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

Wednesday, October 11, 2006

The continuing rise of picocells....but femtocells? Still early days.

I'm impressed by the continued rise in profile of cellular picocells over the past couple of months. I've actually been following the area for about 6 years, and until a year ago, I seemed to be the only person who ever mentioned the technology outside of very specialist "indoor wireless" events. The first mention I heard from a mainstream operator suggesting that they might have a role in FMC / nextgen-FMS (fixed-mobile substitution) was O2, at a conference on wireless VoIP, last November. They were also absent from most FMC infrastructure vendors' discussions, again apart from a few niche applications like ship- and aircraft GSM coverage.

The pivot point seemed to be around 3GSM in February, where I counted at least 10 vendors and silicon companies pitching various types of small base station - mostly aimed at 3G in-fill. Sure, there was still some cynicism about the concept - I remember a senior Nokia Networks representative disparaging special indoor cellular systems with the line "Outside-in always wins".

The next boost, in PR terms, came from the UK low-power GSM licence auctions in May, when people suddenly realised that this probably needed special types of infrastructure installations.

There was a little lull over the summer, but in the last 4 or 5 weeks it's all kicked off again, both for picocells and their svelte sisters, femtocells (aimed ultimately at the residential market). My understanding is that there are numerous major operators trialling the technology in their labs in Europe, North America, and I suspect Asia as well. It's been telling that operators who are publicly hostile to WiFi and dual-mode phones have been cropping up at FMC-related events with surprising frequency.

The usual culprits being mentioned are Ubiquisys and ip.access (the veteran in the market and now part-owned by Motorola since TTPCom's acquisition), with PicoChip named as a clear driver on the silicon side. There's also been a fair amount of reference to assorted Asian vendors, as well as some of the major players in the current home-networking and gateway space.

Also, a number of radio network infrastructure specialists, as well as "convergence gateway" vendors have been adding picocells into their slides, in some cases sharing technology components with dual-mode solutions.

To be honest, I suspect we're going to see a bit of a hype build-up, with various over-enthusiastic forecasts for mass rollouts in 2007 and 2008. I reckon that's a bit premature, to be honest. Looking at the evolution path of recent innovations like dual-mode phones (UMA or SIP), or even WiMAX, I suspect it's going to take rather longer than some of the evangelists suggest - and I'd also say that at present, the residential femtocell concept has yet to prove itself even in the lab, let alone in a realworld trial.

In fact, I think that anyone looking at this sector should probably analyse the pico and femto markets separately - and possibly 2G and 3G as another dimension too. The use cases are different, the proof points are different, and the economics are different.

On the one hand, we're hearing T-Mobile making concrete decisions about 2G network fill-in using what is now comparatively mature & uncontroversial technology, in comparatively simple standalone deployments.

But on the other, the concept of shipping millions of femtocells to home users, magically brushing away an RF management issues, expecting users to simply plug the box into the back of their existing broadband gateway/router, and happily using cellphones indoors on their own private 3G access point is, frankly, a long way off. I suspect that, like UMA, loads of practical issues are about to emerge. Yes, the handset issue will be less important (although I suspect some customisation will still be needed). But the real-world testing of technology (eg femto-macro handoff) and user experience (what happens when someone unplugs the mains-powered femto when they're hoovering... and how does the operator do customer support when someone else's home gateway is in the way) are yet to be determined.

Putting all the pieces together will take until mid-late-07 for "proper" consumer trials. Decisions on early deployments may be made in early-2008, but rollout will be slow (maybe 10's or 100's of thousands), and even if all the practicalities sort themselves out easily, late-08 / early-09 is a realistic timeframe before we can start talking in millions and a potential "hockey stick"

Pico's, on the other hand, are shipping in small volumes now. This will continue to grow steadily with demand for niche 2G fill-in and enterprise applications like the UK LPGSM services. 3G indoor coverage and capacity enhancements for offices & public spaces is also becoming important with the growth of EV-DO and HSDPA. The big variable will be if other regulators adopt the UK low-power spectrum licence policy - although timelines are likely to be protracted given the likely arguments from existing licence-holders.

In any case - I expect the next few months to see a continued ramp-up of interest and (maybe) over-enthusiasm about low-end / low-power cellular.


alex said...

7-8 years ago, some of the large network infra vendors offered GSM Pico base stations with the promise of huge CAPEX savings.
This failed because the higher OPEX for telco-style installation & maintanance far outweighed CAPEX savings.
GSM and GPRS coverage from outside has been indeed more cost efficient, complemented with some indoor signal distribution systems, which are connect also to "big" base stations.

Now it is true that 3G (WCDMA and HSPA) indoor coverage for high-speed data is difficult to achieve from outside, mainly due to higher frequency range compared to GSM900.

But what has changed in the economics that would overcome the high OPEX barrier?

High revenue (potential) for 3G data that outweighs high OPEX?

Drastically simplified operational requirements & procedures?

New business models where somebody else operates the pico base stations?

New spectrum for low power GSM?

It seems to me that several of these factors would have to come together in a disruptive way to make pico base stations viable, and that this is not likely going to happen within a few years.

Without a cause for disruption, I'd see natural evolution from the current situation as "default scenario":
Outside-in best for GSM voice, and ok for 3G data considering the still very low usage & need.
And WLAN in enterprises & homes for high-speed wireless data, expanding very slowly from PCs to mobiles.
And eventually 3G moving to refarmed spectrum in lower frequency bands.

Gabriel Brown said...

The high opex was due to the requirement for leased line backhaul (which kills it dead).

How much of the enthusiasm for this is driven by the ability to increase 3G termination revenues? How fast will regulation erode this aspect of the business case?

alex said...

Yes, leased lines was a killer.

But even for operators with own transmission network to the potential pico sites, installation & maintenance was too expensive.
For example, power outlets are often not in locations which are optimal to get coverage (e.g. under the ceiling). Since GSM spectrum is scarce, number of cells needs to be optimized.

WLAN suffers to an extent from the same problems, which is why corridors etc. often lack WLAN coverage. End-users accept that for data but not for voice.

So what might be viable is pico base stations for hot-spot data capacity, with coverage provided from outside.
But then in many such locations people will use PCs anyway for data(in developed countries).

Meaning I still struggle to see the volume market for pico base stations in the next few years.

Anonymous said...

Over 70% of mobile useage is in the home or office - and the split is about equal. And content - video/music/games/images/webpages - is much more likely to be consumed in the home.

As mobiles move toward Web 2.0, it will become ever-more compelling to do web things on the mobile in the home - especially when your sibling/partner is hogging the PC.

Good 3G coverage in the home is hard. And 3G at 900 is a ways off (but if OFCOM shuts up it will come quicker in Europe anyway).

A Femto could be set at nanowatts of output power, literally not even a sniff going beyond the house, or if it did, a local macro wouldn't see it as it'd be well below the noise floor, especially in even moderately busy areas. This also would mean that the problem of noise rise in macro cells would decrease as those mobiles transmitted on their minimum power step to the Femto. Or my WCDMA RF theory could be rusty, but that's how I remeber it anyway :-)

Femto's could be cheap (<$100), easy to set up and need only a dumb pipe, connecting to the core. Your own private HSxPA link - probably waaaay faster than the ADSL line it's plugged into.

Handover is nice, but not necessary. What % of calls would benefit from handover arriving at or departing the house? A tiny fraction I bet. Like UMA/dual-mode handover, it's a solution looking for a problem.

Consumer pays for Femto. Consumer pays for backhaul.

Operator gets to sell a bunch more content to the user in the home (as well as that other thing...oh yeah, voice), and doesn't need to build more Macros.

Where's the problem? Bring it on I say!



Peter Judge said...

Eight years ago, pico-cells were a high-reliability offer for businesses, with backhaul to match - tht's why Opex was high.

Now we have the Internet, and easy (well nearly) ubiquitous broadband. That allows picos to make a pitch for the consumer, with consumer-level reliability, and consumer-level pricing.

There's going to be hiccups as with other disruptions, and there's a chicken and egg aspect to getting anything into the consumer space, but ultimately, it makes sense.


Dean Bubley said...

Some good commentary here.

Alex - main differences between now vs. seven years ago & the first round of picocells is that this lot are intended to be IP-connected, either via home broadband (for femtos) or an office LAN (for picos). Also improvements in provisioning / management systems, and the requirement for considerably less maintenance with simpler products with dedicated chipsets & lower power (eventually moving to power-over-ethernet in some cases). Another shift is the higher cost of acquiring new outdoor macro sites, especially in urban areas.

Mike - hmm, maybe. There are still a whole bunch of issues to solve - such as configuration / support when a femto is plugged into the back of an ADSL router. Theoretically we could get Home-Broadband-Router-Femto-WiFi-Gateways, but it's a few years until they're massmarket too. A bunch of issues around distribution too, as femtos will be operator-specific & therefore increase things like channel inventory requirements. Also problematic is the 80%+ of cases where there are 2+ operators used by different family members. Two or more femtos? And as for "operator gets to sell a bunch more content"... I'm unconvinced, as that content has to compete on price & quality (and transportability) with the same content accessible on the PC in the next room.

Anonymous said...


to me mike has a point, i think users would prefer to use the mobile device over the pc if it worked. i find myself checking email and news on the phone, just because it is next to me and has no boot up time!!

on the other hand the 2 operators, issue is a big one. also who pays for these devices... operators subsidize these boxes and the handsets...

Anonymous said...

In the foreseeable future, femtocells are likely to stay a GSM/UMTS oriented solution. Qualcomm ain't going to move on the EV-DO side anytime soon and you need their support so that the terminal can handle the hand-off between the macro and the underlay (femto-cell) network. Also, the US is particularly hard since as an operator you need to build the femto-cell boxes to be location aware. Channels vary by location and you don't want to run into issues of a user buying a box in LA and then moving to Vegas. A-GPS in the femtocell box is only a partial solution to that.