I'm at Ericsson's Capital Markets & Analyst event in Stockholm today. There's been a huge amount of discussion about mobile broadband, in terms of traffic growth, moves towards HSPA and LTE, and the demand for backhaul capacity.
The CEO of Mobilkom in Austria spoke about that country's huge shift towards HSPA, to the extent that it's starting to catch up with ADSL use in terms of susbcriptions (dongles again...). Austria is a bit of an outlier in adoption of HSPA, with data traffic apparently now 20x outweighing voice on the network, but it's an interesting indicator of what's coming down the line.
In particular, it's indicating that the leading HSPA networks are now having to deploy an extra set of transmitters on the base stations. For those readers who don't follow this area, 3G UMTS networks (including HSPA), use 5MHz spectrum slices. Most operators have allocations of 10, 15, 20MHz or more, but typically haven't been using all of their theoretical capacity thus far.
The interesting thing for me is just how fast this occurs. Now, Austria is unusual, but even if you assume that most countries will take two years from launching flatrate dongle plans, rather than one year, to fill up the first 5MHz, it's an early indicator of demand ramp-up for spectrum (and capacity) over the next few years. This is especially true as end users get used to higher-speed HSPA, as well as increases in the total number of users. Add in some growth in data traffic from phones with decent browsers or video clients, and it starts to look as if the 2.1GHz 3G band is going to fill up very quickly.
This has a number of implications.
Firstly, there's a short term business case for femtocells - if they can work out cheaper than adding a second or third 5MHz carrier on the macro network. On the flipside, some of the calculations I've seen have suggested that femtos substitute for new extra base stations rather than adding extra kit to existing ones. I'm not sure what the comparative costs are, but I guess that bring up a 2nd carrier must be a lot less.
[Not much discussion from Ericsson about femtos for 3G macro-offload (or indeed at all), to be honest. My take is that as well as potentially impacting its overall integrated base station/transmission business model and bringing in new competitors, I get the distinct impression that the Big E is a little skeptical about some of the femto hype on a fundamental basis.]
Secondly, it means that operators will need to get extra spectrum if they're serious about continuing to drive mobile broadband. 2.6GHz is the obvious big chunk, but refarming 900 and 1800MHz GSM starts to take on more urgency ( as well as expediency for coverage reasons).
Lastly, it means that operators are going to be faced with some unpalatable choices in terms of capex for HSPA - having to choose between fulfilling the need for extra 5MHz carriers in high-use areas, versus continuing 3G build out in areas with no coverage at all. I suspect that this is going to drive a lot more emphasis on EDGE - and EDGE Evolved, as an interim coverage solution in marginal areas, as it wring more life out of existing 2G base stations. Interestingly, the GSA has been talking up EDGE today as well.
On the other hand, the company refreshed its overall take on broadband traffic, reiterating what it said this time last year. In a nutshell, even with massive growth in HSPA & LTE, the world's total mobile data traffic will still only be a tiny fraction of fixed-line IP (<5%), even out to 2013. The main driver of fixed-broadband traffic is seen as IPTV, especially on GPON fibre and VDSL access lines.