iLAND: An Enhanced Middleware for Real-Time Reconfiguration of Service Oriented Distributed Real-Time Systems

Future networked embedded systems will be complex deployments deeply integrated in the environment that they monitor. They will have to react to both user and environmental events, and this may require modifying their structure to handle the changing situations. In a number of domains including industrial environments, this modification of the system structure will need to take place in real time. This is a hard problem that will require novel and paradigmatic solutions involving cross-domain knowledge to build a middleware for enabling real-time interaction between nodes allowing time-bounded reconfiguration. For supporting real-time, this middleware must be vertically architected in a modular way from the network and operating system levels up to the application software and the real-time policies for achieving time-bounded behavior. This paper presents a middleware that addresses these characteristics supporting timely reconfiguration in distributed real-time systems based on services. Experimental results are shown to validate the middleware in an actual small-scale video prototype.

I. Introduction

Current computing environments are becoming complex infrastructures containing numerous collections of nodes of heterogeneous nature forming a cloud of processing points. These nodes sense, process, actuate, and generate information for human operators and/or for other nodes and subsystems. Information captured in different parts of the complex system can be processed and delivered in real time to other remote spots to further process it and generate the required results on time. This is, for instance, the case of a complex video surveillance system with content analysis for monitoring of industrial processes in hostile environments where no human operators are present. In the near future, these infrastructures are expected to perform cross-processing of huge bulks of information in real time.