1) Traffic Classification in VPN tunnels. In this project, we are trying to identify streamed traffic in a VPN tunnel. Since the available information from individual packet is quite limited in VPN's it is a challenging task. However, using machine learning algorithms and especially, (SVM) we believe we can extract some information from streamed traffic flowing inside a VPN tunnel.
2) Available bandwidth estimation. Although this problem tackled by several researchers, up to our knowledge there is still no real implementation of proposed works. Pathload, Spruce, IGI and many others are well known tools however there is still room for improvements. So our goal is to come up with a idea which can be commercialized. This project is supported by Exinda Pty Ltd.
Nevil Brownlee
Wed, 5 Aug 09 (NZST)