Voice and Video

Posts: 105
Registered: ‎04-09-2007

Offloading cellular data to pelicans*

“A funny old bird is a pelican,
His beak can hold more than his bellican.”
Dixon Lanier Merritt The Pelican

While the pelicans have temporarily flown, leaving the Sunnyvale salt ponds to insolent pairs of Canada geese, the Aruba Runners’ Club ruminates this month on cellular data capacity. How much do they (the operators) have, how much do we (the consumers) need, and what dors the future hold for each?

For these are the fundamental questions behind the ‘cellular data offload’ movement. It is not disputed that some areas of cellular networks, notably AT&T in San Francisco and New York, and also O2 in London have experienced congestion due to smartphone data usage, with iPhone owners fingered as the primary culprits. AT&T has the now famous quote that 3% of their users are responsible for 40% of the traffic.

It is stating the obvious to conclude that (data) demand has outrun capacity on these networks, and other operators are facing similar problems – even if they are not hitting the buffers, they see the trends and they are unfavorable. They have three possible paths.

Increasing network capacity is the remedy most operators instinctively favor. Operators have their spectrum, network and brand, and their engineers will be working through HSPA, HSPA+ to LTE, as well as modeling increased spectrum and cell sites – we already se e press releases publicizing efforts in this direction. They should have a good idea of how much capacity they will be able to offer at various dates in the future, and the associated capex plans.

The second remedy is to cut back on data consumption. Our ears are not close enough to the ground to confirm it, but we suspect recent comments from operators are indeed softening up the market for an introduction of tiered pricing models for data use. Leaving aside the questions of customer acceptance, competitive effects and visibility, these plans will be designed to make consumers conscious of how much data they are using, and to send price signals to encourage the more profligate to cut back. Will it be effective? That is not yet clear, because the operators certainly don’t want to discourage their heavy users from exercising these services: that would put the whole industry back. Now, the models need to incorporate the effects of these pricing structures, but also to forecast a rise in the number of smartphones and concomitant data usage. Will AT&T’s brave young 3% become 13%, even 30% - and how fast?

We suspect most operators have been feverishly running these numbers over the past 12 months. They will have a graph showing how fast they can increase data capacity, compared to the data their customers will demand. Our guess is that the network capacity will be insufficient for several years, and perhaps forevermore. How to remedy this?

Only two possibilities exist. Both offer routes to offload data currently carried over the cellular network, shifting it onto broadband landlines. They are our old friends from FMC days, femtocells and Wi-Fi, and they are close cousins.

Femtocells allow the cellphone to use its primary radio, so the client device is simple. However, the slow march of femtocell acceptance – we are only in the trial and limited availability stage, despite 3 years and more of intensive development – shows the complexity on the network side. A femtocell uses licensed spectrum, so its transmissions must be tightly controlled and coordinated with the macro network. There are issues of control from the cellular core network, and the business case: consumers still balk at paying for a device to fix the operator’s coverage issues, even at today’s very subsidized pricing levels.

Meanwhile Wi-Fi offers performance, installed coverage and low cost. It’s pervasive, both in the home and office (the places where there’s heaviest data use) and on every new smartphone. Consumers are happy to buy them at retail prices, and many iPhone users already switch to Wi-Fi where available because of its higher data rates than cellular over-the-air protocols.

This is not to say that femtocells have no future, but they will remain in a niche until the issues above are overcome. Meanwhile Wi-Fi offers an attractive solution for cellular operators. Evidence? Despite the understandable reluctance of cellular network engineers to move traffic off-network, both AT&T and T-Mobile have set up their smartphones to identify and automatically connect to parent company Wi-Fi hotspots. That’s the most telling evidence. The operators would not be doing this if they thought the data capacity crunch was temporary, or if another avenue was available.

Meanwhile many vendors are prancing center-stage with more interesting, even exotic data offload solutions. Just a few of the cast: Kineto, Stoke, Intellinet (Azaire), Bridgeport, Accuris. They all have different approaches, but they are focused on the second and third stages of data offload.

The first stage, as we discussed above, is simply to divert data traffic over Wi-Fi rather than cellular, relieving the cellular network. It’s easy, effective, but incomplete.

A simple offload function like this completely loses the operator’s customer touch – the core network has no knowledge of the device or what the user is up to. So the second stage is to authenticate over Wi-Fi to the cellular core so there is a single sign-on function. Now the AuC and HLR (insert your 3G etc. equivalents, we picked up the lingo years ago and don’t have any appetite to learn more acronyms) can see the smartphone equally whether it’s on the cellular or Wi-Fi network. And it’s also possible for the operator to deliver proprietary services from the core network, assuming they have some, just as they would over HSPA etc.

Stage 3 takes a swing at the big earners, voice and SMS/MMS. If we can deliver these over Wi-Fi, we have a solution that will work where 3G coverage fails, with obvious advantages. And our shrewd and perceptive readers will recognize we have come full circle to FMC. No wonder the UMA and VCC pioneers are back on the trail!

We’re out of space, but if you have an interest in UMA vs VCC vs I-WLAN for cellular offload over Wi-Fi, there will be a white paper out in a few weeks.

Meanwhile look out for those pelicans returning, and ponder that they solve the problem by passing food from beak to bellican but offloading the accompanying mouthful of water*.

Search Airheads
Showing results for 
Search instead for 
Did you mean: