This is my simple attempt to learn a little bit about distributed systems and parrallelization. Each client sorts a chunk of data using mergesort which is the k-merged on the master server. I did some simple testing on one machine seeing the performance difference between one and two clients. The dataset was 100000 random integers. One client took ~.36 secs, while using client took ~.25 secs, so ~.11 secs faster. Since all the programs were running on the same computer, network latency was not a factor in the test results. In general I didn't deal with networking issues in my code as this is a proof of concept.