DMDR stands for Distributed Modular Data Repository
The goal of this project is to create a way to efficiently and securely use the space that so many large companies (and other local networks) waste on client machines (smart-dumb terminals).
An example: Company A orders 1000 new machines in today's computer market to replace some old dumb-terminals that their departments have been using for the past 20 years. When ordering, they try to cut corners to save cost, but find that they end up buying massive amounts of 80GB harddrives where they only need 10GB of space per client machine. Imagine being able to use the residual 70GB (a total of 70TB) for backing up corporate data, serving shared files, and creating a shorter path between transmissions when clients request files (less network hops)!
This is where DMDR would come in. The purpose would be to create a 1st or 2nd tier file system on each machine which could be indexed, secured, and chunk-distributed by a server which keeps records.
We already know there are Distributed File Systems available which would do most of the functionality of the above. However, the purpose of this project would be to master the field of chunk-dispersed data structures, instead of network mirrored file systems such as Coda and MS DFS.
The work involved in a project like this would require research in the following areas:
Currently, related research papers and software can be found at the following sites: