Friday, April 25, 2008

UMA+LTE: It's all good

Recently there has been a flurry of activity around the next generation Radio Access Network (RAN) technology known at LTE. LTE is the ‘all IP’ access network for mobile service providers. With very high throughput (>100mbps) and very low latency (targeting 5msec), LTE is designed to a be a true wireless broadband access network.

With that in mind, people have started questioning UMA’s role in an LTE environment. While there are a lot issues still in flux (including LTE itself), and much work still to be done, what is clear is that the core drivers behind UMA in a 2G or 3G environment still apply in LTE.

Some people simplistically view UMA as redundant in an LTE environment because they both ‘deliver services over an IP access network.’ This view completely misses the value of UMA in a Home Zone 2.0 (femtocell or dual-mode handset) service deployment.

Operator deploying Home Zone 2.0 services in a 2G or 3G network are doing so for three reasons:

  • Create service zone unique from the macro network with the goal of offering differentiated services/billing based on location.
  • Offload the macro RAN, especially for very high speed, media rich audio and video application.
  • Improve the performance, coverage and throughput rate for mobile data services when indoors/in the home zone.

These same service drivers apply directly, and in some cases are even more relevant, to an LTE RAN:

  • Create a unique service zone (still a competitive advantage)
  • Offload the macro RAN (definitely for the projected data rates)
  • Improve coverage/performance (certainly if LTE is at 2 gHz or higher)

As the market evolves, UMA will continue to evolve as well. Just as UMA started as a 2G technology and evolved to support the 3G/Iu interfaces, UMA will continue to evolve into key solution in an LTE-enabled network.

Thursday, April 24, 2008

Cisco to buy ip.Access? Better to wait...

The industry is buzzing today with the rumor that ip.Access has ‘won’ a contract with AT&T to supply femtocells. The reportedly $500,000,000 contract for 7,000,000 units gives you an ASP of $71.43/unit. Ouch.

What exactly did they win? The right to sell femtocells below cost to AT&T?

Frankly I think it's AT&T that is hoping Cisco buys ip.Access. At least Cisco can afford to lose money on every unit and ‘make it up in volume.’

The article notes that the contract calls for ip.Access to ‘eventually price the units at less than $100 apiece’. If half of the units are sold at $100 each, the other 50% need to be sold at $42.86 to get to an ASP of $71.43. Hummm… that’s probably why the article dryly notes the contract ‘could exceed $500m.’

My advice to Cisco: wait… buy ip.Access out of bankruptcy.

The best way to get costs down in the market: Open standards.

Wednesday, April 23, 2008

Is NSN looking to leave its femto 'partners' behind?

An article from DigiTimes today confirms that Nokia Siemens Networks is in talks with several Taiwan-based manufacturers over plans for joint production of femtocells.

NSN has already signed agreements with Ubiquisys, RadioFrame, Thomson and Airvana for femtocell access points. Yet the article quotes NSN general manager for Taiwan, Hong Kong and Macau Mike Wang, who confirms the company is now in talks with ‘other network equipment makers, including Gemtek Technology and Accton Technology.’

Why would NSN need to source their own femtocells? They have signed up four independent femtocell access point vendors. Why sign up Gemtek and Accton too?

Maybe NSN is getting ready to leave its partners behind.

[UPDATE: In feedback from this blog post, I received word that NSN is rumored to be courting femtocell silicon vendors directly, lending credibility to the idea that they are looking to develop their own femtocell.]

Tuesday, April 22, 2008

UMA: The reason behind the Orange/FT - TeliaSonera discussions

Early last week rumors began circulating that Orange/France Telecom may be interested in acquiring TeliaSonera. An article in the International Herald Tribune provided an analysis, but left an odd question unanswered at the end of the article:

“Why would France Telecom possibly want to make [another] European acquisition?”

As the article points out, there’s little geographic overlap, the logic for incumbents merging is ‘muddled’, and its outside France Telecom’s stated goals.

Finally on Friday, things began to clear up. France Telecom Finance Director Gervais Pellissier said that acquiring TeliaSonera could give FT the scale it needs to counter future competitive threats from companies such as Google and/or Apple.

“How do you expect us to develop the LiveBox if we don’t have several million customers using it? How can we push UMA if we don’t have that power?” asked Mr. Pellissier.

Great question! Pushing UMA is a TOP priority for Orange/FT right now and clearly it’s viewed as a technical competitive advantage. UMA gives Orange an advantage in its “Home Zone 2.0” strategy.

Thursday, April 17, 2008

Dreaming of dual-mode and UMA

It's good to know I'm not alone in the world. There are other technology geeks who dream about cool new stuff.

Check out the post:

I have similar dreams about UMA and dual-mode services. In my case, I actually use the HotSpot @Home service. I agree that not everyone has signed up with T-Mobile... yet.

Wednesday, April 16, 2008

A Good Week For Handsets (and it's only Wednesday!)

T-Mobile is starting to make good on their promise to bring 10 new HotSpot @Home handsets to market in 2008. Heck, two hit the rumor mill this week alone!

First is the new Samsung T339, a feature-rich mid-tier handset which continues to push Samsung into the upper echelons of DMH supplier.

Second, a rumor broke on HTC’s Shadow II (shown), the appropriately named follow-on to the Shadow currently available from T-Mobile. This is notable for being the first Windows Mobile device with UMA available from T-Mobile.

All this happens in the same week that the UMA-enabled Blackberry Pearl shows up on the shelves of T-Mobile stores nationwide.

T-Mobile is definitely on a roll.


It came to my attention just recently that a key proponent of the 3GPP Interworked Wireless LAN (I-WLAN) specification has passed.

I-WLAN is a 3GPP specification designed to enable ‘nomadic’ devices (e.g. laptops) to access services in the mobile core network over a Wi-Fi/IP connection.

I-WLAN provides a secure connection and authentication mechanism which is similar to UMA. The fundamental difference between I-WLAN and UMA is that I-WLAN lacks mobility for sessions whereas UMA provides seamless mobility for circuit and packet services between IP (Wi-Fi) and GSM networks.

Azaire Networks ( disappeared at the end of March, just before the CTIA show. While the I-WLAN specification will live on in the 3GPP, it is likely to become one of the hundreds (?) of specifications which are developed but never achieve any commercial success in the public network.

It’s not enough to have a good idea, technology needs to solve a fundamental business problem. UMA lets mobile operators extend their voice and data services over the low cost, high performance IP access network, and create Home Zone 2.0 services. UMA’s continued success is a direct correlation to demand from service providers.

Internet Offload and UMA

“Internet Offload” is the ability for a dual-mode phone to connect directly to the internet through the Wi-Fi radio. The phone which comes to most people’s mind is the iPhone, but Wi-Fi enabled HTC and Nokia E/N series support this feature too.

When the iPhone enters a Wi-Fi environment, the GSM radio stays on (for circuit voice and mobile packet data services), and the Wi-Fi radio can be used for accessing internet services like Google Maps directly.

For some reason, the market seems to think that a dual-mode phone can support either UMA or Internet Offload, but not both simultaneously.

This is not true!

Internet Offload and UMA are fully compatible functions of a dual-mode phone.

In fact the strongest example is the family of UMA-enabled Blackberries (8820, Curve/8320, Pearl/8120) which all support UMA and Internet Offload concurrently.

Here’s how it works:

When a UMA-enabled Blackberry enters Wi-Fi, the UMA tunnel is established. The UMA tunnel is used to transport services which reside in the mobile core network to the phone. The easiest example is voice services. But any data application which is delivered from the mobile core is accessed through the UMA tunnel, applications like MMS or any stream media services offered from the network (MobiTV,…)

A key advantage of using the UMA tunnel for service delivery is mobility. Any application or service which takes advantage of the UMA tunnel maintains session continuity between the Wi-Fi network and the outdoor macro network.

But it’s the way UMA-enabled Blackberries also support Internet Offload which is most interesting. When the phone tries to access web based services like Google Maps, the traffic is routed directly to the Wi-Fi access layer and out to the internet. No going through the UMA tunnel, no burdening the mobile core with non-revenue generating internet traffic. Note this is also how the Blackberry accesses mail services when in Wi-Fi, directly over the internet.

The trick is that the Blackberry maintains a basic router function in the handset. For application which are ‘hosted’ (circuit voice, SMS, MMS,…), the traffic is routed through the UMA tunnel and to the mobile core. For browser based applications, the traffic is routed directly to the internet.

Note there is no 'seamless mobility' for services routed via Internet Offload directly over Wi-Fi. When the phone moves outside of Wi-Fi coverage, the session ends. The session can be restarted over the macro RAN, but there will be a drop in service.

Why is Internet Offload + UMA important?

First, it’s important to clear up the misunderstanding in the industry. Early UMA phones didn’t support the Internet Offload feature because operators specified they wanted UMA-enabled dual-mode phones to work ‘exactly the same’ in Wi-Fi as in GSM. Therefore, all the packet traffic was to be routed via the UMA tunnel to the mobile core when on Wi-Fi, because all packet traffic is routed to the mobile core when on GSM. [It’s in this way that UMA is really a RAN technology, making Wi-Fi an access radio for mobile services.]

With dual-mode phones supporting Internet Offload on the market, operators are again saying they want UMA-enabled phones to operate ‘exactly the same’. But now this means adding support for Internet Offload concurrently with UMA by routing some traffic over the UMA tunnel and other traffic directly to the internet.

Second, support of Internet Offload is a critical component of making the mobile phone an integral member of the ‘connected home.’ There is a movement underway to bring household electronics together into a cohesive fabric. But a mobile phone without Internet Offload capabilities* will be excluded from supporting local connections in the home. Thus the phone will continue to be outside the ‘connected home’ vision.

By highlighting UMA and Internet Offload support concurrently in the same device, UMA devices can now easily fit into the vision of the ‘connected home’.

Now Nokia N series handsets can support UMA along with the existing Internet Offload capabilities. In fact, the Nokia N95 was one of the first handsets to demonstrate support for DLNA, a ‘connected home’ protocol standard.

[* - There are two ways to make the handset a member of the 'connected home'. First is to have a dual-mode Wi-Fi enabled handset with Internet Offload capabilities. The second, pioneered by the Femto Forum, is to proxy a standard 3G handset into the connected home network via a femtocell.]

Wednesday, April 09, 2008

Still on track for 5.5 million

My favorite pessimistic telecom commentator, Dean Bubley, has recently recapped his outlook on network evolution in his 500th blog post.

Given his general skepticism regarding UMA as a viable technology for dual-mode handset services, I suppose I’m honored that UMA was listed as the very first item he chose to list as ‘over-hyped’.

[Side note: considering UMAToday is the only site ‘hyping’ UMA, I hardly consider it ‘over’-hyped, certainly when compared with other things on the list like WiMAX. But I’ll take that as a compliment that we’re doing something right, thanks Dean.]

But given the comments, I think Dean's jumping the gun a bit:

UMA (unlicenced mobile access) - well, it still hasn't got much traction after another 30 months. Maybe 1.5m subscribers at Orange & T-Mobile US, plus some hopefulness around femtocells. My original forecasts (considered horribly pessimistic at the time) of 5.5m UMA homes at end-2009 now actually look insufficiently pessimistic after all.

First, the comment ‘after another 30 months’ seems a bit odd. Orange only launched 18 months ago (October 2006), and T-mobile just 9 months ago (June 2007). True, we’ve been talking about dual-mode services and UMA for some time, but there hasn’t even been one 30 month period, let alone ‘another’.

But second, I think the forecast Dean made relating to UMA-based dual-mode handsets might actually be one he gets right*. I happen to believe 5m units is within range for 2008. At the end of his prediction period (YE 09), I suspect there will be well over 10m units sold, easily meeting his prediction of 5.5m homes by YE 09.

Thanks for the vote of confidence Dean!

[* Dean is a big fan of the ‘naked SIP’ handset revolution, suggesting that this is the way, rather than UMA, that the market will embrace dual-mode services and VoIP.

In this model, subscribers buy/acquire a Nokia E or N series phone and download a third party SIP client onto the handset. The idea is to bypass those nasty mobile operators and make calls for ‘free’. It's the TruPhone model. Of course, there are several reasons why this won’t work:

  1. Spending $500 on a handset to make cheap calls doesn’t compute. Anyone shelling out $500 for a phone isn’t so concerned about saving pennies on phone calls that they’ll download a second client onto a phone.
  2. Putting a second client and a second phone number (the SIP service) on a handset is a hassle. Consumers and prosumers want a single number that’s available all the time, not just when the handset is connected to Wi-Fi.
  3. Oh, you want to run this SIP client over the 3G network as well as Wi-Fi? Round trip latencies for today’s 3G access network are 200 msec, and that’s just ‘your side’ of the call. It does not take into account the rest of distance those packets must travel. That’s an unacceptable delay for voice calls, even free.

For some reason, Dean’s prediction of 220m ‘naked SIP’ phones shipped in 2008 wasn’t captured in his list of ‘over-hyped’ technologies. I don’t honestly know how many ‘naked SIP’ phones shipped, let alone how many are actually use the application he describes. But I'm willing to guess it's a lot less than 22o million.]

Is 3x Enough?

There is still some misunderstanding on the benefits of dual-mode handsets and home zone services. Operators are struggling to find services which offer some type of benefit, either increasing ARPU or reducing churn. I’m not sure what the market expects, but I think a service that reduces churn by 3x is not bad.

According to a slide deck used in France Telecom’s Investor Days (December, 2007):

“Homes equipped with UNIK churn 3 times less on their mobile contract than others” [slide 20, file: 5-France.pdf]

Is there any other service that reduces churn by 3x? Note this claim is from the actual operator who is realizing the benefit. It is unfiltered, unbiased, and objective information.

Maybe just keeping subscribers on your network isn’t enough...

How about a service that causes new subscribers to churn to your network? In today’s hyper-competitive market, subscriber numbers are a zero sum game, for you to win, someone needs to be losing.

T-Mobile recently announced that nearly 50% of the users taking their HotSpot @Home service are new to the operator.

Frankly that’s a staggering statistic. People are churning to T-Mobile to get a dual-mode handset service. To put it into perspective, an early article on AT&T about the iPhone said that “roughly 40% of iPhone subscribers were new AT&T customers.”

So DMH out-churns the iPhone!?!?! Crazy.

I know iPhone and HotSpot@Home are different, but both are bringing new subscribers to the operator’s network. I haven’t seen churn numbers on the iPhone, presumably because the users are still locked into their 2 year contracts, but 3x reduction on DMH services isn’t bad.

I guess the point of this post is: If you don’t have a UMA-based dual-mode handset service to sell, maybe the iPhone is the next best thing?