Architecture and Mechanisms of the Multi-Change Control Layer (MCCL)

点击翻译按钮获取中文摘要

This project addresses the development of a flexible distributed middleware that can autono-mously manage and assure the safety, security and availability requirements of multiple con-currently changing applications which execute on a changing hardware platform and in a changing environment. Although these applications can be functionally independent, they po-tentially share and access common platform components. Because of this resource contention, this can result in conflicting or contradictory goals. The applications specification and require-ments as well as the platform properties are described and formalized in the form of contracts. Based on the contracts, the middleware must evaluate current as well as possible future sys-tem states and find new system configurations in order to guarantee all requirements.It is therefore the task of the middleware to solve these conflicts, to optimize the overall per-formance of the system and to guarantee a correct execution. For this purpose, a model and an execution domain have been developed that are responsible for the modeling, analysis and run-time monitoring. These domains contain the methods developed by the B projects. The resulting middleware has been combined with an existing run-time environment and serves as a foundation for the C projects. This was the focus of the first project phase.In the second project phase, this middleware shall include increasingly open networks and hardware platforms, which involves the cooperation of multiple networked (sub-)systems. The openness and changeability of the components in this distributed system hence necessitates a fault-tolerant and tamper-resistant coordination of the middleware functions. This project de-fines the corresponding architecture and coordination strategies for such a middleware. The major task is the identification and evaluation of requirements for the diverse functions and the search for solutions how the developed coordination mechanisms can meet these. This not only includes the agreement on a new system configuration but, in particular, also the system state evaluation as well as run-time monitoring of distributed, adaptive applications. This pro-ject develops appropriate methods and mechanisms additionally ensuring that all functional and non-functional requirements of the (distributed) system and its applications are met.