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

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

Tuesday, August 29, 2006

T-Mobile UMA.... the worst-kept secret.... plus other UMA bits

Right, back from holiday & onward with matters FMC, VoWLAN, IMS and so on. Already deluged by assorted PR people wanting to tell me about "A new mobile VoIP company offering something unique". So, 37 types of Mobile VoIP uniqueness. Wow.

While I was away, I noticed that people continue to be "surprised" about T-Mobile's imminent launch of a UMA service, with details about its super-secret range of phones. This follows on the heels of some of the beta testers breaking the terms of their NDA earlier last month.

Well, in that case, given that everyone's been amazed that the service will offer the Nokia 6136 as well as the Samsung T-whatever, perhaps I can ensure myself fame & fortune by predicting an appearance by the Motorola A910 as well.

Of course, this is all due to my amazing insider knowledge & crystal ball-reading skills, and nothing whatsoever to do with the photo of the Moto device, along with the Nokia & Samsung ones, from Slide 11 of this presentation given by a T-Mobile exec 8 months ago.

Speaking of UMA, I've been invited back to the lion's den of the UMA conference circuit this week, for an event in London. It will be interesting to see if the tone has changed since the previous conference in Barcelona a few months back.

Also, Kineto Wireless announced support for 3G UMA devices last week, with commentary that it satisfies a clear market demand (which I commented on more than a year ago). At present, operators are stuck with a tricky decision for market segmentation - do we upgrade our best customers to 3G phones? or UMA ones with an FMC service bundle?

What's not obvious from the new protocol stack solution is (a) when phones will actually ship, and (b) whether the phones can actually support 3G-over-UMA/GAN, or whether the device reverts to GPRS-over-UMA when it's indoors. If the latter is the case, although you could ratchet up the GPRS bandwidth to decent speed via WiFi, it would still rely on a 2G core network, with the limitations on applications, latency and so on.

Thursday, August 24, 2006

On holiday.....

Disruptive Wireless is taking a much-needed break for a week or so.....

But a quick anecdote.... just spotted an advertising hoarding for LG's new 5 Megapixel cameraphone. Where am I? Seoul? Helsinki? Singapore? No...... Kiev in Ukraine.

It's interesting - the shops here are full of the absolute newest handsets - Nokia N-Series, assorted high-end Samsungs and so on. It was much the same deal when I was in Mozambique's capital Maputo at Xmas. I guess the handset manufacturers must realise that there's an affluent elite in nearly every country that absolutely, positively, must have the newest (or most expensive) phone.

Friday, August 18, 2006

Another IMS handset software firm gets acquired.....

It looks like the days of independent IMS handset software client specialists may well be numbered.... after the acquisition of Openera by NMS Communications earlier this year, Qualphone has now been acquired by Qualcomm. Qualphone has worked with Nortel and IBM in the past, including participation in an IBM-led trial with Swisscom Mobile on IMS. It has also worked with Infineon to embed PTT functionality into its APOXI handset software framework.

Even more surprising is the low prices paid - just $16m for Openera, and now $18m for Qualphone. Given that these were both quite high-profile specialist players (and in Qualphone's case also had a separate & established business in mobile testing), these are pretty low-ball sums. In what ought to be a rapid-growth field (enabling the creation of IMS handsets), it strikes me as odd that they didn't hang on for a higher price at a later date.

Unless, of course, VCs were unwilling to stump up for another round of investment.....

.... which ties in with my general views that IMS-enabled phones are going to be late in reaching the market, and that IMS handset software vendors are going to struggle, as there are no detailed standards or even generic operator requirements to work against yet. In other words, royalties for "full IMS" software clients on mobile phones won't start to ramp up until 2008-9, and even then it is probable that firms will need huge additional resources in integration and customisation.

So who's next for the IMS phone M&A mill? Ecrio? Movial? Nable? If I was a player in this space, I'd start thinking more about all the other clever things that could be done with SIP-enabled phones apart from IMS applications, and which won't be tied to the glacial pace of IMS handset standards development.

(For more details on IMS-capable handsets, see here for details of Disruptive Analysis' recent report on the topic, authored by yours truly)

Once and for all..... fixed broadband will always be faster / better / cheaper than wireless

I still get people trying to convince me that HSDPA, EV-DO, WiMAX or whatever will close the gap between fixed and mobile broadband speeds.

And I still believe it's arrant nonsense. Copper & fibre are cheaper than spectrum, and unless you live somewhere a long way from the nearest exchange or cable headend, they always will be.

In fact, I firmly believe the gap is broadening, not narrowing. And this post from James Enck over at EuroTelcoBlog underlines the point.

Yes, that's €7 per month for symmetrical 30Mbit/s .

Stick that in your 4G pipe and smoke it.

Thursday, August 17, 2006

Mobile Java being open-sourced - just in the nick of time

According to a few things I've read it appears that Sun is pushing Java ME (formerly J2ME) towards being Open Source.

If so, this is extremely important in terms of handset software going forward. Most mid-to-high end GSM phones feature a Java Virtual Machine, typically used for games, and a few additional applications like dynamic UIs. Generally, Java has been a "limited success" - generating revenues & establishing an ecosystem, but without much success in extending beyond a confined games-oriented domain. At one point in time, various companies were talking up full Java-based operating systems (Coincidence: I got a PR email from SavaJe as I wrote this post), but these have not really got any traction.

An extremely sore point among handset manufacturers has been the need to pay a royalty to Sun of (historically around $0.50-1.00 per phone, I believe), on top of the cost of the JVM itself. While this doesn't seem like much, it is a significant fraction of a phone's total "bill of materials" and hence a major impact on the price and achievable margin on the device. In addition, the fragmentation of the way in which J2ME is actually implemented on phones has landed developers, operators, and device vendors with myriad headaches in porting software across a whole range of handsets. Companies like Tira Wireless have closed the gap with specialised porting tools, but it's still been a bone of contention.

Java's future on the handset as the "3rd party software platform of choice" has therefore been threatened - while it still has a huge developer base, it has been faced mounting competition from other software platforms that can support games, and often much more. Obviously, native Symbian, Windows and Palm applications are alternatives on smartphones, with shipments growing and OS licence fees falling. Linux has also been making a strong push recently, especially with the introduction of a secure "application sandbox" in implementations from TrollTech, although the number of mobile Linux developers remain quite small, especially for games. It's also increasingly possible to run software applications like games inside a handset's browser (eg using JavaScript/XML like OpenWave's MIDAS) or in an abstracted "user experience" layer like Adobe/Macromedia Flash Lite, or solutions from SurfKitchen & its peers.

This move may particularly benefit Motorola, which has been focused on hybrid Java/Linux software platforms for some time. This article quotes a senior Moto exec gushing enthusiastically about the move. I think it's a negative for TrollTech, which is aggressively trying to evangelise the "pure Linux" approach to phones & sway developers away from other platforms.

Overall, I think this move has the potential to entrench Java more firmly in future handsets at a time when its future was looking shaky - especially as this is a critical time as the industry moves towards more sophisticated handset platforms employing IP/IMS connectivity, multi-tasking and web-type functions.

T-Com's dual-mode offer.... actually, I think it's on the right track

I noticed a couple of posts from Andy Abramson and Richard Statsny about the new FMC dual-mode offer from T-Com in Germany, based on their custom-made TC300 handset.

Richard & Andy are pretty ambivalent about the proposition, referring to it as "fake FMC" because it uses call forwarding, and essentially combines two separate services (a mobile phone, and a fixed-VoIP connection with a "cordless" handset) in one device, rather than a one-number, one-service device of the type proposed by UMA enthusiasts.

I'll take a contrarian stance. I actually think this (along with NeufCegetel's attempt in France) is closer to the leading edge of FMC/dual-mode than some other propositions. It's about taking a home broadband service + VoIP + SIP, and leveraging the handset's form-factor by adding on cellular, rather than vice-versa (taking a cellular service, and leveraging broadband by adding on cheap local WLAN connectivity).

It uses two numbers rather than one - a distinct advantage in markets like Germany where there is a very low penetration of outbound cellular calls vs fixed, and very high fixed-to-mobile termination fees. Nobody wants to call from a fixed-line phone to their friend's mobile number at an extortionate rate per minute - they'd much rather call a fixed-line number. This is why the established German "Cellular HomeZone" services, like O2 Genion and Vodafone ZuHause also feature dual numbers.

The device itself is quite sleek - certainly among the nicest/smallest dual-mode phones out there. I understand that T-Com actually got a proper industrial designer to create a "want one" shape & look, and it shows. The corollary is that it's too small/light to feature a big battery, so I can quite believe that using WiFi drains the battery. (Someone told me the handset is made by FoxConn in Taiwan, but I haven't had confirmation of this).

It doesn't feature "seamless roaming". Big deal. This is one of those things that gets attention FAR beyond its actual importance. People will use it on what, 5% of calls? At best, it's a nice-to-have. At worst, it's a distraction. Sure, in the long term it's important, but for early dual-mode services, it should be priority #17.

In my view, T-Com has made a sensible decision to prioritise the introduction of SIP, rather than get sucked into the "roaming by any means" philosophy by hitching themselves to the dead-but-still-warm UMA.

Speaking to T-Com, my understanding is that this is definitely pitched as an initial, basic, entry-level phone. It's a "new" WLAN home cordless phone, which also has a cellular connection. Coming from the fixed/IP side of the Deutsche Telekom, it has good WLAN pedigree features like WPA2. No, it doesn't have whizzy mobile-phone colour & camera & multimedia.... but who cares? It looks quite cool, and that makes up for its lacking megapixels. There will apparently be follow-ons using handsets from Nokia's N-Series or others coming in the future.

Even more clever is T-Com's ability to sell the product into non-broadband homes. It can act just as a "WLAN cordless phone" attached to an ordinary PSTN (or ISDN, this is Germany...) line, again doubling as a basic cellphone when out & about.

Instead of hooking everything back to a mobile HLR & seeing the WLAN as a cellular "extension", it uses a mobility management function located in the T-Com fixed network, almost using the T-Mobile network as a sort of "internal MVNO" rather than having a fully-integrated network core. It's a very fixed-network operator view of FMC, and one that I think makes a lot of sense. To my mind, FMC is a great tool for the more aggressive fixed/IP/broadband/VoIP service providers looking to re-substitute minutes from the lazier, legacy, IP-phobic mobile operators.

Overall, I rate this pretty highly as a first attempt at FMC. It will be interesting to see what T-Com's 2nd and 3rd iterations are like, especially as Deutsche Telekom moves to a converged core network over time.

Wednesday, August 16, 2006

Multiplicity rules in mobile

A theme I've mentioned before is that of "Multiplicity". (I also sometimes call it "Divergence", largely to annoy the more religious convergence zealots). I'm increasingly convinced that this is an important future trend, and that vendors and operators need to understand its implications. This is a corollary of the segmentation strategies employed by operators & manufacturers - the more tightly segments are defined, the more likely that any individual will associate with two or more of them. (So, for example, you might be a music-lover with additional obsessions for checking work email

A couple of conversations I had yesterday with vendors point to the fact that this is (slowly) becoming accepted, with functions such as multi-device management being integrated into their boxes' capabilities. Chipset, operating-system and software client vendors are also talking up their multi-platform capabilities (mobile phone, VoIP phone, broadband modem etc). I'm also seeing the first signs that the obsession with Powerpoint slides featuring "Ones" (one device, one number, one bill, one identity, one operator etc) has peaked.

This morning, Multiplicity has even become visibly trendy.....

Nevertheless, I think we are a long way before operators or other constituencies in the mobile industry fully embrace - or even evangelise - multiplicity. Finding service providers which make it easy for you to get a multi-SIM offering is still tricky. I don't know any operator that offers some form of "personal multi-device management service", perhaps recognising that you'll also use a work-provided phone or BlackBerry from another operator. There's so much battle for subscription market share, that companies have failed to realise that there is possibly more value in integrating/federating across multiple different providers.

I suspect it will be 2007 before this trend really becomes evident, but it's something I'll come back to as it evolves.

Tuesday, August 15, 2006

Fuel cells in mobiles - Forget it, it ain't gonna happen.....

I've long thought that fuel cells in mobile phones were never going to happen. Yes, they pack a high energy density, but the idea of using a methanol-fuelled device by my head has never really appealed, especially given the ongoing risks of "normal" high-energy batteries spontaneously combusting.

I'd also been skeptical about the notion of taking cells with flammable contents on board aircraft. And yes, I know about less-flammable things like borohydrides, but given the hoo-ha of the last week about any liquids being taken on-board flights, I think we can finally lay this idea to rest. A pity, as it sounded kind of cool, but I think this is one area where practicality will get in the way.

Tuesday, August 08, 2006

OK, I don't usually take potshots at my peers....

.... but this piece of analysis/spin/marketing made me grin.

"MMS can be seen as a very popular and successful service"

... on the basis that people are now using it occasionally. (Even me, since it seems to have a decent user experience on my new phone).

Just a shame that operators have subsidised something like a billion cameraphones to the tune of around $10-20 a piece (apparently about $9 for the camera module and then a bunch of bits of software & integration work), plus had to invest $XXbn in MMSCs and other bits of infrastructure, and $YYbn in developing & marketing their photo messaging services.

It's pretty difficult to get any reliable stats on MMS traffic, but the anecdotal evidence I've seen points to a maximum of around 150-200m being sent globally per month (excluding the exceptional-as-usual Japanese). Verizon's one of the leaders, at 62m per quarter (ie 21m per month), while another big advocate Telecom Italia Mobile made a grand total of €49m from MMS in 2005. Scaling that up means MMS is (optimistically) a $2bn business in 2006.

Yeah, great success, given a cumulative investment of the order of $20bn+

Sprint, WiMAX and the fallout... and device thoughts

Hmm. I wonder how long before someone acquires IPWireless, now that it's big chance has evaporated, with Sprint's decision to go with mobile WiMAX rather than UMTS-TDD or Qualcomm/Flarion's technology.

This line in the press release is interesting: "...will provide voice service using the core 3G network. The 4G broadband network will offer a complementary, high-bandwidth service driven by data centric devices" as what it seems to suggest is that Sprint is deciding to focus its new EV-DO RevA network (due to be deployed from end-2006) for voice - maybe cellular VoIP. Reading between the lines, this seems to suggest that a broader move is occuring, polarising networks to device categories:

- 3.5G networks = predominantly cellphones, with a smattering of data cards, moving to VoIPo3G over time, and also supporting handset-centric applications like downloads, messaging etc. That is, mostly things that work with the operator's own services most of the time (ie on-net, either for voice or content/app portal, or, presumably IMS type stuff)

- 4G = predominantly data devices (in Sprint's words "computing, portable multimedia, interactive and other consumer electronic devices") - probably with a mix of "high speed pipes" (about time we had mobile VPNs, methinks) and in-house/partnered services, but generally more IT/Internet centric philosophy. I suspect it will be relatively independent of IMS's interference, and certainly not too walled-garden centric. This also confirms my view that WiMAX phone form-factor devices are unlikely to appear until 2010 in any meaningful numbers, if at all. Oh, and I'm willing to take a bet that the early Sprint WiMAX devices won't be SIM-based for authentication.

One interesting question is where this leaves smartphones in the future. Are the going to be confined to the "voice+a few other bits" category of 3.5G? Or will they evolve more into the higher-end device segment? I think that if this type of network divergence is replicated elsewhere, they start to look uncomfortably positioned on the fence.

Another interesting element is Samsung's involvement. Clearly, the Korean telecom establishment's decision to steamroller through the commercial deployment of WiBro is going to yield huge dividends for network infrastructure exports (and possibly IPR), just as its handset business looks on shaky ground.

More thoughts on all this in due course....

Whatever happened to WiMAX for cellular backhaul?

At various points over the last couple of years, I've heard WiMAX advocates talk enthusiastically about the possibility of using their favourite technology as a cheap & effective way of improving mobile backhaul.

So what's happened? As far as I can see, the need for faster, cheaper, better (ideally non-line of sight) backhaul has grown tremendously. The demand certainly seems to be there.

And a never-ending stream of alternative approaches is appearing, from fibre & metro-ethernet connections, to various forms of multiplexing & compression, even industrial-grade DSL backhaul.

But speaking to one of the largest WiMAX equipment vendors last week, I got the distinct impression that this is one of the "use cases" that hasn't made it as far as the initial marketing pitch for the technology. I suspect this is partly down to the limitations of existing interfaces (and physical space) at cell sites. But it also seems to be a distinct aversion in many traditional mobile carriers to the idea of deploying another wireless network, especially just for "spot" solutions.

Tuesday, August 01, 2006

Mobile phones features: operators, users & subsidies

Keith over at TeleBusillis has an interesting point about reclaiming analogue FM radio spectrum for digital uses.

He also talks about whether operator-subsidised mobile handsets should include features like FM radio, which generate no additional revenue for the carrier. He mentions Bluetooth in the same context, pointing out that it generates little extra revenue (some extra in-car calls, but I'd guess no more than 1% of total mobile minutes are on Bluetooth headsets), but enables easier sharing/piracy of ringtones or content. (It probably also generates expensive customer support calls).

This is a central issue in the handset business at the moment. In many (but not all) countries, mobile phones are heavily linked to the operators' business models. Often they will be subsidised, customised & distributed by the carriers, especially in markets like the US and Japan (China and much of the rest of Asia is the opposite, however). Some will even be designed by the operator from the ground up, and manufactured by the ODMs.

Keith concludes that handset manufacturers (he singles out Nokia as probably the earliest advocate of FM radio in mobiles) "cannot go winding up the operators and playing them for fools " .

I can sort-of see his point, but at the same time the operators also recognise that the phone has two "personalities" from the end-user viewpoint. Sure, it's a "service device" intended to be part of the mobile communications package you get from an operator, integrated with & optimised for billable services like phone calls, SMS, and possibly some of this other stuff that not many people can bothered with....

... but it's also a piece of consumer electronics / jewellery / computing hardware. Nokia refers to its N series as "mobile computers". Others draw comparisons with fashion accessories, or imaging devices. This is what customers want - and, if necessary, this is what the operators will need to "bribe" them with in order to persuade them to use their services. It's not just FM radio and Bluetooth - it's titanium casings & laser-etched keys, 3 Megapixel cameras & memory cards. I mean, the calculator software or the clock doesn't generate revenue either, but it's expected in the phone. Going forward, WiFi may also be included in some devices for non-operator functions (and will have to work well for these purposes, as well), as well as SIP for applications outside the operator IMS as well as inside it.

To some extent, the advent of the RAZR killed the idea that operators had "won" over manufacturers. In one swoop, Motorola proved that the "shiny thing" philosophy won over a large proportion of consumers. "Give me one of those.... or else I'll churn to an operator that will". Hardware is more valued than software, which in turn is more valued than services. Ouch.

The subsidy issue is a thorny one. Up to a point, I have some sympathy for the operators. Subsidising something like Bluetooth or WiFi that generates no revenue (or even helps cannibalise existing revenues) has got to hurt. Ditto subsidising something that generates costs like support calls, or higher returns (step forward smartphone OS's...). My view is that subsidising something gives the operator some moral authority to say how you use it. But not as much as some of the US operators, that cripple Bluetooth & otherwise act dictatorially over "their" phones.

The question is one of balance. How can an operator minimise subsidies, whilst not increasing churn (or, worse, persuading people to buy "vanilla" handsets not optimised/customised for any operator's value-add services). It's a problem that will only get worse, as Moore's Law & other technology scale economics mean that the average handset will bundle in more capabilities (memory, power, processor, display, software, connectivity....) faster than the operator can exploit.

Historically, subsidies were about either (a) acquiring new customers, or (b) keeping them. Now, you also have (c) persuading them to use more of your cool new services, by giving them customised applications and user interfaces. But customers don't naturally want phones with 9 icons, 8 of which are trying to sell them stuff - you have to give them a certain amount of other cool stuff "for free", or else they'll just exploit the scale economies & processor price/performance curve and buy one SIM-free.

I chose my last two devices based on capabilities I wanted - decent camera, memory slot, good browser, aesthetics etc - and then looked at which carrier subsidised it most, with the fewest restrictions on usage and the least onerous service charges. But then I live in the UK, which has 5 main mobile operators and numerous MVNOs, and so they compete aggressively for my custom. It's also helped by a strong non-operator retail presence (AKA Carphone Warehouse) and a decent channel for SIM-free phones via firms like Expansys. Other markets are less (or more) open.

Bottom line - operators may be handset vendors' main direct customers, but the end user often ain't stoopid. There's going to be an ever-greater bunch of stuff that phones can do, which doesn't generate revenue for the carriers. Sorry, but that's the real world. Black chrome on a razor's handle doesn't directly help Wilkinson Sword sell more blades, but it's still subsidised as part of the whole package, helping them compete against Gillette & minimise "churn" amongst the shaving population.