Chapter 6 - Conclusions

6.1 Review of Project Aims

In review, here are the aims of the project:

"Produce a powerful yet small master, controlling computer, and slaves, participating computers, to perform the work normally done by one computer." The project proves that it can do work across many computers but its still in its very initial stages. There is much room to grow as discussed in the chapter on Evaluation & Improvements.

"Incorporate existing technologies, like CORBA and distributed file systems, thereby maximum software reuse." A significant portion of this project uses a lot of existing technologies. ORBacus provides a CORBA and CODA provides a distributed file system are a few examples of software reuse. The time for development of this project were significantly reduced due to their use. The table below describes the advantages and disadvantages of using ORBacus versus ACE+TAO.

CORBA Library	Advantages / Disadvantages
ORBacus	+ Support was quick and informative via mailing lists
	+ Configuration and Compiling was easy.
	+ Documentation easy to find and read
	- Only supports basic set of services. (4)
	- Name Service does not support mutlicasting
	- Pricing model prohibits using project for commercial purposes (even testing) without license.
ACE+TAO	+ Supports a large set of services (11)
	+ Name Service supports multicasting
	+ Support was quick and informative via mailing lists
	+ Free
	- Configuration and Compiling were difficult.

"Perform test on simple projects to confirm this proof of concept actually works." The project does complete this goal but not to the level desired. A simple driver was compiled but it could not be time due to the way in which the master and slaves are started. So in light of this there is no real solid evidence beyond seeing the system in action that this goal is completed.

6.2 Final Thoughts

After discussing this with various people many thought that is could be used. It could fit well as a middleware application existing in a server farm. At this point middleware applications are becoming a more desirable feature of a distributed system. Many company are looking for ways to cut costs when installing new systems and are looking to terminal servers. So as client systems contain less processing power with their file systems distributed across the network, more applications will rely on specialty servers. These specialty servers will bridge the various network services with the client terminal station. DGCC will be one of those applications providing a powerful, reliable and fault tolerant system for developers of the future.