John Linton
We have, with only a few minor glitches, now completed the cut over of our NSW IP bandwidth from SX via Verizon to SX via Optus. As far as I'm aware there were no negative issues noticeable by any customer and the engineers responsible for the smooth transition are to be congratulated on doing an excellent job - every time the network gets 'touched' I fear for the worst so this is one time when my forbodings were not realised. I was interested to see, while we had both the old' and the 'new' bandwidth on line simultaneously, just how much bandwidth our customers would utilise while there was so much available - depending how much was actually 'switched on' at any particular time there was a theoretical resource of well over 6 gbps at different times.
There was no clear answer and in any event the whole 6+ gbps couldn't have been used as we don't have that much 'customer side' connectivity. However I was interested to see that our customers seemed able to use as much bandwidth as we provided on various occasions while there was almost two gigs or so more bandwidth than we would usually make available in the off peak times. So now we have cut off the Verizon feeds the net upgrade of 'pure' IP across the network is moving over the last week from just over 2.5 gbps to just under 3.5 gbps which is easily the largest single upgrade we have ever put in place. On top of that there is another 1+ gbps of cached 'bandwidth giving a peak usage availability of around 4.5 gbps - quite a long way away from the 10 mbps we started with in February 2004.
I have mentioned before that Exetel's broadband users have accelerated their increase in bandwidth usage quite considerably over the past 12 to 18 months and this exercise has served to underline that the increase may well be gaining more momentum. So I was a little puzzled when I was sent this yesterday:
http://www.itwire.com/content/view/25717/127/
Vocus was one of the IP bandwidth providers we seriously considered and went very close to selecting so we understand a reasonable amount about what they offer and we certainly know the pricing they offered us - which was very, very good. What puzzled me was (knowing the pricing offered to Exetel but not assuming that TransACT bought at it - they may well have obtained much better pricing than we negotiated) a $A500,000 per year doesn't buy you very much pure IP bandwidth - even assuming that TransACT have negotiated much better pricing than we were able to do. So it was this comment, which up to reading this I would have believed was true:
"It’s very satisfying that a company with the size and reputation of TransACT now sees Vocus as mature enough to be a selected as a major provider.”
I know nothing about TransACT's reputation (I assume it is highly reputable) but if the clear implication in this press release is that a $A500,000 pa spend on IP is a major part of TransACT's IP provisioning then they can't have many IP/data users - which doesn't make any sense. So what is a 'sizable' communications company in terms of customers? I am assuming that an ACT user has similar downloads as other users in NSW (as opposed to Victoria where average usage is much higher). Not that it matters in any way other than I was surprised at what it seemed to indicate about a company that is quite prominent in the 'media' and has, apparently, invested what appears to be a great deal of money in its own infrastructure and bid for the Labor NBN1 and is, apparently, lobbying for a chunk of Labor's NBN2. I do believe that TransACT is much larger than that IP spend would indicate......unless......
.....unless TransACT is buying the IP bandwidth from Vocus at a much, much lower price than we were able to get it appears to me that TransACT wouldn't have more than 10,000 or so users which is a fraction of the number I have always assumed they had and therefore I can't believe its a sensible estimate. So the, pretty clear, implication behind the press release for me has to be that we must have bought IP bandwidth at a much higher price than TransACT has negotiated and that, presumably given the ownership structure of Vocus, much higher than the ISPs who are going to get IP bandwidth from Pipe are going to achieve later this year.
It now deeply concerns me because over the IP transition we've just gone through we had been looking so closely at what sort of bandwidth is going to be required over the coming months and the just completed upgrade coincided with the release of the new ADSL1 plans (yesterday) and the new ADSL2 plans (later today) on which we have calculated the new lower plan prices based on likely usage indications and cost assumptions we are now contracted for. Have we (I suppose I should more correctly use 'I') got the buy pricing, and therefore the sell pricing, completely wrong?
It is a little unnerving to read an apparently innocuous press release and then feel a 'cold shiver' go down your spine as your sub-conscious starts to tell you that you need to re-look at what you casually read and then dismissed because maybe it means something entirely different. I can't go back and renegotiate the IP pricing now nor can I change the residential broadband parts of our FY 2010 business plan (which is only a handful of days away). So what the problem for Exetel might now be is that other ISPs are going to be buying IP bandwidth at substantially lower prices than we can buy at and therefore significantly impact our ability to remain competitive as the next few months deliver those possible cost savings to our competitors.
I badly need a holiday, but if it is the case that we have bought badly then I may not be getting one for a very long time.