||Trust-based accounting in grid systems
||Due, Peter Birkestrøm
||Sharp, Robin (Department of Informatics and Mathematical Modeling, Technical University of Denmark, DTU, DK-2800 Kgs. Lyngby, Denmark)
||Technical University of Denmark, DTU, DK-2800 Kgs. Lyngby, Denmark
||Distributed systems which offer access to computing resources via the use of the Grid paradigm, it is important to be able to find a suitable ''price'' which is to be paid for the use of these resources.
In traditional centralised accounting systems in computers, the price is simply a function of the number of resources used, the length of time for which they are occupied and possibly the urgency of the computing task must be completed. In turn this means that the price will depend on the load on the system and the Quality of Service desired by the user.
In Grid systems, where previously unknown users may start to use the system, the situation is slightly more complicated, because it may not be possible to trust all the users to fulfill their part of the contract, i.e. to keep within the limits of resource usage which they have asked to be allocated. Similarly, a user may not necessarily trust a provider to provide correct results or to supply the agreed resources. A just price for resource usage may therefore depend on the risk which the user and the resource provider are willing to run, and to the degree the provider and user trust one another.
In this M.Sc. project, the aim was to develop a system for trust management which can be used to evaluate a reasonable price for use of computing resources, with due consideration for the degree to the trust relationship between the site and user and vice versa.
The fields studied was the Grid architecture for Computational Economy, Beta trust reputation and revenue management. It lead to a model formulation for price calculation based on a trust value and lead to a heuristic for how a site and user should negotiate with each other. An simulation program was implemented to investigate the price calculation and to find the optimum resource allocation from an economical point of view.
||Department of Informatics and Mathematical Modeling, Technical University of Denmark, DTU : DK-2800 Kgs. Lyngby, Denmark
||Grid systems; Economical Grid; reputational trust; revenue management; optimum capacity allocation; price calculation.
Creation date: 2006-10-06
Update date: 2012-12-18