Friday, February 01, 2002

Kevin: Great point. I agree with the statements on these assumptions, but it kind of reminded me of some time I'd spent in a working group looking at Internet2 possibilities. There was a lot of general agreement that "the telcos" couldn't see the benefits of I2, just as it took them years to realise the value of the net the first time around. But my response was to remind them that the telcos are just a business like any other. Where they can see income they'll invest a proportionate amount, but where they can't see it they won't make the investment. There's nothing mysterious or particularly backward about this, other than perhaps their lack of vision.

But given the sharp decline in the whole telco industy over the last 18 months it is a brave person at any telco that asks their bosses to punt on something as (to them) speculative as I2.

OTOH, I'm not sure that I agree that telco engineers are deliberately assuming that point source broadcasts will drive the networks of the future. I've recently been working in this very field, in fact my last task was to design a complete telco architecture for our national carrier, and our reigning assumption was that we would never be able to predict where the traffic would be originating nor where it would be terminating (except in very gross terms, a la CBD's, core services nodes, major corporates, etc). Even where we felt certain that the bulk of traffic would be addressed we built in large variability over the whole net to account for rising unpredictability in the future.

One example we used was Weta Studios, the production house owned by Peter Jackson, and makers of The Lord Of The Rings series. This studio is situated in a suburban house in a non-commercial area of Wellington (New Zealand) and during the production process has been transferring gigabytes of footage in and out to a number of other locations. There is little, if any, precedent in the historic growth numbers for anything of this sort. And once the movies are made - who knows? The traffic may disappear, never to reappear at that node. There are real economic implications for a telco that invests too much in such situations and can never recoup that investment, let alone make a profit from it.

I don't know about every telco engineer, but rest assured that the clued up ones are well aware that the linear growth patterns of a voice only PSTN are no longer any use in designing the networks of tomorrow.

The key problem is that no-one, but no-one, knows what will be useful.

One suggestion I'd make is to look at the growth of blogs and try to put together some useful numbers on users, traffic volumes, life-cycle of a blog (life expectancy might be useful to know?) so that a telco could make some judgement about how much these sorts of uses might impact them in the near future.

I could imagine a situation where a telco that owns an ISP would see some benefit in offering cheap (or even free, as with usenet) Blog services to their customers, to raise the value of the overall service bundle. Having done so, and having seen some advantages in terms of increased customer numbers, higher network utilisation, decreased churn(which might be very important to them with increasing competition!), etc, they would be more inclined to build such assumptions into their architecture and design processes.

No comments: