[Author’s Name]
[Institution’s Name]
Essay on Component-Based Software Engineering
Component-Based Software Engineering aims at decreasing development time and costs by creating applications from reusable, easily connectable and exchange-able building blocks - the components. The rules that these components have to follow are defined in a component model. Obeying the rules permits safe connection and substitution of components. Current component models like Microsoft's COM (G. Eddon, H. Eddon: Inside Distributed COM. Microsoft Press, 1998) and Sun's Java Beans are based on a uniform execution platform (i.e. the Win32 API and the Java Language and Virtual Machine) and some structuring conventions for components.
These enable the composition of components by matching functional interfaces only. Non-functional issues, e.g. how long it take to run a component and how much memory it consumes, are not addressed by these component models. It is assumed that a component can use nearly unlimited virtual memory, that there are no other timing constraints apart from the user's patience, and that one component's resource usage will not have any influence on other components. Both assumptions, the negligible effects of non-functional properties and the uniform execution platform, do no longer hold in the field of distributed embedded real-time systems. In a real-time application functionally correct results have to be computed in time. A late result is as useless as a wrong result. Moreover it must be possible to predict beforehand if an application will always pro-duce results in time.
Resources that can be used by an embedded application are usually scarce. Available processor time and memory are very limited due to hardware costs. Thus a component can easily influence others simply by consuming too many resources. To make matters worse, there is a variety of different hardware architectures and operating systems on the embedded market which creates a large number of diverse execution platforms. Many embedded applications are also distributed across multiple nodes in a network. ..........
