Dynamic QoS Management in Distributed Real-time Embedded Systems
Citation: Joseph Loyall and Richard Schantz. Dynamic QoS Management in Distributed Real-time Embedded Systems. In Handbook of Real-Time and Embedded Systems, Insup Lee, Joe Leung, Sang Son (Editors), CRC Press, 2008. ISBN: 1-58488-678-1 and 978-1-58488-678-5.
Formats: PDF
Abstract
Increasingly, embedded systems are part of larger distributed real-time embedded (DRE) systems in a wide variety of domains, including military command and control (C2), avionics and air traffic control, and medicine and emergency response. DRE systems combine the stringent quality of service (QoS) requirements of traditional closed embedded systems with the challenges of the dynamic conditions associated with being widely distributed across an often volatile network environment. Traditionally, embedded systems have been able to rely on their closed environments and self-contained bus architectures to limit the dynamic inputs possible and could rely on static resource management techniques to provide the QoS and reliable performance they need. The environment of distributed, networked systems is more open with heterogeneous platforms, where inputs can come from external devices and platforms, and dynamic, in which conditions, resource availability, and interactions can change. Because of this, achieving the necessary predictable real-time behavior in these DRE systems relies on the ability to manage resources end-to-end, map system level requirements to platform level controls, aggregate conflicting requirements, and adapt and reconfigure to changing conditions. Middleware, such as CORBA, is being applied to these types of applications because of its ability to abstract issues of distribution, heterogeneity, and programming language from the design of systems. CORBA, specifically, has spearheaded this trend because of its development of standards supporting the needs of DRE systems, such as RTCORBA, FT-CORBA, and Minimum CORBA. This chapter describes research efforts to develop middleware for providing dynamic, adaptive QoS management in DRE systems. We describe some of the issues in providing a middleware platform for QoS adaptive systems, the middleware solutions we have developed as part of our research, and case studies applying them to the DRE system context.
