An architecture for a cached policy-based decision making systems

Abstract

Nowadays, majority of systems that involve some kind of decision-making process use policy-based computing where decisions are being made as a result of a policy evaluation. Policies are usually formed by system experts who describe in a specified Policy Definition Language how the system should respond to a change of its context information. However, depending on the policy complexity needed to make a decision may become an issue, especially when embedded or real-time control systems are concerned. In such systems certain level of policy complexity may result in ability of the control system to return a valid decision before a deadline. Hence any architectural solution that would resolve or at least minimise the issue would be more than welcome. In this paper we propose an architecture for a policy-based decision making system in application to an embedded control systems. In our solution add to a typical policy-based system an intermediate component — a cache memory which ultimately leads to increasing of the whole system performance. A number of scenario-based performance tests show potential performance gains.