科研之友
微信
新浪微博
Facebook
分享链接
摘要
Autonomic and self-* systems of all kinds are expected to operate in a stable and predictable way. This in a natural way implies the need for pre-deployment design-time verification of the decision making logic, meant by a process of proving that the logic deployed into the autonomic service/software component meets all requirements and will not lead to any unpredicted or unwanted behavior. There is actually no simple way that can be used for verification of the decision making logic used self-* and autonomic systems as only methods that use some kind of formalism can in an objective way verify the logic correctness. There is one problem, however, with this kind of approach - the formalism may vary depending on the way the decision making logic is expressed (for example, it may be in the form of a policy). In this paper, an approach to verification of policy decision making logic is based on its transformation into the corresponding Petri net using the proposed methodology. Consequently, the resulting Petri net can then be verified using Petri nets formalism and these verification results can be used to assess policy correctness. Two policies were used to illustrate the approach: ACSPolicy controlling the basic operation of an Air Conditioning System (ACS) in a car and ACSFullPolicy being its more sophisticated version allowing the controlled ACS to work in three pre-defined modes. The second policy shows how easily the proposed methodology scales up allowing the analysis policies of any complexity level.
- 出版日期2016-10
