ABSTRACT
A distributed computation involves a number of processes communicating with one another. We observed in Chapter 2 that interprocess communication mechanism is fairly complex. If we want to develop algorithms or build applications for a distributed system, then the details of interprocess communication can be quite overwhelming. In general, there are many dimensions of variability in distributed systems. These include network topology, interprocess communication mechanisms, failure classes, security mechanisms. Models are simple abstractions that help understand the variability — abstractions that preserve the essential features, but hide the implementation details from observers who view the system at a higher level. Obviously there can be many different models covering many different aspects. For example, a network of static topology does not allow the deletion or addition of nodes or links, but a dynamic topology allows such changes. Depending on how models are implemented in the real world, results derived from an abstract model can be applied to a wide range of platforms. Thus, a routing algorithm developed on an abstract model can be applied to both ad-hoc wireless LAN and sensor network without much extra work. A model is acceptable, as long as its features or specifications can be implemented, and these adequately reflect the events in the real world.
