Mitigating Provider Uncertainty in Service Provision Contracts

Abstract

Uncertainty is an inherent property of open, distributed and multi-party systems. The economic viability of mutually beneficial relationships between the constituent parties in these systems relies on the ability of each party to effectively quantify and reason over uncertainty in order to facilitate rational decision-making. Service provision in Grid systems is one such relationship, in which uncertainty is experienced by the service provider in his ability to deliver a given quality level due to inherent behavioural factors, such as load fluctuations and equipment failures, and due to statistical factors relating to the use of past empirical data for future prediction. Inability of the provider to effectively quantify and reason over these behavioural and statistical uncertainties can result in errors in the estimation of quality levels consistent with business objectives. Emblematic consequences of such errors include loss of revenue, inefficient resource usage and erosion of consumer trust. To address this, we propose a utility model for contract-based service provision which extends common economic utility models and facilitates explicit reasoning over the uncertainties in quality levels. We couple this model with a monitoring policy which enables the mitigation of statistical uncertainty in quality levels under the constraint of costs for information acquisition.