INSIGHT: The maths behind unlimited data
- 20 November, 2014 07:26
The commoditisation of consumer and business grade internet is to a large extent now complete, and enterprise/corporate internet is close to following suit.
My definition of commoditisation, and hopefully I have nailed this, is when services are largely uniform between suppliers from a consumers point of view and price is the primary determining factor in the purchasing decisions made by consumers/business managers.
In recent times we have seen unlimited data on business plans in both NZ and Australia become the norm.
They don’t suit every business, some companies just don’t do the volume of traffic to justify it, but for those that do, the option is now available at very competitive rates.
In recent times SNAP has released new plans, as has DTS, and MyRebublic has made an entry into the NZ market with unlimited plans the only option.
In the days of slow circuit speeds, which you could argue Australia still lingers in, providing unlimited data was curiously enough fairly uncommon.
ISP’s purchase international and domestic transit by the Mbps (speed), but traditionally sold data by the GB (data volume), meaning that that if an ISP had to buy more bandwidth it was likely to be because data usage, and therefore revenue, was increasing in a near linear fashion.
Unlimited plans, assuming no rate shaping is enforced by the ISP, only restrict a client’s usage /speed by their physical access, and therefore the factor that determines the potential overhead to an ISP of providing unlimited data comes down to the speed of that connection.
If a client has an unlimited plan on a 200Mbps UFB circuit, they represent quite a high potential cost if you need to ensure that same amount of domestic and international transit is available to them. That of course is why contention ratios exist.
Some ISP’s may truthfully state that they don’t rate shape or limit connections (and neither does DTS), but not one can say they don’t contend services to keep costs at an acceptable level.
You could argue that unlimited plans over slower speed circuits cost less and therefore the cost to consumers and ISP’s is proportional, but the difference between unlimited plans over 100Mbps and 200Mbps UFB circuits is about $50, and that in no way represents the difference in potential costs to an ISP of providing those services.
ISP’s have set price points, DTS included, that now push clients towards higher speeds with no data caps, and the factor that will determine the difference between client experience between ISP’s will be how much they contend their services to maximise profit/minimise loss.
Australia on the other hand really has a bit of a gap in the market between ADSL2+ and symmetrical copper/fibre services, with only a few suburbs or buildings in metro areas having access to an intermediary service such as VDSL or NBN.
Because of this, the majority of unlimited business plans are over SLA backed enterprise/corporate grade connections and unlimited plan rates are largely calculated p/Mbps in the same way HSNS (or equivalent) services are in NZ.
In either country, if you want uncontended services that are capable of running flawless business critical cloud based applications, you will be paying for unlimited services by the Mbps. The good news is that he price per Mbps is now largely around $10.
What impact will unlimited plans have on the market?
This shift in the market places the emphasis on smart network design and scale.
I am a bit of a broken record when it comes to talking about scale but I struggle to see how small ISP’s can compete in the NZ market when they need to deploy POP’s in each UFB region while being able to afford enough bandwidth to bring on unlimited UFB clients at the same rates as their bigger counterparts.
A few years back people would have said that good service and easily accessible support would justify a service premium, but in 2015 my view is that those aspects are expected to be provided at the same low rates by all ISP’s.
As for network design, when you are selling low cost data, you need to route it at an even lower cost. This means caching, peering, and being smart about which carriers you work with.