Towards Autonomic Computing Using Goal-Based Translation Strategy in Heterogeneous Distributed Environments

摘要

This paper presents a new architecture to implement autonomic features in complex heterogeneous environments. This three-layer architecture model eliminates the need for a common shared language and heavy global standards by using semantic negotiations between different layers. For this reason, a collection of high level cognitive agents make specific policy based decisions and an adapting interface layer sends them to the system elements. Using the received decisions and state information, each element tries to achieve the decision (goal) by its capabilities and local knowledge. To provide such local knowledge and semantic based functionality, we focused on dependability and its attributes as the key core of local autonomous and how they can affect the initial goals: self-configuring, self-healing, self-optimizing and self-protecting. Local elements will accomplish the goal by using the received state information and its dependability diagram. Such goal-based translation strategy will provide a better scalability and flexibility for today and future heterogeneous changing distributed environments and interwoven systems.