Is Internet offload the next bottleneck for mobile broadband?

This article about H3G Austria implementing a "direct tunnel" for data caught my eye today. It fits into an emerging theme I've been seeing about the need to offload mobile data traffic from transiting core network elements unnecessarily. Given that some stats suggest that 95% of all 3G data is going to/from the Internet, it makes sense to "dump" as much of it straight onto a connection to the web rather than routing it through expensive (and capacity-limited) core nodes like SGSNs that add no value.

I heard much the same story about femtocells last week, and I've also heard it mentioned about UMA-type WiFi services, sometimes called "Split tunnel architecture". It makes particular sense for roaming data traffic from PCs, for which there is zero value in backhauling via the home network (and especially for H3G, which doesn't charge data roaming fees to on-net customers).

Now in theory, much of this capability to provide a "flattened" IP network architecture should arrive with LTE, and more specifically its counterpart the Evolved Packet Core (EPC, formerly SAE, System Architecture Evolution). But given the timelines, it makes sense for the more data-centric operators to move ahead sooner.

I have a strong suspicion that offload / direct or split tunnel / (assorted other similar terms) will become the next "big thing" after the current backhaul bottleneck is fixed for mobile broadband operators. There will probably be a few different architectures, which I guess will dovetail with specific operator instances of local (femto/WiFi) radio offload, and macro transport connectivity (owned / 3rd-party).

Operator-specific vs. Vanilla handsets

I just had an interesting chat with the sales guys in my local branch of Carphone Warehouse. I was inquiring about the new SonyEricsson C902 phone, and so asked them which networks it was available on. They replied that it was supplied in with either O2 or Orange contracts. (Interestingly, the CPW website also mentions T-Mobile, and it should be noted that the company no longer resells Vodafone).

Now, I'd recently tried out another phone which had a horribly slow and clunky operator-specific UI on top of the usual slick midrange SonyEricsson menu and app software. So I asked the CPW staff "Which version of the phone [Orange or O2] has the better software?". The reply surprised me "They're the same, we supply the phone with the generic software".

Now the reason for this can't be logistics (ie desire to avoid stocking separate versions of each phone) as I'll bet that each one still has physical modifications like operator logos. So I'm wondering instead if there has been some push-back from customers about operator-specific differences between phones. I've thought for a while that the handset review magazines and sites really ought to compare between operator variants of the same phone. And anecdotally, I'm certainly aware that some UK consumers are certainly aware that devices' capabilities and useability differ.

Obviously in instances where operators have exclusive rights to given handsets - or more-material customisations - such comparisons can't be done. And in markets like the US and Japan, it's quite common for many devices to be very-tightly specified by a single carrier.

But for popular devices like many Nokias and S-Es and Samsungs, do the mobile operators really want to compete on whose version of a given device is best? Obviously the quick answer is "Oh no, we'll compete on the services available on our version of the phones". While there may well be specific custom client software (to access custom services), this ignores the more in-your-face changes to top-level menus and UI that some carriers insist upon. It's minor things like populating the home screen with unchangeable links to stuff you don't want, or locking the browser's home page, that customers will be irked by.

There's also another question here - will the generic versions of phones supplied by CPW work well with all of each operator's existing service portfolios?

Carnival of Mobilists on LBS

While I'm turning into a big believer about location-based capabilities in mobile phones (GPS, maps etc), I'm much less convinced by the argument for monetisable location services.

I can quite imagine buying a Nokia device which comes with 4GB of global maps built in, plus GPS. I can quite imagine continuing to use Google Maps on mobile, with its clever autolocation capability.

But I can't imagine ever wanting "local search", or paying for "where is my nearest X", or trusting any "local" restaurant reviews where the venue has had to pay to be included in a directory. Possibly there may be some geo-specific advertising (step forward Google again), but in terms of actually getting money out of my own pocket for a billable service? No.

Nevertheless, there's a number of very good articles and different opinions courtest of this week's Carnival of the Mobilists, edited this week by Rudy de Waele. (It also references my post from last week about applications on smartphones).

Will broadband providers throttle femto backhaul traffic?

I've heard a couple of suggestions at the Avren Femtocell conference in London over the last 2 days that a key part of the business model is that "the customer pays for their own backhaul". I've also heard various suggestions that laptop 3G traffic (dongle or embedded) might be offloaded onto the femtocell & fixed broadband.

Where this falls down is in situations where the broadband provider and the femtocell provider are different companies. So perhaps Orange femtos linked to BT ADSL, or AT&T femtos on Comcast cable.

There are two issues here:

• Extra traffic load for the ISP
• Competitive or net-neutrality arguments about value transiting "my pipes".

(a third issue - how the mobile operator gets QoS on 3rd-party broadband, is also involved).

Now if the femto traffic is just "best efforts" mobile voice being offloaded, I wouldn't expect most of the broadband providers to worry about the incremental traffic load. But if it's 3GB a month of "dongle" traffic from a PC, going to a mobile broadband provider, that's a different issue. In that case, I'd expect to see traffic-shaping being used, perhaps throttling femto traffic to 100kbit/s or whatever level is applied to other bandwidth-hogging services like BitTorrent.

The competitive argument is a trickier one, and will vary according to markets. In countries where all the main operators have both fixed and mobile arms, any attempt by one to unilaterally block (or charge for) another's femto traffic could lead to a nuclear-style "mutually assured destruction" scenario.

But in other markets where there are dominant ADSL or cable providers that don't have mobile arms, I'd expect some to become pretty aggressive about their femto traffic-management policies. It's not as if it's hard to spot femto backhaul traffic on the network, as it goes to a fixed IP address of the gateway & will have a recognisable profile.

We may see some broadband ISPs attempt to negotiate with mobile operators to guarantee QoS on their lines, for a a price. Some may even return some proportion of cash to the end user, so they don't end up "paying the operator's backhaul costs".

Less than 1% 3G traffic from smartphones

A month back we got the stats from Finland that showed that 92% of 3G data traffic was generated by PCs, 4% from Symbian smartphones and 4% from all other sources.

I've got some more anecdotal data today from another advanced cellular operator (from someone who manages the 3G radio network). He said that his traffic was made up of:

- c70% HSDPA / HSUPA cards & USB dongles in PCs
- c20% older WCDMA 3G cards in PCs
- 8% from fixed-wireless 3G terminals (routers & wireless backup etc)
<1% from all handsets

Now presumably this excludes a bit of 2G data from Blackberries & unlocked iPhones and ordinary WAP phones, but it still points to the fact that from a radio network (and therefore capex investment & spectrum & dimensioning point of view), operators should care about 2 main things:

- Circuit voice & SMS traffic from phones
- Packet data traffic from PCs

I had an argument with vendor yesterday about whether this picture will change as people adopt more smartphones, but I really can't see it migrating much for several years, especially given that 3G dongles are still just at an early phase of growth. I reckon that a small number of PCs & iPhone-type devices generate enough traffic to absolutely swamp everything else for some considerable time.

I'll draw a graphic when I get a chance, but basically there's an inverted pyramid with 10% of devices generating 90% of the traffic - and I think that the skew is likely to get worse, before it maybe evens out in a few years' time with the adoption of more iPhone-grade devices.

Airvana touting Femto-aware devices

Just saw an interesting presentation from Airvana, which was talking about "hybrid networks", whereby both devices and macro networks are "aware" of femtocells and collaborate to maximise capacity and minimise interference.

It's the first public conference presentation I've seen (apart from my own) which has used the term "Femto-Aware Handset". The speaker was suggesting that these should become important after 2010.

However, while I'm impressed with Airvana's forethought about devices, I was conscious that the debate around femto-aware "phones" may get mired in discussions about interference management and 3GPP R8 tweaks to the interfaces involved.

But following on from my last post, I'm now convinced that the majority of data traffic on femtos will come from devices which also have WiFi. 3G dongles, iPhones, top-end smartphones, MIDs and so on.

Although the Airvana representative thought the balance would shift to non-WiFi devices, I disagree. While I don't think that more than a small fraction (maybe 10-20%) of mobile devices will have WiFi in them, I expect those to be the devices that will generate vastly disproportionate amounts of traffic - almost certainly above 80% of total 3G usage, and in many cases 90%+.

This will mean that simple changes to the handset chipset and protocol stack will not be enough. There will need to be higher-level software adaptations to manage "policy" - when to use femtos, when to use WiFi, when to use macro networks and so on. This will need to reside in the handset OS, probably in the Connection Manager. Some applications (eg operator, QoS-managed services) will work best on the femto, while others (eg straight-to-Internet vanilla access) will work better on WiFi. Both sorts might need to be run simultaneously on the same device.

This sort of issue needs to be considered now. I cover it in depth in my new report, and it's also something I've been discussing with advisory clients of mine. Please contact me on firstname.lastname AT disruptive-analysis.com if this area is of interest.

Femtocells and 3G dongles - a paradox?

I'm at the big Femto conference in London this morning. I'll write up more during the event, but a thought just struck me.

We now know that the largest amount of data on 3G networks emanates from PCs, not phones. Dongles are causing the macro load, plus a few high-end devices like iPhones. So it therefore follows that if femtos are being used for offloading data traffic from the (expensive & congested) macro network, then the majority of that offload data should also be PC/dongle-related.

But if the mobile broadband hype-merchants are to be believed, then 3G (or LTE or WiMAX) broadband will substitute for fixed broadband in the home. In which case, exactly what is the femto going to be attached to?

Or if someone does have ADSL or cable at home... what's the justification for the extra cost of a femto to connect the PC, rather than using cheap and simple (and faster) WiFi, integrated into the router or even wired ethernet? The devices generating all the data traffic in the home home - PCs, iPhones, top-end Windows Mobile or Symbian devices - all now have evolved & usable WiFi.

HSDPA on the fringes of coverage

There's a lot of talk about whether 3G dongles with flatrate data represent direct like-for-like competition against home DSL or cable broadband. My view is that at some levels this is true - especially for people unable or unwilling to take out long contracts for fixed-line connections, eg students or people moving home regularly. There are also segments of the market that want both fixed & mobile connections for different reasons, and a zillion sub-segments with unique requirements of their own.

That's all well and good, but discussion recently about fixed-line broadband (in the UK at least), has highlighted a lot of unhappiness about headline advertised data rates ("8MB Broadband!", "Superfast 20MB cable"), where the figures quoted (usually with "up to" in small print) don't match the reality. In those cases, it's usually a limitation imposed by the length/quality of the copper, so that if you're living 4km from the local exchange, you get a lousy speed. Although obviously there's a lot more to it around contention rates, backhaul and so on, there is at least a chance that you could be given a "guaranteed" data rate rather than a hypothetical one.

My experience with HSDPA recently has been a little patchy. In particular, I'm becoming increasingly aware of one of the limitations of the technology - the way it dishes out much more capacity to those people in the best radio conditions. It's not just that the signal degrades with distance / indoor penetration (ie the equivalent of DSL line length), but the fact that the system deliberately biases itself towards people with better radio reception.

This exacerbates the difference between the best and worst grades of service - to a greater degree than happens in the fixed world. On ADSL, you'd have to be pretty unlucky not to get 2-3Mbit/s on a supposed 8M service (ie maybe 70% less than headline rate). But on HSDPA, it's quite possible to get a 90% or more reduction in speed when you're on the fringes of coverage, versus when you have 5 bars on the signal meter.

I notice this regularly - I have 2 branches of Starbucks near me on Baker St in London, perhaps only 200-300 metres apart. One has seating in the basement, the other at ground level. Sitting in one of them, I'll regularly get 1Mbit/s+ with high reliability & low latency (this is on Hutchison 3, and I haven't got above 2Mbit/s at all). But in the other, I'm often lucky to get 100kbit/s, and often the network will just seem to 'hang' for 20-30 seconds. I'm not sure if they're both in the same cell, but the difference between 1-bar signal and 5-bars is at least a 10x change in throughput.

Does anyone else experience something similar? Or is this an issue that's going to cause customer service issues, or perhaps drive the need for femtos, or 900MHz 3G? In either case, it strikes me that any regulatory approach to selling "What you pay for is what you get" broadband is really going to struggle with mobile.

2 events next week: Femtocells & Billing/Pricing

I'm going to be participating in a couple of interesting industry events next week.

The first is the industry conference for Femtocells, run by Avren Events. I went to the one last year and it was literally standing room only. It's colocated with the Femto Forum Plenary so all the great & the good should be there.

I'm around on Tuesday & Wed, and will be taking part in a panel on low-cost femtocells. If you're there, I'll be easy to spot, as I'll be the one wearing body armour in case anyone violently disagrees with my stance on the need for femto-optimised handsets. Or indeed, if anyone takes exception to the fact my femto market forecasts are a lot less bullish than other analysts - I'm expecting a baseline of 8m shipments in 2012, with the chance of a bigger ramp-up thereafter if all goes well.


The other event is totally different, but also should be very interesting. Run by billing vendor Highdeal & moderated by myself, it's a roundtable on the evening of Weds 25th. It will be comparing Billing & Pricing strategies across multiple different industries - telecoms (especially mobile), finance, and transportation. Although these industries all overlap, there are some interesting parallels to be drawn, and lessons to be understood.

The event is mostly aimed at journalists & other analysts but others may attend as well. The invitation form & info is here.

Blending UnNP and femtocells

Interesting video demo of some femto applications here, courtesy of ip.access. They use a couple of different devices (Nokia N95 and a Samsung Windows Mobile smartphone) via the femto to control either a home media server, or a SlingBox.

One of the interesting aspects is the adaptation of UPnP to work over 3G (Universal Plug'n'Play, as specified as part of the DLNA standard for home networks). Having asked about some more details, I found it interesting that getting the UPnP demo working over 3G is not currently possible on the N95, as the client software in that phone assumes it can only work over WiFi.

Now, wouldn't it be useful to have a handset OS connection manager that could tell the application about the availability of a UPnP-suitable 3G connection (ie via the femto when the user is at home)? And wouldn't it be good to get UPnP support in non-WiFi handsets?

(Of course, if the phone has both WiFi and 3G, the "policy management" of in-home networking gets a bit more complicated as well).

These are some other types of subtle "femto optimisation" approaches to mobile phones. DLNA and full-featured connection managers are both covered in the "Femtocell-Aware Handsets Report", published yesterday by Disruptive Analysis. More details & contents pages are now available here.

Received party pays for mobile in Europe?

Some interesting comments being reported by various media this morning, suggesting that the European Commission might be open to the idea of some mobile operators moving to a "receiver-pays" model - perhaps akin to the US.

This would get around some of the complexities emerging around interconnect & call termination fees, as we move towards FMC and various hybrid technologies like femtocells & dual-mode WiFi.

Put simply, in the past in Europe there has been a fairly reasonable justification for higher fees charged by mobile operators to terminate inbound calls - the extra costs of spectrum and radio network. This has been reflected in separate number ranges for mobile, and regulator-specified termination charges that can be 10x that charged by fixed operators.

But now, the gap in termination *costs* to the operators is narrowing. As fixed calls move to mobile, the underlying expense in running the network is shared between fewer minutes of use, while the opposite is true in mobile. Further, some supposedly "mobile-terminated" calls actually terminate (via a mobile number) on WiFi, femtocells, fixed-VoIP softphones, or voicemail boxes. There is no reason for the termination fees in these cases to be the same as that terminated on the expensive cellular macro network.

Unsurprisingly, the mobile operators are a bit unhappy about this, with various comments about how this might "distort the market" or "confuse consumers".

But by a strange twist of irony, I've been presented this morning by the most egregious "distortion" of the current caller-pays/mobile-termination fee model. I am participating in a 7-way conference call tomorrow, hosted by a mobile operator. There's only one participant from the operator involved, and some of the other participants will probably be outside the UK. Yet the dial-in number is a UK mobile number (obviously, one of the operator's own). Not a freephone number, not a fixed number that can be called at fractions of a penny per minute, not even a list of international mobile dial-in numbers. So 6 people will each be paying extortionate amounts against one mobile-terminated leg of the call. Cost-based? Yeah, right.

That sort of abuse of mobile numbering has got to be the best single argument for receiving-party-pays I've come across yet. Maybe the European Commission's got this one right. Or maybe I should get myself an 09xxx premium-rate number and insist that any future briefing calls from mobile operators have to use that.

Edit: apparently the mobile dial-in conference call service is being positioned as a value-add service, along the lines of the fixed-line "free to set up" conference services that use semi-premium number ranges like 0870 in the UK. That's sort of fair enough for occasional ad-hoc confcalls among friends or small businesses, but I would have thought that a large company (especially a telecom company!) would be able to have a cheaper & more efficient platform based either "in the cloud" or (shhh!) on a PBX. And even the 0870 conference guys have options for international dial-ins or even Skype access.

New Report: The Case For Femtocell-Aware Handsets

I mentioned a couple of weeks ago that I was working on a report assessing whether or not femtocells might, in fact, require the use of revised mobile phones.

The full research report is now available from Disruptive Analysis, titled "Femtocell-aware Mobile Handsets: Will cellphones need to be optimised to support future massmarket femtocell business models?"

The report is an in-depth analysis of what should be changed in handsets, to make them perform better with femtos. Please contact: information AT@ disruptive-analysis.com for more details. (Journalist enquiries are also welcome). A table of contents is here.

Before I get a hitman from the Femto Forum knocking on my door, I should be clear upfront that for early, basic deployments of femtos, handset upgrades shouldn't be needed (as long as the user's existing 'legacy' phones are 3G-capable).

My view is that the femtocell industry is currently focusing on the short term – getting initial trials in place, developing standards, and securing commitments for early commercial deployment. These are all critical to validate the market, raise the profile of the femto concept, and stimulate finance and investment. One of the central marketing propositions is that femtocells can work with normal 3G handsets.

But while focus is good – and the industry does not want unnecessary distractions – there is a risk of medium-term disappointment if certain future problems are not addressed early enough, even if this muddies the waters of the short-term marketing message. Handset development cycles are so long that works needs to start now. Already, femto proponents are talking up massmarket business models that go beyond simple indoor coverage and macro-network offload. They are talking about 10’s of millions of subscribers, and new “in-home” services for users, that exploit fast and cheap local mobile connectivity.

(Disruptive Analysis' forecasts in the report are less bullish than some - an installed base of around 30m femtos at end-2013)

As the market gets beyond "Femto 1.0" business models, handset innovations become more important. In part, this relates to complexities in managing the radio environment and mobility between femtocell and macrocell networks. Various optimisations are desirable, especially when dense deployments of femtos occurs. These drive changes in areas such as the way the phone “selects” cells on which to register. There may also need to be ways to offer provisioning and “guest access” on femtocells. Some of these changes to handsets are being enshrined in 3GPP's Release 8 specifications for what it calls "Home NodeB's".

But the medium-term hopes of the industry are also that people will use their cellphones differently when in range of femtos. There will be different applications and different behavioural patterns when people are at home – perhaps content backups, podcasts or even advertiser-sponsored TV programming. The mobile phone may need to linked to TV, PC, HiFi or other items of domestic technology.

This report argues that if the phone will be used differently, it needs to be designed differently as well. Standard phones can work with femtocells, but they are not optimised. Certain applications may only work when the phone is within femto range – but they need to know when that is. Yes, some services can be notified by the core network that the user is “at home”, but that approach doesn’t scale to a wide base of operators, application developers and handset/OS vendors. The phone needs to be “aware” of the femtocell, ideally both in the radio and the application platform. Elements like the IP networking "connection manager" in the phone are key. Other aspects like memory allocation and power management may need to be revisited too - even perhaps the phsyical design of the device.

Changing such elements is not quick. The handset industry is much more complex and slow-moving than many in the wider wireless business understand. The mobile industry habitually trips up on these issues. Phones - both hardware and software - are extremely difficult, time-consuming and expensive to get right. It often takes 2-3 years for changes in handset architecture to reach commercially-sold handsets, and another 2-3 years to reach a broad range of devices and reasonable penetration within the user base. And there are huge challenges in getting the user-facing applications and useability right - the network interfaces are much easier to standardise and develop than the user interface.

In recent years, other mobile network innovations such as UMA and IMS have faced delay & lack of market acceptance, in part because of a lack of early focus on the practicalities of getting handsets ready. The fact that the network protocols are in place does not mean the whole proposition works in the hands of the user. I remember talking to an IMS advocate a couple of years ago - and pointing out that many of his suggested use cases for IMS implied a multitasking OS on the phone - something that even today is far from ubiquitous.

Disruptive Analysis believes that the femtocell industry needs to be much more open-minded about the need for modifying and optimising handsets – and to be alert to the huge time and effort it will take to achieve. It is necessary to look at everything from physical design, to network protocols, to testing, to the phone OS, to femto-specific applications. Although many femto advocates fear distractions could delay immediate market acceptance, early consideration of these “2nd order” problems is necessary for longer-term success.

The report predicts a demand for at least 50m femto-optimised phones by 2013. And that number could be much higher still (100m+) if there is widespread use of "shared" femtocells, or if some of the more optimistic observers' baseline femto shipment forecasts hold true.

The "Femtocell-Aware Mobile Handsets" report is about 85 pages in length, and is priced starting from £1500 / US$3000 / €1950 for a 1-5 user licence. For sales and additional details, please contact: information AT disruptive-analysis.com

Hotel WiFi - a tale of two conferences

I was at two conferences last week - the Informa Handset World event in Berlin, and the Osney Media Mobile Web 2.0 show in London. I'll comment on the two events' outcomes in separate posts, but one quick observation:

• The Berlin event was held at the Berlin Palace Hotel. There was no delegate WiFi, and Swisscom charged a hideous €22 per day for its supposed "business grade" service. It was slow and unreliable - lots of "404" errors on normal website like Yahoo & the BBC - (I think the DNS server was based on an abacus). It needed me to reboot my PC after I moved between conference rooms. and access points. As a joke, a slightly cheaper €17 "economy" 256kbit/s version was available, although without signing up it wasn't clear exactly what it blocked - eg Skype, IM, email attachements etc. Unsurprisingly, given it was Swisscom, there wasn't an economic option for Informa to have provided delegate access for free.
• The London event was run at the very impressive Mayfair Hotel - which is part of the Radisson chain. And which provides free WiFi at all its European properties. It worked flawlessly.

I pity hoteliers who decided, maybe 4-5 years ago, and before the growth of WiFi, to sign up for long exclusive contracts - especially with Swisscom, which I think I'd single out as the greatest villain of the European WiFi industry. It's experiences like mine that will make people stop using WiFi in favour of 3G modems, as roaming prices come down.

I'd exhort hotel chains to look at renegotiating their WiFi contracts, as otherwise it will start to cost them lucrative conference business. And I'd recommend to the cellular operators that they start renting out 3G-backhauled WiFi AP's to conference organisers and hotels.

I'd recommend that conference organisers specifically try to avoid venues with "legacy" WiFi providers that have locked in the hapless site owners to 10-year contracts.

And lastly I'd recommend to enterprises that they watch out for 3G modem tariff plans that give free or flatrate roaming (Voda has one, and 3's dongles are totally free on its partner networks). A recent trip to Stockholm cost me the princely sum of £zero for 2 day's fast access - and has earned 3 UK a boatload of referrals from me for new customers.

It's a shame really, as WiFi when done right is excellent for laptop connectivity. But €22 per day for a lousy service makes a mockery of the whole industry, and makes some recent predictions of its demise as a public service (which I generally disagree with) more likely to become reality.

Apps on smartphones = Geek. Apps on Facebook = Fun. Why?

I'm listened to a panel session at Handsets World last week about device personalisation. One of the panellists was convinced that it's all about adding applications to make the phone "your own".

However this runs completely counter to most "Normobs"' worldview. Normal people think that downloading anything to a phone (except a ringtone or Java game) is utterly geeky. Yes, among small groups of people this has changed somewhat - perhaps iPhone owners (although many of those are geeks anyway), or teenagers using a Java messaging app to avoid paying for SMS. But mainstream users? No way.

But another speaker made the very insightful point that many people who would recoil in horror at installing smartphone apps are entirely happy to stuff their FaceBook pages with all sorts of random plug-ins.


It made me think a bit about what the differences are behind this divergent attitudes.

The most obvious is platform. People have grown up installing applications on PCs, from the minute they take them out the box, through to the familiarity of the Installshield process and frequent updates to OS and various client applications. Conversely, people have grown up using phones, as, well, phones, or maybe MP3 players or cameras. It takes a long time for attitudes to change - and especially when people don't think of things as "programmable". There's no reason you couldn't install aps on a camera, or a car, or even a toaster, but it would feel weird. People don't think of these items as using software at all. I remember the weird looks I got when I told people that my car's electric windows had a software bug that meant it had to go back to the dealer.

The second difference is in virality. On mobile, you can't really recommend installable apps to friends or acquaintances. You don't know what phone they have, if there are any operator customisations to the UI, or even whether their networks' policies permit the app's use. It's no use saying "everyone with an iPhone / Symbian / Java MIDP3 device can get it" - most people neither know nor care what devices & contracts their friends have. On FaceBook, or on a PC generally, you can be pretty sure that everyone you forward a recommendation to can benefit from it. Until there's a reliable "vector" for viral adoption of mobile apps, it'll remain a minority sport for enthusiasts only.

Thirdly - and specifically with regard to FaceBook - it's a web service, not a downloaded app. Yes, you have to agree to assorted permissions and access privileges. But there's no "are you sure you want to install this?" trauma. It's interesting to see the push towards browser-based web services and platforms like Nokia Ovi, which should take the pain out of driving handset usage by avoiding the need to install device-side software.

My question to the Personalisation panel was about the realities that I see in the marketplace, rather than wishful thinking. To my mind personalisation is still about exterior of the device (shape, colour, style) and also the combinations of multiple devices. Having a red BlackBerry as well as a black N95 says more about you than any amount of software installed on an iPhone.

Amusingly, as I got off the plane on the way back from Berlin, I saw a girl in the row behind me - she had two (inexpensive) pink phones - a Moto KRZR and a small Samsung, I think. And a pink portable CD player. Now that's personalised.

Tip of the hat to Carnival of the Mobilists

Thanks to Bena Roberts for linking through to my 900MHz refarming article from last week.

Some other interesting posts on the list too. This one is good on sources of mobile stats.

Although I vehemently disagree with the "local search" concept reheated on this post. No, I don't ever foresee asking my phone "what pizza restaurants are nearby?". As I've said before, mobile search is a red herring. Search on mobile, yes, but local search or this fatuous notion that "mobile users want to find, not search". Please. It's just wishful thinking from people jealous of Google.

iPhone 3G - what's Apple got against cameras?

Has Steve Jobs lost the plot? Although details are vague, it seems like the new iPhone has the same sub-grade 2MP camera as the original one. That was already obsolete a year ago, when most top-end phones already had 3MP cameras. Now, most mid-top tier devices have 5MP and many have xenon flashes.

This isn't just me talking either - a Nokia keynote speaker at the Handset World conference this morning from singled out camera capability as *the* most important purchase decision criterion for many customers.

GPS is a nice-to-have. WiFi is a nice-to-have. A decent camera is a must-have.

So, I've reluctantly already added the device to my list of also-rans of phones that look good, but fall down on a deal-breaker element. It doesn't make the grade as a potential day-to-day "personal" phone for me, although it might be good business device.

At Handsets World & Mobile Web 2.0 this week

I'm at events in Berlin & London this week - Mon-Weds at Informa's Handset World, where I'm chairing a session on Handset IP Connectivity, Power Management & Storage. And I'll be dropping a few hints about the femto-aware handset issue I mentioned the other day. And on Thurs I'm at Osney Media's Mobile Web 2.0 Summit.

I should be pretty vocal & visible at both, so please come & say "hi" if you're there.

So... you think that femtocell users won't need new handsets?

There's been much joy in the femtocell industry this week, as 3GPP has decided on a standard architecture for connecting HSPA femtos (they call them Home NodeB's) into the network. This should make life an awful lot easier, given that there have been assorted proprietary and semi-standard approaches suggested up until now. The consensus among operators is that standardisation is mandatory for widespread large-scale deployments.

The output is a hybrid of the UMA/GAN approach advocated by Kineto, and some of the other "RAN Gateway" approaches advocated by the likes of ip.access, NSN and others. It's called the Iuh interface, and it's pretty much kicked the IMS/SIP approach into the long grass for the medium term, at least for 3GPP networks - although that architecture is more likely to be adopted by CDMA operators.

This is in the latest version of 3GPP's specifications, Release 8.

So far, so good for the femtocell proposition.

But unfortunately, the picture is likely to get a little murkier. I've been working for some time on what's going to be a rather controversial report. This post is a bit of a teaser before its publication in the next week or two.

In a nutshell, I think that one of the fundamental assumptions about femtocell business models has serious flaws.

Although existing "legacy" 3G handsets can work with femtocells, they are not optimised for them. There are some enhancements also likely to be in 3GPP Release 8.... and quite a lot of other considerations outside R8's scope that the vendors, and the standards and requirement-setting bodies haven't considered yet.

I have.

If you're interested in getting a heads-up on "The Argument for Femtocell-Optimised Handsets", keep an eye on this blog. And if you pre-register your interest via email at information AT disruptive-analysis DOT com, I'll let you know as soon as the full report is published.

Handsets, handsets everywhere. Yet not a single one I want.

Maybe June will be the month I finally replace my main personal phone. I've been utterly exasperated over the last 11 months that there has not been a single phone that's ticked all the boxes. I'm still using the SonyEricsson K800i I got from O2 almost two whole years ago (despite being on a 12 month contract....).

Sure, lots of phones have come out in the meantime that are superficially attractive, but they've all fallen down on one ground or another.

My wishlist for a new personal handset (not a work device):

5MP cameraphone with a proper flash
HSDPA
GPS (this has come up my list of wants recently. Not sure about aGPS though)
Looks quite cool aesthetically
Comfortable to hold & use
Decent browser
No horrible operator-customised menus
Preferably candy-bar, or maybe slider
Preferably WLAN - although it's got to be fully open if it's there
Preferably no full OS like Symbian or Windows (Fine for a work device but not personal)
Preferably keypad (not QWERTY, and only a touchscreen if it's excellent)

Extras like big memory & high-end music player are an added bonus.

Phones I'd thought about all lost out one way or another:

SonyEricsson K850i - nasty semi-touchpad screen for softkeys, no GPS
iPhone -2MP camera, no GPS, no 3G (I'd probably live with the touchscreen) & too expensive
Nokia N95 - clunky and I really don't like S60 on my main device
Nokia N82 - huge, S60 and awful keypad
Nokia N78 - OK size, but even worse keypad & 3MP camera
LG Viewty - no keypad, no GPS
Assorted HTCs - OK as a work device, but no 5MP cameras & I don't want Windows on my personal phone
Samsung Soul - getting closer but no GPS
LG Secret - same story as the Soul
Motorola - not been tempted by Z8 or Z10
Blackberries - camera limitations & operator-locked WiFi and connectivity

But maybe, just maybe I'll find something over the next month or so, although I think I might have to compromise a bit. It's starting to get embarassing going to mobile conferences with a battered-looking 2-year old phone.....

SonyEricsson C902 - looks OK but has no GPS
SonyEricsson C702 - aGPS but only 3MP camera
Nokia N96 - maybe I'll grit my teeth over the UI. Looks a lot better than the N95.
HTC Diamond - maybe I'll grit my teeth again... the UI looks better than normal WinMob
Nokia 6220 Classic - looks promising

.... although if some of the rumours are true, I might end up with a 3G iPhone.

Or maybe I'll just keep the K800i for another year.....

UMTS at 900MHz becoming a reality

With all the recent focus on the US 700MHz auctions, and the ongoing 2.5/2.6GHz band auctions, it is worth catching up with another less glamorous trend in spectrum (OK, I'll admit that radio frequencies aren't that glamorous to begin with, but you'll get the point).

According to the GSA there are now 3 active UMTS900 networks, and 7 capable devices. That's Elisa in both Finland and Estonia, and now also AIS in Thailand. Various other operators have announced intentions to roll out the technology, and there are ongoing consultations in many parts of Europe.

(The UK's is mired in dispute - unsurprising with 2 incumbent GSM900 licence holders and 3 more large 1800/2100 operators each hoping to get a slice. It's a bit of a nightmare for the industry to sort out.... and made even more complex because of T-Mobile's reported threatened legal action to delay the 2.6GHz auction until it's resolved).

Interestingly, some consulting work I did about a year ago basically suggested that most handset manufacturers and their silicon/RF suppliers thought that UMTS900 was an inevitable "done deal", while only few had progressed that far on UMTS2600/LTE2600.

UMTS900 is ideal for operators needing to meet 3G coverage obligations, and it should also get much better indoor penetration than 2.1GHz. Currently announced devices with UMTS900 support include Nokia 6121 and N96, the HTC Diamond and assorted others. Interestingly, it seems to be being supported in quite a lot of phones first rather than just modems/dongles - reflecting its suitability for coverage rather than capacity.

It will also be very interesting to see whether operators deploying UMTS900 are as aggressive on alternative options like WiFI dual-mode and femtocells for enhancing indoor coverage.

DoCoMo dual-mode consumer VoWLAN launches

DoCoMo actually launched the first operator-based dual-mode WiFi/cellular service several years ago in 2004 - based on SIP and aimed at enterprises. It was called Passage Duple and was based on the NEC N901iL handset, and has subsequently been followed up with a replacement N902iL.

However, Japan has been pretty slow about extending dual-mode into the consumer arena. This changes with the introduction of the new N906iL "onefone" and a new service called Home U. It's not immediately obvious what the architecture behind it is, but given the previous SIP approach, I suspect it's some sort of pre-standard VCC rather than UMA.

Extending mobile presence - the need for context and device state information

I speak to a broad variety of vendors and operators about presence in a mobile context. It crops up repeatedly both for operator-centric developments like IMS and RCS, and also for independent applications like mobile VoIP and corporate unified comms.

In theory, mobile phones ought to be good platforms for presence information - both rudimentary connection details (online, offline, idle etc) and some personalised state or emotional content as well ("Dean is currently thinking about lunch" and so on).

In addition, the network is sometimes able to add some extra information about the user - maybe location cell ID, whether the user is on a call, whether they are connected via fixed or mobile and so forth.

All this is useful, and currently highly under-exploited. There is a lot of work needed just to prove the current user experience and gain adoption. Although I can't see anyone actually paying for presence directly, it potentially enhances other services, and could also generate increased traffic and call completion.

But at the same time, I think the mobile presence industry is missing a trick. The presence client on a phone ought to be able to pick up a lot more information about the device's - and the user's - context. In particular, it would be fantastic if it could pick up details from the underlying phone APIs, and enable that data to be exploited by clever applications in the network, or directly by a contact. So for example - whether the phone is on charge, or is running low on battery. Or whether the user is using a Bluetooth headset. Or that the accelerometer can detect motion that looks like walking or driving. Or that the memory for SMS's is almost full.

These are just some ideas, but you get the picture. There are all sorts of clever things that could be achieved with this approach.

(Yes, clearly privacy is a huge concern here, but let's assume that we can simultaneously invent some form of reasonably-effective permissioning system - perhaps an easier version of that seen on sites like Facebook)

Now clearly a fundamental issue is that phones differ in many of these regards. I don't think there's a standardised API for % battery remaining for example, and the SMS memory may be split between phone, SIM and memory card. Different OS's and models of phone have very different capabilities which would make creating "common denominators" extremely difficult.

But maybe we could start with a couple of basic aspects - perhaps standardised by OMTP or OMA - and then incrementally add more features over time.

Unfortunately, there seems to be no easy way to catalyse development of this type of concept. Bodies like the 3GPP shy away from dependency on mobile phone-resident software, often naively believing that everything of value can be done from the network core. It may be that we will need to rely either on operator-customised handset platforms (DoCoMo, maybe?) , or third-party software providers, probably from VoIP or IM backgrounds (Truphone? Skype? social networking clients?). We're getting there first with location context (Loopt, fring and others do this), but that still not providing data at the level of "device state".

On the other hand, one of the downsides of any extensions to current mobile presence technology would be the amount of extra signalling traffic involved. I'm already hearing that the IMS-centric view of a "presence enabled phonebook" is running into scaling problems. If you have 100 entries in a contact list, and you expect all of them to be continually updated, this causes a huge amount of (typically SIP) traffic. It also requires presence clients to be running continually in the background on devices, with consequent implications for power consumption.