If mobile data billing takes too long to change, device-based intelligence will fill the gap

My recent posts have discussed the issue of femto-originated data traffic being charged against the same caps and policies as macro traffic. I think that this is likely to be seen negatively by many consumers - it is fundamentally unjustifiable, as the cost-bases are so different.

It reminds me of one of the problems with many implementations of earlier voice FMC services such as UMA or VCC - where a mobile call is terminated via WiFi and the user's own fixed broadband network rather than the inbound operator's cellular infrastructure, the justification for keeping the call termination wholesale fees the same diminish. In that case, the customer (and the originating operator) is paying not just for a network it's not using, but for spectrum as well.

The problem is that it's very difficult for billing and rating engines to take account of cost differentials involved in a given call or data stream - not least because most operators don't really have neat and accurate per-min or per-MB costs, and it can vary according to lots of specific circumstances. The cost of connecting to a voicemail server is different to connecting to a phone; the cost of a call in a rural area is different to a city centre and so on.

So operators just take a broad average, reflected in the retail and wholesale prices and hope that people don't nit-pick too much - at least not so much that the regulator gets involved.

While I think that's fair enough - we're never going to get precise cost-based pricing even if we want it - it doesn't excuse the most *egregious* examples of iniquity. The femto vs. WiFi traffic against mobile data cap issue does seem pretty egregious, because there's an immediate A-to-B comparison available to the end user.

The industry spends a huge amount of time telling us that "customers don't care what technology they use, they just want to get connected". Which is fine, but only if there's no glarig differences in price and performance.

This comes back to my assertion that there is a need for holistic approaches to mobile data traffic management - combining all the various bits of the puzzle to offload / charge / optimise mobile traffic so that this type of head-smacking event and potential PR/loyalty disaster doesn't occur.

Although we still haven't got a detailed comment from AT&T (and Vodafone does the same) about their billing system, it is not uncommon for any major changes to take a considerably long time to implement. That's especially true in cases, such as here, where AT&T was also moving its whole system from flat-rate charging to tiered-and-capped. They may well have been aware of the femto vs. cap unfairness, but decided that it would take too long, or involve too much risk, to address that at the same time as the more general switch-over in policy and charging.

This situation is likely to be repeated throughout the data traffic management arena, in various scenarios. It is *theoretically* possible to hook various bits of the policy/charging/optimisation infrastructure up to the radio network, or other bits of the carrier's machinery. But it's not easy, it's not cheap, and it's not quick.

That leads me to think that there will have to be some short-term workarounds that are easier to implement. Top of my list is smart connection-management clients for smartphones and laptops, which have awareness of tariffs, data consumption and different access methods. "Are you sure you want to connect via the 3G femtocell? I can easily register with this nearby WiFi if you want free connectivity".

It will be interesting to see if the slowness of creating sensible end-to-end traffic management and billing priniciples in the OSS/BSS leads to the device intelligence taking the lead again. I'm tracking a number of infrastructure vendors who are looking more closely at the role of the device as a key actor in determining mobile data routing, and I suspect their collective work will pay off.

For more detail on the technologies available for holistic mobile broadband traffic management (including device connection managers), please consider purchasing my new study, from $350.
NEW Mobile Broadband Traffic Management Paper