Speaking Engagements & Private Workshops - Get Dean Bubley to present or chair your event

Need an experienced, provocative & influential telecoms keynote speaker, moderator/chair or workshop facilitator?
To discuss Dean Bubley's appearance at a specific event, contact information AT disruptive-analysis DOT com

Tuesday, January 27, 2009

Enterprise UMA: Dead on Arrival. Enterprise Femtos: Come back in 2012

There seems to be a sudden flurry of activity around enterprise FMC version 2.0. I'm expecting to see a large amount of further hype over the next few weeks, especially around MWC in Barcelona.

In my view, it's all smoke and no fire.

Meru Networks has inexplicably partnered with T-Mobile US to offer a corporate version of the dual-mode WiFi UMA system. Its "distributed single access point" architecture makes it slightly less unsuitable than the more complex WLAN networks from other vendors, that would necessitate AP-to-AP handoff. It is highly notable that competing WLAN vendor Aruba put out a white paper about UMA a year ago which I discussed here, but has been very quiet on the topic ever since.

Quite why T-Mobile is still flogging the dead UMA horse is a separate question. It doesn't take much imagination to guess why it's made no mention of subscriber numbers to Hotspot @ Home. If it was even slightly successful, it would be shouting from the rooftops.

I've written many, many times over the past 4 years before about UMA's deficiencies for the general consumer marketplace. In the enterprise, there are even more challenges - security, PBX integration (please don't give me the substitution pitch yet again), devices, roaming, numbering, operator lock-in, WLAN vendor lock-in, WLAN coverage, TCO, management, monitoring and above all channel to market and support. Then there's the question of how the enterprise can use the WiFi on the devices for non-carrier applications. Oh, and most of the most popular dual-mode WiFi smartphones (except RIM's) don't have UMA. It's not in the G1, most obviously.

The only upside I can see is that unlike most of the PBX-centric approaches to corporate FMC, you can actually get SMS to work OK.

In a nutshell, I'm surprised that Meru is wasting its time and money on this. Yes I'm sure there will be a handful of demo sites, and maybe even some sort of commercial launch. But I'm utterly certain that in a year's time it will have had precisely zero effect on the overall enterprise comms market.

Enterprise femtocells are another emerging hot topic - but one which has slightly more "legs" - eventually. I've been hearing escalating buzz about this for the last 6 months or more. I first wrote something on this in May 2007, when my conclusion was that 2012 seemed like a reasonable timeframe for massmarket commercial offerings. Maybe I'm being optimistic given the economy, but I still reckon that's about right.

What appears to be happening is that the femto guys seem to have learnt some of UMA's lessons, and "get" the complexities of corporate integration. I've heard talk about PBX integration from day one, awareness of the issues around firewalls and network management, and definitely some cognizance of the difficulty of getting and retaining a channel. I'm expecting the lead routes to market to be hybrid fixed-mobile operators with large integration units and an loyal established base of corporate PBX customers.

It's still not going to be easy though - there's all sorts of little hidden issues to sort out in large buildings / enterprises (eg access to ducts and cables, do you need power-over-ethernet for femtos, what are the compliance and Sarbanes Oxley implications of 3rd-party infrastrucure etc). I'm expecting some others to go for the other end of the spectrum and start talking about SMEs. In theory, yes, SMEs could go "all wireless". But in reality it's not that simple, and there's the usual challenges of 10,000 channels to get to the buyers, high cost of sale per customer for little margin, and niggly annoyances like integrating/replacing landlines for faxes and security alarms.

To give an example of the complexities - what happens if a business gets femtocells installed on its premises... and then one of them gets stolen? Or hacked? Whose responsibility is it? I see lots of revenue for lawyers....

There's also a little bit of fuzziness about how "enterprise" gets defined here, and I expect to see a few "corner cases" emerge before the real massmarket. In my view, the main categories are:

- SME sites (small offices, shops etc)

- Private large-enterprise sites (corporate HQs, warehouses etc)

- "Public" business locations with lots of visitors (big retail stores, government buildings, HQ lobby areas etc)

The first point to make is that femtos are very poorly-suited to visitor-heavy locations, as they are single operator only. I'm not sure that a council would want to improve coverage or costs for just 25% of visitors. Obviously, they could confine the benefits just to employees, though - assuming the signal doesn't mess up macro-network coverage for visitors with interferece, deep inside the building. There's possibly some clever stuff that could be done with roaming - so maybe deals can be cut which enable Big Corporation Inc to say "cheap calls for all our overseas vistors while on the premises" although that's not as easy as it sounds either.

There might also be some other "hotspot" business cases for cafes or hotels or airports (eg the "free mobile broadband" business models I discuss in my recent report). I could certainly imagine a sign on (for example) Costa Coffee, saying "Free HSPA access for T-Mobile users". Is that "enterprise femto" if it's the coffee chain paying the bill?

I reckon that ip.access knows more than most in this sector about the corporate market. After all, it was pitching the concept of enterprise picocells back when I first met them in 2001. I remember mentioning to the CEO at the time that it wasn't as easy as it looked, and that the usual problems of attaching an operator-owned box to the corporate-owned LAN would apply (demarcation points, VLANs, firewalls, support, management, monitoring again etc). Over time, I was proved right, and the majority of its picocells in business environments (I think) have been put in singly, often with a completely separate backhaul connection rather than sharing the LAN. They've done some analysis on picos vs. femtos which I'm inclined to agree with as it reduces the need for things like femto-femto handoff - but neither is easy, either technically, operationally or commercially.

So, I wouldn't be surprised to see a few instances enterprise-ish femtos appearing in 2010-2011. But I'm not expecting to see "Company X deploys Operator Y's femtos throughout its new campus site and integrates with IP-PBX vendor Z's infrastructure" for a long time yet. It's not just a case of hanging a couple of cellular access points off an existing LAN.

No comments: