LeraoSoft - Knowledge Distributed

Final Report

ieee-presentation

mimir-presenation-final

ieee-presentation

Sorry i lost the signed version and unsigned versions of our team contract.

sr-project-meetings

Mimir Demo Video

Also as usual all of our documentation can be found at http://compsciguy.homeftp.org/svn/srdesign/Documentation

Here are a few charts of our progress we have made.  In the first chart we show what we had at the beginning of our development in the second chart we show where we are now.  The y axis was how much data was used to rebuild and the x axis is the trial number.

As you can see in the first chart we have a lot of points above the 50% mark meaning that if we lost half the data we would not be able to rebuild.  In the second chart you can see there are none over that mark.  The points tend to be somewhat normally distributed meaning even in the second chart there is a chance that it would take more data to rebuild but we can say with certainty of 99.9998 (from experimental data) that we can rebuild with 50% data loss.

The changes made to the robust soliton distribution are extensions from ones done at Helsinki and at UC Santa Barbara in that we added at least one spike and changed the initial low degree from 1 to 3.  We hope to be publishing more on this later.

As for speed here are the results for 1mb with just local encoding/decoding times are in seconds and throughput is in MB/s.  Also Throughput is calculated by dividing the initial size of 1mb by the speed in both cases.  this means that for each case our network usage would be more since we are creating 4 times the original size in data.