Emerging Patterns in Adaptive, Distributed Real-Time, Embedded Middleware

We have been developing Quality Objects (QuO), an adaptive middleware framework that supports the development of distributed object applications that can measure, control, and adapt to quality of service (QoS) needs and variations in a system. We have applied QuO to the problems of distributed real-time embedded (DRE) applications, including a streaming video dissemination application and a dynamic replanning avionics application, both described in [3]. During the course of developing QuO and of applying it to DRE problems, we have both used many documented patterns and identified new patterns for network centric computing. Some of these are variations on existing, documented patterns, while others appear to be newly discovered. We have seen these patterns appear over and over in the development of our adaptive, QoS-aware systems, which suggests strongly that they are indeed new and reusable patterns. Time and additional investigation will determine whether these are also lasting and universally useful patterns.