A unique storage problem occurred in the College of Science at RIT. A need for additional centralized storage was needed, but there were no additional resources allocated for centralized infrastructure. However, there was an abundance of free disk space on the lab computers that existed within the college. After calculating the total amount of hard disk space for each lab, it was found that there was approximately 36 terabytes of unused disk space per lab. The challenge was how does one utilize this space, yet still maintain reliability and availability of files? This thesis describes the development of a proof of concept system which utilizes free storage space on lab computers within the college. The system automatically distributes a file to enough computers so that it may be retrieved even in the case of computer failures. It also is able to periodically determine whether or not the file is needed on additional computers within the system in order to maintain availability of the file. The proof of concept system was implemented and tested to demonstrate the effect on individual clients within the system. Results have shown that while it has little effect on CPU and RAM utilization, it does have the ability to utilize all available network bandwidth.

Library of Congress Subject Headings

Peer-to-peer architecture (Computer networks); Computational grids (Computer systems); File organization (Computer science)

Publication Date


Document Type


Student Type


Degree Name

Networking and System Administration (MS)

Department, Program, or Center

Computer Science (GCCIS)


Barido, George


Note: imported from RIT’s Digital Media Library running on DSpace to RIT Scholar Works in December 2013.


RIT – Main Campus