I am currently corresponding with a few engineers from Sun about my masters thesis, Dynamic Network Data Compression. But since I have this blog, I thought I'd share my ideas here in search for input, thoughts, criticism etc. to get as big a perspective as possible. So read along and tell me what you think - all feedback is appreciated.


More and more data is transferred across networks. And the recent popularity of virtualization shows that machines are often underutilized. The combination of these two factors leads to the conclusion that an algorithm to compress network data will optimize transfer speeds using already available computing power, hence make no difference to the user.

This thesis focuses on analyzing how to maximize network data transmission speeds using well known compression algorithms by monitoring the available processing power on both the server and client side and the instantaneously available bandwidth between these.


Adding to the above (part of my thesis abstract), I will be implementing this dynamic compression on a project in which I know the file types and sizes. The project is a speech recognition program where I will be performing the speech recognition remotely. Based on the bandwidth between the platforms (e.g. 512 kbit/s) and the available processing power on the platforms (based on system speed and current utilization), the compression which will give fastest possible file transfer speed is determined and the file is sent using these parameters.

My main concern is that there is not much research in this area. This means that I am either working on a very interesting project or I am walking on very thin ice! Which is it?

L