Some interesting facts and figures about the strains of data traffic load on today's mobile networks. With femtocells being positioned as an essential element of the solution, do the facts stack up?
Are mobile networks really under stress?
Reading annual reports from mobile operators takes some real dedication - parts of them can be very dry indeed. But one analyst makes the assessment of Vodafone's recent report, concluding that there's still plenty of spare capacity. Only 5% of their cellsites run at more than 90% load during the busy hour. I wondered if this could be because Vodafone UK only started selling the Apple iPhone last month, and we know that this can dramatically increase traffic far more than email only smartphones. Vodafone forecasts an increase from 20% to 40% for the smartphone share of new handsets sold (noting that it's already 40% in the US).
Despite this projected growth, Vodafone believe that their 3G HSPA+ network has enough capacity available to meet its demands for the next 2-3 years.
My own experience (statistically meaningless I know) is that 3G USB data speeds remain patchy (meaning I just don't use my data dongle, instead relying on the iPhone when on the move) - this isn't specific to any UK network. Voice and text service is generally excellent, mobile broadband data variable and USB data extremely so. The accessibility of data services has meant that I've actively moved towards the smartphone in recent months, instead of using my laptop in certain situations.
Are dongles or smartphones to blame for this data overload?
Another analyst classifies the problem into two geographic regions depending on availability of fixed broadband and affordability of smartphones. In countries without easy access to wireline broadband, such as Latin America, Eastern Europe and parts of Asia, low priced USB data dongles have been very successful - perhaps too much so.
But in regions such as Western Europe and North America, it's the smartphone which has caused the greatest strains. ATT have perhaps seen the most publicity on this, most directly as a result of their exclusive iPhone deal. Om Malik suggests that Skype have held back launching their service over 3G on the iPhone (it only supports Wi-Fi today) due to concerns about network quality. Or perhaps it's because Skype have just announced a deal with Verizon to integrate their service on their network instead. Regardless, ATT have committed to an extensive investment program of almost $20 Billion this year to address data capacity as a key priority.
Is it volume of data traffic or the overhead?
Last week I heard one US operator reinforce a point I made some time ago. It's not just the volume of data that is overwhelming networks (i.e. how many GBytes of traffic carried) but the signalling load from many applications. Dave Nowicki of Airvana calculates that smartphones place 8 times more signalling load that laptops.
Unfortunately, not all apps are well behaved and can place considerable strain as they continually open and close data sessions, perhaps sending only small packets of data each time. I heard of one misbehaving app that was designed to open a new PDP context (data session) for each new page called up. As you can imagine, this has dramatic impact on the network by using it in ways it wasn't designed to cope with.
Data (and voice) pricing are about to undergo a major sea change
ATT's CEO reported that some 3% of their users consume 40% of their data traffic, largely down to the (ab)use of unlimited data tariff plans and sounding the death knell for unlimited plans in the near future. Sprint CEO said last year "As we move into 4G, it'll be much less about minutes and more about gigabytes as the main driver of what customers are buying per month, because it's going to be VoIP oriented. Minutes will be largely irrelevant." Verizon CEO has also been wondering if they will resemble an ISP in the near future.
Femtocells solve both traffic and signalling load
Since domestic femtocells need a wired broadband internet connection, they aren't likely to be so useful in those countries where USB data dongles are the cause of most traffic. In those places, the operator may choose to use femtocells to provide capacity directly in areas of high congestion as part of their own network equipment.
More likely, domestic (and enterprise) femtocells will be popular to meet this burgeoning demand from smartphones and their applications. By handling not just the traffic volume, but also signalling overhead locally on a low cost system, operators can benefit from both offloading their high cost macrocellular networks and from the improved customer experience femtocells provide.
Domestic femtocell data usage shouldn't be charged the same way as outdoor use
This also impacts considerably on the proposed new data tariff plans and pricing approach being taken by network operators. Surely it's unreasonable to charge for data traffic sent through the domestic femtocell - retaining this with unlimited usage seems much more appropriate here. Competitively, domestic Wi-Fi already matches this price point. Instead, use of mobile data service when out and about makes much more sense with some combination of usage cap and traffic shaping to control and prioritise resources effectively.
This would encourage both users and app developers to optimise their use of capped data plans, and might also be perceived as a bonus to potential femtocell customers - i.e. we give you a femtocell, you get unlimited data use at home but your outdoor use is capped.
More on this in a future post...