Mobile data offload with femtocells

Rupert Baines

Rupert Baines, VP marketing at picoChip, on the femtocell scam run by mobile operators.

It would have been difficult to miss the recent controversy over Vodafone and AT&T counting traffic used over a subscriber’s femtocell towards their overall data allowance. Congestion on the carriers’ networks is a big problem and until recently it seemed obvious that femtocells offered a good solution, or at least part of it. And yet both operators have decided that (for now at least) data traffic on your femtocell is still part of your plan and counts towards the data cap (and so potentially if you exceed the quota then femtocell data usage is charged at full rate).

According to Signals Research Group, data on the macrocell Radio Access Network (RAN) costs around £5 per GB. But on a femtocell as architected today, offloading the expense of the macrocell but with data passing through the carrier’s core network, this drops to around £2 per GB; and on a femtocell with ‘true offload’, £0.04p per GB (avoiding the core too).

Femtocell scammers

So, even though using a femtocell reduces their operating costs by two thirds, Vodafone and AT&T will charge customers full price, by counting this traffic against their subscribers’ data caps.

As a rationale, the carriers point to the legal requirements on them to monitor and control traffic; primarily, the need to implement Lawful Intercept (LI). Legally, they must support wiretap technology on 3G to be able to monitor all data usage for illegal activity.

What this means is that all 3G traffic has to go into the core network for authentication first and cannot pass directly to the internet, as is the case with WiFi. Transmission from the femto to the wireless operator’s core is by the ISP, but the carrier pays for routing (via its own infrastructure, primarily a Serving GPRS Support Node or SGSN) and peering/termination (instead of the ISP).

Confusing times

Of course, the LI rules aren’t applied to WiFi. The logical conclusion of this situation is that the security services’ right to listen in actually depends on the specific radio in the handset. With 5MHz CDMA they can listen; with 20MHz OFDMA they can’t. VoIP-over-3G is different from VoIP-over-WiFi. The paradox increases when LTE is added to the mix: a 20MHz OFDMA signal at 2.6GHz (LTE) will be treated differently to a 20MHz OFDMA signal at 2.4GHz (WiFi)!

But I realised long ago that ‘logic’ is not the key principle in regulation. The technological solution for operators is something called SIPTO (Selective IP Traffic Offload). Using this descriptively titled approach, the carrier can select, by looking at the nature of the traffic, exactly what should be offloaded and what needs to travel via the core. So ‘Mr Black Hat’ gets routed, and tapped, in the carrier core; ‘Mr White Hat’ gets sent straight to the internet, with no interception. In contrast, WiFi sends all traffic to the internet, unmonitored and untapped, whatever it is or whoever is involved.

There are complexities: according to the LI rules, operators have to keep traffic records, so they might choose to gather the signalling and URLs, but offload the content itself. Or they might need to prevent certain users (for instance under 18s) from accessing inappropriate content on their phone.

Controversial question

Whatever the technical solutions, I would think a saving of two-thirds must be worth having from the carriers’ point of view. If I were them I’d offer unlimited data femtocells and accept the small cost. Certainly if, like AT&T, an operator has been offering an ‘unlimited’ voice service, it would make sense to make it unlimited for femtocell-data too, generating a clear profit into the bargain.

It’s a controversial question; one that carriers are wrestling with, but that they should be able to solve straight away, at least as far as customers are concerned. Reducing costs by 70% is worth doing.

Softbank, with its free femtocell pitch in Japan, has got it right in this respect. Although AT&T and Vodafone have made excellent steps in bringing femtocells to consumers, on the data cap question they have a long way to go.

picoChip provides technology products that enable new types of cellular communications service and that can cost effectively improve existing services. http://www.picochip.com/