分布式实时系统的自适应资源管理中间件

Adaptive Resource Management Middleware in Distributed Real-Time Systems

  • 摘要: 提出了一种基于实时CORBA的自适应资源管理中间件服务及基于最优收益的QoS自适应机制,向开发人员提供了一套面向实时任务的QoS编程接口,并通过资源自感知和QoS自适应使应用级资源分配独立于底层平台。自适应资源管理中间件降低了分布式实时系统中应用程序开发成本和复杂度,并使系统在不确定环境下能动态优化资源分配以适应外界变化。

     

    Abstract: The platform-specific code in the distributed real-time systems (DRTS) is a main contributer to the complexity and cost of application software development. It reduces the maintainability and transplantability, but is not capable of ensuring the performance of real-time tasks. In this paper, an adaptive resource management middleware (ARMM) service based on the real-time CORBA and a QoS (quality of service) adaptation mechanism based on the optimal reward are proposed. A set of QoS APIs that is orient to real-time tasks is provided. The user-level resource management and software development are separated from the hardware platform by resource self-profiling and QoS adaptation. Thus the cost and complexity of application software development of DRTS is reduced, and the resource allocation under unpredictable environment can be dynamically optimized according to the outside variation.

     

/

返回文章
返回