Design Guidelines for a Monitoring Environment Concerning Distributed Real-Time Systems
While the dependability requirements of distributed real-time systems are expanding, there is currently no framework for defining and mapping these requirements into the system design and operation. A method of controlling and achieving the dependability level is real-time monitoring, which measures the degree of requirements fulfillment, relates it to the pre-defined, measurable system-level expectations and dynamically adapts the system, based on quality metrics, risk analysis, cost evaluation, control theory, neural networks, data acquisition and system knowledge management. The book deduces a framework to reveal, define, quantify, measure, analyze, design, implement, test, monitor and enhance dependability (functional and non-functional) requirements of a distributed system with real-time constraints. It presents how the framework can be applied throughout all life-cycle stages, under varying constraints and with maximized cost effectiveness. An overview of the applicable tools and methodologies is given and an integrated and generalized architecture for ensuring continuous fulfillment of system requirements is proposed. The framework provides a multilevel specification mechanism to establish the preservation of system requirements. This ensures the correct functioning of system through adaptations at run time. Among the benefits are controlled access and coordinated resource sharing in accordance with service-level agreement policies, multi-stakeholder interest preservation, transparency with respect to location, naming, performance etc., achievement of quality of service on demand, decentralization, seamless integration of resources and applications, as well as increased predictability.
Publication Date: 4/1/2004