor of Philosophy in Computer Science Department of Computer Science, Faculty of Science and Agriculture, University of Zululand, 2008.