It's an interesting game the ISP business... Or in AARNet's case the I2SP business or the A3SP business depending on whether you're looking at the AU or the US perspective - next generation networks regardless.
Certainly something that has come to light in the roundup of the roadshows (and congratulations to Steve T from Adelaide by the way on winning the iPod!) is that one of the core objectives of our video strategy is to remain as endpoint agnostic as possible, to make sure that the core applications and services we're delivering work across as many people's preferred tools and techniques (software and hardware) to help them get their job done.
This has come to light many times in the recent weeks. I've been pushing heavily the open source philosophies and toolkits that are available for software video conferencing, because to be frank, they're the ones that are really doing the most interesting work in terms of advancing our goals above.
Some of the toolsets we've been espousing the virtues of still have a long way to come, but they show what is possible in terms of pushing the envelope just that little bit further. Just a few examples:
I've been investigating what we can do about
Skype, not in terms of replacing it's P2P abilities in the video space, but for integrating it into the 'ubiquitous video' vision for the
National Video Conferencing Service. However upon reading the Adium blog it details some of the background why this is a little harder than it appears:
What about support for Skype? Unfortunately
this is something we are completely unable to do at this point in time.
Skype is also a closed protocol. If it were an easy thing to break,
other apps would have supported it long ago. This doesn't even take
into account that there are also some VERY serious legal issues in
regards to this.
The reason that Skype has gained such acceptance is quite a simple one - IT JUST WORKS. There's no need to know about firewalls or the various types of
network address translators in order for a call configuration to be successful. Skype achieves this transparency by utilising encrypted peer to peer connections which automatically detect and adjust for these variables. Plus it's multi-platform, plus it's got video, and very shortly will feature "high-def" video, 640x480 to be precise, which is the best full motion you're going to get from the majority of today's webcams.
The downside of this of course is that network administrators cannot control the traffic which may be generated by a single, well connected node. These nodes are referred to as supernodes and can ramp up to consume almost all available bandwidth whilst the connection is otherwise idle. This is obviously
a bad thing (TM) for well connected research networks, or at least their customers who are paying the commodity traffic bills.
Skype have heard the call and offer a webpage on configuring networks to assist in
controlling the creation of supernodes. Unfortunately most of these (except for maybe implementing a transparent HTTP proxy) require intervention from the end user to configure, and certainly we would expect that a company that relies on supernodes for the continued quality of service of their product would hope that not too many people intervene!
In the meantime, the action is afoot in regards to making H.323 work almost as simply, without the need for the supernode hierarchy. The ITU's H.460 is such a standard, which strangely enough is being implemented into the fast-moving open source products first. We'll be watching developments in this field with interest, and see the future of collaboration as a much easier place.
Watch this space for updates in the 'appliance' world of high definition conferencing as well as details on the new multipoint booking system released to the roadshow over the last couple of months. And don't forget Megaconference is on Thursday and Friday next week.