Lookup NU author(s): Dr Christopher Smith,
Professor Aad van Moorsel
Full text for this publication is not currently held within this repository. Alternative links are provided below where available.
Uncertainty is an inherent property of open, distributed and multi-party systems. The viability of the mutually beneficial relationships which motivate these systems relies on rational decision-making by each constituent party under uncertainty. Service provision in Grid systems is one such relationship. Uncertainty is experienced by the service provider in his ability to deliver a service with selected quality level guarantees due to inherent non-determinism, such as load fluctuations and hardware failures. Statistical estimators utilized to model this non-determinism introduce additional uncertainty through sampling error. Inability of the provider to accurately model and analyze uncertainty in the quality level guarantees can result in the formation of sub-optimal service provision contracts. Emblematic consequences include loss of revenue, under-utilization of resources and erosion of trust and reputation. We propose a utility model for contract-based service provision to provide a systematic approach to optimal service provision contract formation under uncertainty. Monitoring policies to enable the derivation of statistical estimators for quality level are introduced, with analysis of their resultant accuracy and cost.
Author(s): Smith C, van Moorsel A
Publication type: Conference Proceedings (inc. Abstract)
Conference Name: EMAGS 2007. Workshop on Economic Models and Algorithms for Grid Systems
Year of Conference: 2007
Publisher: Institute of Information Systems and Mangement, Universität Karlsruhe
Notes: Also presented at PMCCS-8, the 8th International Workshop on Performability Modeling of Computer and Communication Systems, September 20-21, 2007, Edinburgh, Scotland.