John Linton
I commented yesterday about what I saw as an anomaly in the June half year ABS statistics on broadband use in Australia - a monthly growth of around 1% in total Australian ADSL users. I only singled out that aspect of the report as my interest had been in how wireless broadband had grown in that period and that number really surprised me.(TPG/SPT's annual report on the same day announced that TPG was signing up 8,000 new ADSL services a month which would have meant that TPG accounted for 20% of the total ADSL market over the last six months of FY2008 - a truly magnificent achievement).
What also surprised me was the, apparent, fall in the amount of data that the various ISPs reported which fell despite the actual number of users increasing. Now, I could understand if the amount of data rose more slowly as I would think that the new' ADSL users are, by and large, migrating from dial up and would not use very much in their first 6 - 12 months of broadband. Similarly the wireless users would not be heavy down loaders because of the cost of downloads (and uploads). But to actually fall?
I was still 'playing around' with those strange figures when I got two calls within 40 minutes of each other from suppliers of IP bandwidth. While they represented opposite ends of the 'spectrum' in terms of apparent size and reputation the pitches were almost identical:
"due to bringing forward a major capacity upgrade we are able to offer Exetel amazingly good pricing on SX IP if........."
While I'm always grateful to be offered lower prices for essential infrastructure I am always very cautious of 'out of the blue' offers - especially when they are at prices lower than my 'research' has indicated would be available. So I said what I always say: "please put the offer in writing", thanked them for thinking of Exetel and ended the call. Right now Exetel can't consider any major change to our IP bandwidth providers as we have only relatively recently signed new 12 month supply contracts with our long term suppliers. Pity about that as it turns out as the prices being offered are around 30% lower than our current buy pricing and that sort of saving would be very useful right now.
The third, coincidental, incident (perhaps that should have been 'coincidence'?) was a call from Steve asking me to look at some data on the PeerApp MRTG which, although it was a 'mistake' (in the network configuration which was immediately rectified) gave a dramatically improved view of what the new topology of the network could deliver once we have completed the last of the reconfigurations. In essence the new configuration of the network will allow Exetel to triple the P2P traffic that can be delivered from the PeerApp cache by the end of October. If we activated the 'transparent proxy' capabilities that are a standard part of the PeerApp (which we have never used) then the amount of traffic deliverable from the cache would dramatically increase to over five times what is currently being delivered.
Interesting scenario. If other ISPs are using proxies and similar equipment/software more than they have in the past then it might account for a decrease in the reporting of bandwidth usage (depending on how they answered the bandwidth questions which, in turn, would account for why two IP providers are suddenly looking to reduce the costs of IP bandwidth). Then again it might mean nothing at all and I've concatenated three incidents in to meaning something that they don't mean at all. What it does mean, inevitably, is that caching will continue to reduce the amount of 'raw' bandwidth needed by the customers and therefore continue to reduce, at a steeper rate, the cost of delivering the end user services - providing that the end users don't continue to increase their bandwidth usage faster than the cost reductions can absorb - which may well not be the case.
So the ABS statistics seem to indicate a slowing demand for ADSL and at least two IP wholesalers seem to indicate their is a slowing demand for IP. But SX is quadrupling the capacity of its trans-Pacific cables, Telstra is doing something similar and Pipe is planing to add a completely new IP cable. Or are we just about to return to 2000 and the height of the insanity of the dot com boom predictions of cable demand which resulted in IP prices falling by 80% over a two year period?
Someone has to have got it wrong - surely?
Added to this 'conundrum' is the dichotomy between ADSL2 wholesalers like Optus apparently running out of ADSL2 port capacity and being reluctant to add new hardware to increase ADSL2 port capacity (probably part of the NBN 'tender' debacle) and at least three other ADSL2 port deployers lowering prices and seeking to buy up 'wholesale' customers to use their spare ADSL2 capacity or, failing that, buying up other ISPs to use their idle ADSL2 port capacity. All these companies publicly indicate that their businesses are 'booming' but the above referenced three major indicators of user growth overall seem to be saying the opposite.
I will be interested to read the next set of ABS figures to see if there has been a 'correction' to the trends evident in the June half year report. I will also be interested to see what the price of IP bandwidth is in February 2009. Anyone want to take a guess? My guess would be sub $A100 per mbps.
I guess it's best to go back to the old standby of believing only 50% of what you read and absolutely nothing you hear. So if you're still reading this then you better delete every second sentence.
Trouble is which sentences do you delete?