RapidIO: The Next Generation Communication Fabric For Embedded Application

1.8. AN ANALOGY

The basic problem that RapidIO is designed to solve is the movement of data and control information between semiconductor devices within a relatively complex system. The RapidIO architecture is targeted at applications where there will be tens to hundreds, perhaps thousands of devices, all communicating with one another within a single system. It is optimized for this size of connectivity problem. It also assumes that efficiency is important. By this we mean that the overhead introduced by using RapidIO should be minimal. There are two dimensions of overhead. There are the extra bits of control information that accompany the data that is being sent across the interconnect. This includes the destination address of the data, the type of transaction that is intended, the error checking code that is used to validate that the data was received correctly and other important pieces of information. Overhead also includes the amount of work that needs to be done to send the data across the interconnect. The RapidIO controller might do this work, but the software that interacts with the RapidIO controller might also do it. In a system that includes hundreds or thousands of communication links, efficiency becomes an important consideration.

We have discussed RapidIO as a technology that replaces buses. We have also mentioned that Ethernet, a local area network (WAN) technology, may also be sometimes used as a bus replacement technology. The following analogy helps to understand ...