Speaking Engagements & Private Workshops - Get Dean Bubley to present or chair your event

Need an experienced, provocative & influential telecoms keynote speaker, moderator/chair or workshop facilitator?
To see recent presentations, and discuss Dean Bubley's appearance at a specific event, click here

Wednesday, April 16, 2008

Why use QoS when you don't need it?

I've just watched a presentation at the Telco 2.0 event by a network vendor. Like many I've seen before, it asserted that the network would increasingly become "application aware" and behave differently according to what type of service a user is trying to obtain.

I've got some sympathy for this view. If you're a quadplay operator with dedicated home gateways, and customised & locked-down phones and set-top boxes, then you should be able to easily spot, and define QoS parameters for (say) IPTV or traffic destined for a femtocell gateway.

But the concept breaks down when devices are open and smart. The network may well be application-aware. But increasingly, the applications are going to be network-aware. This already happens on PCs (your streaming media player can identify the connection type and speed, and self-configure appropriately), and will extend to mobile handsets as well.

Software will be able to access (perhaps from a particular website, or peer-to-peer) information about specific network conditions policies. Or it may be able to measure and deduce them itself
in near real-time. The application will know what network connections are available to it, and will be able to present the options to the end user. It will know what needs to be realtime, and what can afford to be delayed.

So, for example an application might be able to know that at a given time in a given place, best-effort connections are actually good enough. So it can initiate a VoIP call, or a low-latency game, using a cheap & basic connection. But if quality shows signs of degrading, the app could ask for a higher-QoS connection and upgrade on the fly. Or if it reasonably suspects that the network is deliberately blocking or degrading the service, it can advise the user, or look for an alternative method of connectivity. (Or maybe initiate a realtime auction among multiple service providers for its desired connection).

In the real world, in the short term, there will obviously be a balance. People are lazy and often take what they're given - especially if its free or subsidised. They won't configure software, or want to be too intrusive, even if it's telling them about ways to save time or money. Often, people will accept (or have no option but to accept) locked-down devices. But over time, processing capability moves to the edge of the network as a consequence of Moore's Law.

Anyone who thinks that the network will always be able to identify and outsmart applications is naive. Applications will (in some cases, but not all) become able to play the network at its own game and win - and the incentives for creating these game-players evolve in parallel with the networks' intended control platforms.

No comments: