How did you come to be involved with Femtocells?
I've been involved with Digital Mobile Communications in Mobile networks since 1985, 3G technologies since 1990, and wrote one of the main industry texts “The WCDMA Design Handbook” in 2003. After some time with Philips and Nokia, I set up an independent training consultancy in 1995. I've always been interested in the small base-station concept which has been around for perhaps 20 years.
During 2003, I first conceived a full 3G cellsite in a small box with SIP backhaul. In July 2004, I filed some patents and formed 3Way networks in August with 3 colleagues. We weren't involved in the naming of the femtocell - at that time Home Basestation or Home Node B were more commonly used.
In April 2007, Airvana acquired 3Way to complement their CDMA femtocell product, and I became their VP Systems for the UMTS femtocell.
And what's the background of Airvana?
The company has been a supplier of CDMA 3G basestations, using EV-DO technology, and shipped tens of thousands of cellsites across the US and elsewhere. These have been sold through channels, such as Nortel, so customers may not have been aware of our brand name and engineering behind it.
Airvana saw Fixed Mobile Convergence (FMC) as a growth area. They identified the need for a femtocell, and reused their existing CDMA technology to build one. When considering a UMTS femtocell, they looked at building their own or buying a company - and bought 3Way as a result.
Both markets help each other, and the higher volumes help to reduce costs when buying from suppliers. However, the standards are sufficiently different to require a different approach.
For UMTS, our guiding principle was to own a complete solution. This led us to buy commercial silicon and implement our own baseband algorithms and upper layer software, rather than using a dedicated chipset. We're not using a DSP, so we don't have issues with high power consumption.
What femtocell architectures do you support?
3Way was SIP based from the start. It became very obvious that the UMTS operators wanted IMS in the long term, but had issues with immediate deployment:
- How would it integrate into the core network? There is no IMS in the core today.
- Mobility management between SIP and standard 3G core network.
So we quickly switched to an Iu based solution working with Nokia Siemens Networks (initially as 3Way and then Airvana), subsequently the femtocell industry adopted the Iu architecture also.
We've continued to work with Nokia-Siemens, and have actively participated with them in developing the Iuh based standards. We see a clear and easy path to migrate to SIP in the long term.
In the CDMA camp, there is much more demand and alignment with the SIP based approach. We have partnered with Alcatel-Lucent, Hitachi and Motorola in this.
The femtocell standards will be in a solid position during Q1 2009:
- There are a few issues to tidy up, which will be finished before the end of that quarter
- A more detailed object data model is required for TR-069 (the remote management protocol)
- It generally takes around a year after the standards have been agreed for the industry to have complete product. So expect lots of trials towards the end 2009 with standards compliant commercial product to follow shortly thereafter. Expect a press release for the "first Iuh call" shortly after the standards are fully ratified!
There is an IOT (Interoperability Testing) group within the Femto Forum. This group may organise "plugfests" in due course - it's early days and it may take some time to organise these. During the first year, expect more one-on-one, followed later by a mix-and-match approach.
What's your forecast for femtocell pricing?
This is largely driven by market volume. We're comfortable in getting down to lower price points, because we have our own stacks and baseband processing. Things are moving out to the right, so we're unlikely to hit high volume much before the end of 2009.
In some ways the industry was caught out because it didn't expect the femtocell standards to be agreed so quickly. Telco’s are often quite conservative - no-one wants to be the first, but would quickly follow once demand and operations are proven.
Route to Market - Airvana has chosen to sell through others. Why?
Operators want a single system integrator - an owner for the end-to-end solution:
- Their supplier must be large enough to handle all aspects, such as back-office systems integration.
- The femtocell touches almost every component of the operator's network, so operators want partners who can manage this complexity.
Moving forward, as the standards become more widely adopted, and femtocells become just another CPE (Customer Premises Equipment), this may change over the next three years.
On the far horizon, residential gateways are where the market is headed. It’s an important future for Airvana. In the short term, we work directly with the operators, priming the market. We also work with Thomson and Pirelli to add value to both their solution and ours.
What are Airvana’s USPs (Unique Selling Points)?
Our single top selling point is owning our complete technology solution. It’s not necessarily the easiest path to market, but in a fast-emerging space like femtocells, it will give us the greatest flexibility to adapt to changing requirements. Second is our partnerships. We realised early on that this cute little box was part of a larger system. That explains why we’ve been working closely with so many infrastructure and CPE vendors.
We’re also working on stuff that I can’t share with you today, so that will have to be an excuse for a future chat with ThinkFemtocell!
Does Airvana also address the Enterprise market?
That’s a broad term – it’s not really a monolithic marketplace. There are many different opportunities. We see more of an evolution path. We’ve also seen demand for an “open access” solution from operators, to be used within retail outlets etc.
Unlike domestic femtocells, which would normally be restricted to use by members of the household, using what’s called a “Closed Subscriber Group”, an open access femtocell would be a remotely managed device providing service to any user of the owning network.
I’m not convinced by the “super-femto” story, and don’t see a huge opportunity for very large devices, although I’ve heard some operators clamouring for up to 32 channel “super femtocells”!
I believe the sweet spot is somewhere between the current 4 channel and 32 channels.
What are Airvana’s plans for LTE?
We follow all of the standards and technologies. There’s a question of when LTE will really happen – will the “credit crunch” limit investments or stretch timescales?
I don’t see LTE happening in the femtocell side for 2-3 years. It’s likely that initial products would be 3G/4G dual mode when deployed. The timescale and format will be driven by the operators.