Distributed Real-Time Systems
In Chapter 1, the enabling concepts behind real-time systems were introduced. In this chapter, the concepts that extend a real-time system to a distributed real-time system (DRTS) are introduced.
Classically, a DRTS is a collection of several real-time subsystems interconnected using a shared data network. Each subsystem, in the simplest case, could be a processing element comprising a processor and memory. The entire DRTS may be viewed as a loosely coupled distributed system with definite response time, that is, real-time functionality. Architectural issues of a DRTS mostly focus on the interconnection of the constituent subsystems and also the ways the functionalities of a DRTS are governed. First, the interconnection topologies are visited.
Topologies of Interconnection
A topology defines how the components of a DRTS are connected. The interconnection may be physical or logical—two nodes may be physically connected but they may never communicate. The basic physical topologies widely used in a DRTS are bus, star, and a ring as shown in Figure 2.1.
The bus topology consists of a single network cable on which the individual nodes are connected by shorter cables. This is most suitable for systems where the subsystems are arranged along a line; a typical example could be an assembly line. A star topology, on the other hand, is the most common in DRTS applications. The switch acts as an interconnection providing a path between any two nodes. Logically, this is a strongly connected graph with an edge existing between any two nodes. On the contrary, a ring topology provides physical segmentation of the network.