My photo

Bill St. Arnaud is a R&E Network and Green IT consultant who works with clients on a variety of subjects such as the next generation research and education and Internet networks. He also works with clients to develop practical solutions to reduce GHG emissions such as free broadband and dynamiccharging of eVehicles (See . View my complete profile

Tuesday, May 22, 2007

Why research into the future of the Internet is critical

[I highly recommend taking a look at the YouTube video being distributed by the FTTh council and also listening to radio interview of Jon Crowcroft listed below. The FTTh council video is ideal for politicians and policy makers as it provides a very high level perspective of the Internet and its current challenges. It offers compelling evidence of the coming "exo-flood" of data that will hit the Internet, largely due to the distribution of video. They rightly argue that today's Internet is incapable of supporting this tsunami of data, particularly in the last mile - and that new network architectures and business models are required.

Jon Crowcroft's interview is quite interesting in that he claims that heat loading at data centers will make distribution of video through traditional client server models impossible, and that peer to peer (P2P) will be the only practical way of distributing such content. In testament to that fact there is an explosion of new companies gearing up to deliver video and other content via P2P including Joost, Vudu, etc. This is why many argue that the traditional telco NGN architecture with IPTV is doomed to failure. But P2P imposes its own sort of problems on today's Internet architectures as witnessed by the attempts of many service providers to limit P2P traffic or ban it outright.

If P2P is indeed going to be the major mode of delivery of data, especially for video then we need to explore new Internet architectures. As Van Jacobson has pointed out - too much network research is focused on the Internet as a traditional telecommunications medium of "channels" from A to B. As a result Internet research, especially in the optical and network world is largely about topology optimization, network layers, reliability, redundancy, etc

Universities to my mind should be at the forefront of exploring ways to build and deploy a new Internet - not only for the research community, but for the global community as well. But sadly many universities are also trying to restrict P2P traffic, and in some cases act as the snarling guard dog for the RIAA and MPAA. Students at universities are the early adopters of new technology (much more so than their ageing professors)- rather than discouraging their behaviour -we should see them as an opportunity to test and validate new Internet architectures and services. See my presentation at Net@Edu on some thoughts on this topic

P2P may fundamentally reshape our thinking of network architectures such as enabling the end user to do their own traffic engineering and network optimization to reach the closest P2P node or transit exchange point.

Thanks to Olivier Martin for the pointer to YouTube Video and Dewayne Hendricks for a multitude of other pointers -- BSA]

FTTH Council Video

Jon Crowcroft Interview

A good article on the current challenges of P2P running on today's networks

A very interesting read of why IPTV is doomed to failure:

Ohio university bans P2P

Van Jacobson's talk

How Universities can play a leading role in Next Generation Internet

This Internet TV program is brought to you by ...
Joost, the Internet television service being developed by the
founders of Skype, has lined up several blue-chip advertisers,
including United Airlines, Microsoft, Sony Electronics and Unilever,
as it prepares for its introduction.


NY times article on Vudu and its P2P plans