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

Thursday, January 31, 2008

Thanks and a couple of recommended links....

I owe a couple of "tips of the hat" to:

IP Convergence TV, edited by Jon Arnold, which includes a Guest Opinion column by yours truly on VoIPo3G

and

My estimations on Mobile Internet/Browser users get a mention on week's Carnival of the Mobilists over at a blog called SkyDeck (which I hadn't previously come across). It's got some other interesting articles too, including this one about multiple phone ownership in the US.

Cheers guys.

How do you recoup spectrum from the public sector? Let them sell it.....

Great idea from Ofcom, the UK regulator, this morning. Here, as in many countries, there are huge parts of the spectrum (often in prime useful bands) that has been reserved for the military, or other government-sector users. Typically, these organisations have so much that they don't have any particular incentive to use it particularly efficiently. But at the same time, they'll push very hard indeed if there's any threat of it being taken away from them - "national security" always wins arguments.

But there's another approach.... let these organisations trade or sell their spectrum, which might be worth £billions. Given budget constraints for the military in particular, the notion of investing a small amount in optimising their usage of existing spectrum, could mean they could sell enough of what is "spare" to buy another aircraft carrier, or better equip troops for ongoing conflicts.

Tuesday, January 29, 2008

Is it just me, or is webcasting completely useless

I'm currently listening in on a webcast vendor briefing - you know the sort, where you have to register & have slides pushed to you in a window on your desktop along with streamed audio, plus an option to submit questions by typing in them.

It's a terrible user experience - I'm restricted to the slide currently being 'pushed' (so if you're 5 mins late you can't review what you've missed). I can't flip through the rest of the deck to see if there's good content coming up. There's no download PDF / PPT option. I can't expand the slide size, so small fonts don't render properly and so on. I also needed to keep refreshing the window as the slide transitions didn't seem to work.

Sure, it's dependent on the system provider (in this case it says Thomson in the corner, which I guess is the managed / hosting provider). Some others have slightly better interfaces. But it's a solution looking for a problem - similar to the useless "magazine reader" software that gets used increasingly widely.

The old way works better - either email a PDF/PPT, or a link to download it. Then use a normal phone call or audio webcast.

Monday, January 28, 2008

Nokia / Trolltech acquisition - first thoughts

My kneejerk reaction to Nokia's purchase of Trolltech this morning was "Wow, Nokia's just bought a handset Linux OS vendor..... does this mean we'll see S60 on Linux, or a Linux base, or a multi-tasking Linux-based alternative to S40?"

But then I remembered a bit about Trolltech's roots, and the fact that its Linux OS/framework for handsets sits alongside an older business that is more about developer tools. Reading through the press release hints much more about enabling Nokia to assist developers create web-based "applications that work in the Internet" and also run on PCs and (Nokia) phones. "will enable Nokia to accelerate its cross-platform software strategy for mobile devices and desktop applications, and develop its Internet services business"

The more I think about this, the more I realise that Nokia is starting to get into a position where it can leverage its 40% of the handset market, by creating services that don't work as well (or at all) on its competitors' devices, except at the top end where they have full-spec browsers. This isn't a battle for the smartphone space, but featurephones. Realistically, and future high-end app developer will port across Symbian, WinMob, maybe Apple & Android & BlackBerry - or alternatively, they can run in the browser and exploit the inevitable presence of flat(ish) data plans.

But it's much trickier to do a cross-platform play in the featurephone space. Java is fragmented, browsers are often inadequate, Flash is sparsely supported. Plus, most mid-tier users outside N America & Japan use prepay tariffs with often-poor data availability. This is an Ovi play, and a pitch to offer useful third-party hosted services to mobile operators whose own internal services innovation has been weak. Essentially, Nokia wants to take an aggregation role between software developers (who have cool ideas but no easy way to deal easily with 100's of operators) and the MNOs. This seems to be a similar direction to that taken by Microsoft and especially Google of late.

That said - I'm still curious about when S40 is going to get a proper refresh. I'm increasingly convinced that full multi-tasking capability is mandatory for delivery of 'serious' nextgen IP services, Web 2.0 things like widgets and so forth. Some apps need to run "in the background", while user behaviour means people tend to flip between multiple services.

And another open question is what the platform(s) will be that support these apps - it sounds like Internet / Web-Services based, rather than IMS, for example. Since the spin-off of the infrastructure business to NSN, the handset/software part of Nokia seems to have lost a great deal of enthusiasm for IMS, preferring instead to focus on Web 2.0. It will be interesting to see what happens when Ovi migrates towards session-based services - whether it will pursue a "naked SIP" path, IMS, or something else again.

Thursday, January 24, 2008

FAO: AR/PR people - My 3GSM schedule is now full

Slightly off topic - but in the interest of saving time all-round, any PR/AR folk hoping to organise meetings & briefings with me during 3GSM in Barcelona next month, I'm sorry but my diary is now full. I'm currently receiving 10+ invitations a day, so if you read this before hitting "send", can I suggest you may wish to leave me off the list.

Exceptions may be made on a highly cynical & mercenary basis for people who are current or probable future paying customers of Disruptive Analysis.....

Monday, January 21, 2008

Using your mobile phone instead of a bunch of keys? No thanks.

This comes straight from the department of "Silly convergence ideas"....

Suggested by the CTO of a major operator at a conference this morning: Putting the function of physical keys (for actual door locks, not crypto) onto your mobile phone.

Yeah right. So when I divorce my operator (ie churn) do I need to change the locks on my house and car?

And if I accidentally leave my phone in my other jacket & walk out & slam the door.... I haven't got either my keys, nor the phone to call a friend or locksmith. Or if I lose the phone etc etc. Having separate (diverged) phone+keys gives redundancy & backup. Or if I've got 2 or 3 phones.. do I have to remember which one has my house keys in it?

OK, I guess I can see this working for certain cases - perhaps replicating a hotel key-card function on a phone. But even that is fraught with difficulty (roamers and handset diversity, basically)

And yes, there might be a few other niches (locker keys or padlocks, perhaps). But not "mission critical" keys for my front door or car.

QoS and bearer-aware applications - when is best efforts good enough?

I've been in a few discussions recently talking about QoS, IMS, differential classes-of-service over IP networks, and the GSMA IPX.

The usual topic of "guaranteed QoS" vs. "best efforts" always crops up, usually with VoIP or video lauded as the most QoS-able services. There's usually an assertion that customers will be prepared to pay more for assured quality. There's also usually a recognition that Internet or supposed low-quality IP connections are sometimes pretty good.

Obviously for certain applications it is true that customers will pay for QoS - 2 CEOs discussing a merger are going to want a premium connection, and the marginal cost of a QoS-guaranteed IP connection is tiny in comparison with the inherent "value" of their conversation. But for someone wanting to make a 1-hour call to their granny in Australia, it's not so obvious.

Then there is the question of whether any QoS / CoS level is chosen by the application or the network. This fits with a much wider question of whether networks will tell phones how to behave, or vice versa, in the future.

My own view is that that over time, the end-devices (and applications they use) will start to win out. Applications & devices will be able to examine the available wireless (or wired) bearers, and decide which to use based on an algorithm which perhaps compares bandwidth, latency, ownership, QoS, security and other criteria. Another issue here is whether QoS applies to the radio network as well as the core - the best QoS mechanism in the network (or in the IPX) is no use if you're out of coverage.

Surely it makes sense therefore for application to first check whether, at a given time & location, whether best-effort Internet access, or low-tier operator-guaranteed QoS is sufficient. If best effort is good enough, then great. If it is not good enough, the application can then request a better QoS level. There's no point paying for QoS when it's not needed.

Saturday, January 19, 2008

Mobile vs Fixed Internet Users #2... it's all about the browser

Let's play word association.

How about the word "Browser"?.....................

............ I bet you said "web", or "Internet" in response.

But that's not the full story, and this is why I think that Tomi Ahonen & I have been disagreeing so strongly over some analysis of the numbers of mobile & fixed Internet users. He asserted that 30% of global Internet users were using mobile access only, based on some top-down analysis of total Internet & PC users, and assorted survey data reporting mobile users' Internet behaviour preferences. I responded that that seemed completely at odds with bottom-up statistics that I could find, plus I have my doubts about any survey that purports to accurately the sample the whole global mobile user base, especially if it's sponsored by a vendor for PR purposes.

(In general, I don't trust survey data unless I know exactly the wording of questions asked, who's been asking them, how the sample was selected, what the purpose the study was - and ideally if I have the entire raw output rather than a carefully-selected summary of findings. My default belief is that 90% of respondents don't know what they're talking about or have the questions poorly-defined - and the other 10% lie. I'm open to evidence to the contrary, but it needs to be robust. I trust Government stats & surveys a bit more than others)

But the disparities in some of Tomi's data against my own, plus the fact that some survey responses have a high WTF? quotient, made me think about what's underlying data vs. measured reality.

I think the answer is this:

On a PC, browsers are almost (but not entirely) used to access the Internet, specifically the WWW. Yes, there are some other applications such as corporate intranets, accessing embedded standalone web servers in bits of computing gear for configuration, kiosks linked to local government private networks and so on. But basically on a PC, it's quite easy to write & distribute "native" applications as there's one OS platform that's 90% of the market, plus Macs if you want full coverage, and Linux if you're really enthusiastic. The only reason (historically at least) you'd use a browser was to access the web. This is changing a bit with the rollout of widgets, web services and the like, but for now it's a fairly reasonable assertion. (Main exception is the enterprise, where a much higher % of browsers are used for non-Internet tasks).

On a phone though, as any mobile developer or operator knows, platform fragmentation is hideous. Unless you're an operator able to tell your handset manufacturers exactly what platform to deploy (eg Verizon/Brew, or DoCoMo/Symbian or /Linux), most apps will struggle to go beyond 10-20% of addressable handsets. Maybe a few more if you're prepared to put up with Java's limitations and fragmentation as well.

In other words, on a phone, there's a big argument to use the browser as the only (relatively) standard data access platform. This means that mobile browsers get pushed into service for a lot more non-Internet access purposes than is the case on PCs - especially access to operators' walled-garden portals and content download facilities. (Yes, alternatives like "on device portals" also exist, but they've had comparatively low traction).

Now don't start telling me that Operator Portals are "the Mobile Internet" or any of that nonsense. They're not. Internet is short for "internetwork", which means the connection of two or more separate networks. Accessing an operator's portal from that same operator's customer's phone is no more an internetwork than a set-top box talking to a cable operator's head end, or my switching on the electricity at the wall for my toaster. It's one network, not several. There's no "inter".

If the term "The Internet" had been trademarked, any operator daring to call a walled garden by a derivative of the name would have been receiving nasty letters from lawyers quicker than they could say "per megabyte charging".

Now in some cases the walled garden does offer a narrow path to the Real Internet, or perhaps even a Google search box. Increasingly, operators in developed countries are offering "real" Internet access (albeit perhaps filtered/proxied in some way) and/or Internet email interconnection. (Separately, although email is clearly an Internet application, I suspect that most people tend to think Internet = WWW. But there are probably too few email-only mobile-Internet users to make a difference to the stats). It's worth noting that prepay services often don't permit web access, or else it's prohibitively expensive (especially on GPRS because of the inherent cost of data transport).

Although are grey areas, I think that in most cases it's possible to classify Operator Portal use as Internet/non-Internet. Roughly speaking, any content/resource that's accessible from any other Internet access = Internet, while any proprietary/operator-only content is not. So, for example:

- Google via a portal-embedded API = Internet
-Accessing news headlines prepared & edited especially for the portal = Not Internet
- Downloading ringtone from WAP menu = Not Internet
- Using browser to initiate POP3 access to ISP email box = Internet
- Pressing operator portal icon on phone menu & visiting home page = Not Internet

Over time, more grey areas will undoubtedly emerge - mildly operator-customised versions of Amazon & eBay, specific partnerships with Yahoo and so forth. Is initiating a Skype call from a 3 Skypephone, terminating on a PC, mobile Internet access? That's where it starts getting vague, but most of today's users aren't there yet.

With all this in mind, I've decided to have a bash at revisiting my Mobile/fixed Internet segmentation table, but this time I've done it twice. Once for "All Browser Users", and once for "access to the Real Internet". Again, there's some health warnings to go with these - I'm aiming for order of magnitude data, so don't read too much into 1m there, or 300m in another. Could well be 400k or 350m in reality - I don't have all the source data to hand. (And if I did, I probably wouldn't give it away for free on my blog....). Once again, I've aimed to make it all add up to the 1.3bn Internet users figure, which seems (fairly) robust, but which is also perhaps worth revisiting another time.

[Separately, it's also worth noting that total # of PCs in use is now >1bn, rather than the earlier 900m estimate. Interestingly, Forrester is forecasting 2bn in use by 2015 which would essentially give PC-Internet accessibility to pretty much anyone with a phone. I suspect we'll end up with more or less the entire planet having access to the Internet via multiple platforms]




But in addition, I've recalculated it if you include ALL browser access. So this includes operator portal access, and drastically increases the total number of mobile "Internet" users to more than 800m, which fits in with some of Tomi's cited stats. It also slightly bumps up the number of PC-based users, as there are a small proportion of extra non-Internet browser use cases (eg information kiosks, corporate applications, access to secure government private 'internets' etc)



The addition of non-Internet browser usage definitely skews the results more towards mobile. I estimate that 52% of "Browser users" use mobile handsets at least some of the time, and that 27% are mobile-only or mostly-mobile browsers. On the other hand, I reckon that 93% of "Real Internet" users have access to a PC at least some of the time, and that 82% of users are "PC-primary".

Using this approach, it's possible to view the world through different lenses - you can either pitch it as a mobile-growth story, or one that says that PC are still central to Internet access.



Thoughts & feedback very welcome. And if this type of analysis is of further interest, please contact me via information AT disruptive-analysis.com

Enterprise FMC - return of the picocell?

Forget about enterprise femtos - in general, they're too unproven, low-capacity and potentially difficult to manage to be a realistic proposition for a good few years, apart from maybe a small shop or home office.

For now, in corporate environments (or other 'indoor' locations like onboard ships) the operator-centric corporate FMC solution of choice is the femto's bigger brother, the picocell. They're more expensive, but much more "carrier class" in various regards.

(As discussed in 2 previous posts over the last couple of weeks, I don't buy the notion of operator-centric dual-mode WiFi/cellular FMC in the enterprise - although it's a good solution where the company itself anchors it to a PBX & manages it).

This article caught me eye. It's about a new US operator called Strata8 which is deploying a CDMA-based pico solution with PBX interconnection. Reminds me of Teleware's Private Mobile Network in the UK, Spring Mobil in Sweden, and a Swiss operator in&phone - although those three use GSM.

What Strata8 appears to do is to get rid of the usual blinkered hype about "binning the PBX" and instead looks to use & exploit it for the benefit of the enterprise's mobile users:

"Strata8 Networks Hosted Cellular provides two significant solutions to address and control enterprise cellular use:

- Utilize the enterprise PBX to terminate cellular calls at a significant cost savings compared to cellular carriers. This cost savings is magnified for workers who make international calls from their cellular phones.

- Take advantage of free calling between Hosted Cellular subscribers and between those subscribers and the enterprise PBXʼs extensions. Even calls between office locations are free."


The idea of "hosting" the cellular element, rather than "hosting" the PBX is very clever indeed. The main questionmark - as is often the case with picocell-based "indoor" or solutions, is whether the company can sign a nationwide roaming deal. This is often a major stumbling block, and has been one of the main reasons for a lack of traction in the UK market using the innovative low-power GSM spectrum licences.

Friday, January 18, 2008

Mobile vs fixed Internet users - a first pass estimation

Following on from my post yesterday about the real truth about fixed/mobile Internet access, I thought I'd have a very quick attempt at quantifying the actual state of things. I've done a back-of-an-envelope estimation, segmenting the estimated 1.3bn Internet users. I've tried to break things down by both device type (PC, mobile phone, various mixes) and access mechanism (cellular, fixed connection, WiFi, various mixes).

NOTE - updated post & estimates now at 2nd post here

Now clearly there's a lot of guesswork here. And yes, I've missed out some oddities like TV or games-console access, and fixed-wireless services like pre-WiMAX or satellite modems.

As I said, this is a first attempt. To be honest, the main objective here was more about coming up with the general design of the matrix & the enumeration of two suitable axes & categories. I've basically filled in number to add up to 1300 without detailed background data, but in such a way that there aren't too many alarm bells ringing.

The units are (million individual users) Any input or feedback welcome.... [apologies for the legibility. I'm sure there's a better way to upload the table....]

EDIT - one possible issue is around the 1.3bn users. If you include people who use WAP browsers on handsets to access operator walled garden portals (eg to select games/ringtones to download) then there may be more in total. In my view, it's not the 'real' Internet unless you go via the portal to a website like Google, either by typing in a URL or by the portal using an external web services API or similar. [by the same token, if you use a browser to look at HTML on your own PC, or as an interface to a corporate Intranet, that ain't 'real Internet' use either]


NOTE - updated post & estimates now at 2nd post here




Wednesday, January 16, 2008

Sorry Tomi, I disagree

I know Tomi Ahonen pretty well - we regularly encounter each other both virtually on our respective blogs, on the ForumOxford mobile discussion board, and in real life at conferences.

Some things we broadly agree on - the importance of SMS, the non-importance of many location-based services, and so on.

But on one area, we definitely don't see eye-to-eye - the relative importance of PC and mobile-based Internet access. Now, there's a million definition shades of grey here, so it's difficult to be definitive. Does a 3G modem in a laptop count as PC, mobile, or both? What about a dual-mode handset connected via WiFi to home broadband? Or via a femtocell? There's also the metrics of users, usage, traffic, sessions, money, purchases etc etc.

It's complex.

But whichever way you slice it, I have to flat-out disagree with his assertions in an otherwise great "state of the nation" post here about mobile in 2008.

"So out of all 1.3 billion internet users, only 37% access exclusively by personal computers (desktops and laptops). This includes all access from internet cafes and computer labs and shared PCs. Another 33% of internet users access by both PC and mobile. And already 30% of all internet access in 2007 was exclusively from mobile phones"

This runs completely counter to my "smell test". It just doesn't pass some basic "sanity" filters. Who are these 30% of Internet users who *only* use mobile phones? Which country? Which demographics? That's 400 million people who supposedly only ever get online via mobile. Against that backdrop, even the 20 million or so 3G laptop users is a rounding error (most of whom use other means as well, anyway). This can only mean actual mobile phones, over actual mobile networks.

How many of these 400 million people have you met? I knew precisely one in the UK last year - who was just using the browser on a SonyEricsson featurephone, until she moved house & got ADSL.

NOTE - my new estimates for mobile/fixed Internet user base are here

I find it hard to believe that *anyone* in either Japan or Korea *never* uses a PC for Internet access - that implies that they don't have access to PCs at home, work or school. Yes, maybe some Japanese *mostly* use their phones, but that doesn't add to that big a chunk of either the 33% or 30%. But lets be generous - maybe there are 10 million - maybe even 20 million - mobile-only users in those countries.

Which still leaves over 300 million of these mysterious cutting-edge handset web warriors to account for.

In Western Europe, the number of mobile-only Internet users is tiny, with the possible (small but very fast growing) exception of people using laptops with 3G. But most of those also have LAN access at work, or also use WiFi which to my mind counts as fixed-wireless access as it doesn't go via a GGSN in a mobile network, or (usually) involve authentication against a mobile operator's HLR. I doubt that there is a single user of FaceBook or MySpace that hasn't set up their page on a PC, even if they sometimes access it on a handset.

I haven't been to India, so I don't know the details first hand. But given the usual mismatch between prepay/low-end-phones/Internet access I'd be surprised if there are that many there either. Anyway, judging from this article, there's only about 50m Internet users in India in total. Let's be generous, and say that maybe there's another 10-20m mobile only users there (if any reader has a better view, let me know)

My gut feel (and I haven't done this methodically, I'll admit) is that there *might* be 50 million mobile-only users worldwide. Maybe 60m if we relax some definitions - laptop+3G, include some use of operator WAP portals or walled gardens (well, it's using the browser.....)

Overall - I reckon it's more like 3% than 30%.

Tomi - I'm happy to change my mind on this & apologise if you can point me towards this group of 400 million. I know you always try to evangelise what's possible in mobile. But I seriously think you need to take another look at your numbers in this case.

EDIT - some more data

From the Korean National Internet Development Agency, a brand new 400-page report on "Survey on the Computer and Internet Usage", gives the statistics of 34.4m total Internet users in total. 98.6% use a desktop PC some or all of the time. 51% of mobile phone users have ever accessed the Internet, and 25% have done so in the month previous to the survey. Given that 84% of Internet users have mobile phones, this equates to 21% being regular mobile Internet access users.

So basically, we can write off Korea as being a candidiate for supplying more than a tiny handful of this huge volume of mobile-only Internet users. Maybe 0.5-1.0m at the absolute most.

For Japan, it's difficult to find a full analysis from 2007. But from 2006 "The latest research by Ministry of Internal Affairs and Communication shows that there are more mobile Internet users (approx. 69.23mil) than PC Internet users (approx. 66.01mil) in Japan for the first time in history. It also shows 57% (approx. 48.62mil) of the Internet users (approx. 85.29mil) are using both PC and mobile devices to access Internet." So therefore 21m were mobile only, and 18m were PC-only. OK, that's a bit larger than I'd estimated, but not that much.

So, together Korea & Japan represent 20-25m mobile-only users.

Some good 2007 China data, from an official source, is here
96% of Internet users use desktop PCs, although 27% use phones & 21% use laptops so there's quite a lot of overlap. Analysing the stats suggests that maybe 2-3% are mobile-only. Out of 162m users in total, that gives, at most 5m and quite probably more like 3-4m

Indian statistics seem to be a bit contradictory. This article claims 3x the number of mobile users as fixed, but only seems to focus on subscriptions (and not multiple users/PC) while this report suggests a huge % of Internet cafe users who obviously don't 'subscribe' at all. This report from an Internet/Mobile trade association doesn't mention mobile access at all. This presentation suggests 13m "regular" mobile Internet users, but also seems to suggest a lot of "mobile Internet" access is actually to walled gardens via WAP, which to my mind is something a bit different. Sifting through all of this, I reckon my original estimate of 10-20m mobile-only users looks possibly a bit high: 12-15 sounds better.

So... scores so far:

Korea - <1m
Japan - 20-25m
India - 12-15m
China - 3-5m

Total above - 35-45m

I can't see any way the rest of the world is going to add up to another 50m+ mobile-only users. Russia? Indonesia? Brazil? other low-PC penetration markets? Doesn't see feasible so far.

So if the total number of Internet users worldwide is 1.3bn, the proportion using cellular only is well below 10%.

Joining the 3G dongle generation - what a bargain

I've written quite a lot recently about the uptake & user-friendliness of 3G USB modems for laptops. But while I've played around with quite a few demo ones, I've actually now gone out & put my hand in my pocket to buy one for my own 'official' use.

It's a Hutchison 3 HSDPA service, using a ZTE MF622 modem.

It is, quite frankly, astonishingly good value. Sure, the sign-up process in-store for a new business user account in 3's local store is a little clunky - the company mainly focuses on consumer users and their online provisioning/sales systems reflect this. The staff had to phone head office to find out how much the actual modem should cost me on my chosen tariff (12 months, 3GB/month cap) as although their pricing info had details for a consumer on the same plan, it wasn't in the system for businesses. In the end, they concluded that the modem is free for all business users irrespective of contract length (it would have been £70 if I'd subscribed personally over 12 months, or free for 18/24 months).

So how much does this cost? £12.77+VAT per month (basically $25). For 3GB, and no messing around with over-picky terms of service. Skype works fine (and is actually cited on the 3 website as a suggested use case). Streaming audio works fine. One quick test gave me about 700kbit/s downstream, though it allegedly should go at 2.8M if I sit next to a cell tower.

And if I travel, I can use it without roaming premium on 3's other partner networks (Italy, Sweden, HK etc). Off-net, roaming is about £3/MB for most European countries, or £6/MB for more obscure places. (ie painful, but no worse than 3's peers, as it reflects what I suspect are heinous wholesale data interconnect prices).

Going back to the pricing scale I proposed ages ago, this counts as a bargain. For comparison, Vodafone's price is double the price, and has more onerous (ie anti-VoIP) terms and more expensive roaming, albeit with perhaps more robust corporate-grade account signup & management. And it's certainly cheaper than using a hotspot at a conference or airport a couple of times a month. (For reference - T-Mobile UK charges £20 a month for its unlimited hotspot plan, albeit with a 30GB cap)

The only downside of it so far is that it seems to eat battery power rather quickly, especially where I'm on the edge of network coverage.

Tuesday, January 15, 2008

Enterprise UMA - a thoughtful analysis

Aruba Networks has come out with an impressive & comprehensive analysis of what would be needed to get UMA service working on a full enterprise-grade WiFi installation.

The bottom line is that it should be feasible.... but with quite a lot of "ifs", some stated, some implied.

- If the corporate IT manager is prepared to configure the WLAN security to recognise UMA devices and deal with authentication of devices & traffic & firewall accordingly
- If they define correct SSIDs and set up the WiFi configuration on the handsets appropriately
- If they have good site-wide coverage of the WiFi network
- If they're prepared to ditch their PBX in entirety and switch over to a service provider's hosted centrex offering
- If they're willing to accept a much smaller range of handsets

This fits in with the post I made last week. Basically, it's still a niche of a niche, although it's good to see the first sensible analysis of what needs to be done if you're insistent on this route.

Some factors to consider:
  • Are any UMA operators able to sell & support all this? If not, what are their channel partnerships?
  • What happens to enterprises that want to retain some PBX extensions (eg for call centres or desk-bound staff)
  • What happens to multi-site or multi-national corporations that have locations not covered by a single operator, or a mix of ones with good & poor cellular coverage
  • When doing ROI calculations, what are the assumed lifetimes on dual-mode handsets, and how does annual depreciation and TCO compare with normal PBX handsets?
  • Will there continue to be a decent range of handsets in future? In particular, there are still no Nokia Series 60 UMA phones
  • What's the migration strategy from PBXs - or is it an all-or-nothing forklift upgrade?
The cynic in me wonders what's tempted Aruba to put all this work in.... a relationship with T-Mobile US or maybe Orange? Or has Kineto stumped up for this on a professional service basis?

Differential pricing for handset and laptop data - and a solution?

I've noticed an interesting trend recently to have different data tariffs applicable to 3G laptop modems, Internet-capable handsets and PDA-type devices.

In general, the phone-based data services are more expensive in per-MB terms, and more constrained in terms of service (VoIP-prohibited etc). Certainly in the UK, there are no low-cost data-only 3G contracts. And increasingly, the operators are loading the phones with all sorts of software and applications to dissuade you from accessing the real Internet.

T-Mobile only sells Web'n'Walk these days in conjunction with a quite pricey Flext voice & SMS contract - which is fine if you're a one-device sort of person, but personally I like to have a mobile data-only device for email & web browsing, separate from my phone. A two-device strategy also means it's billable to Disruptive Analysis Ltd rather than Dean Bubley Esq to make life easier for my accountants (and reclaim the VAT).

It's intensely annoying that I can't buy a sensible data-only tariff for handheld devices (with or without a subsidised device), which also comes with sensible roaming charges & ideally some form of WiFi-friendliness.

(I'm presuming here that I can't buy a £10 a month Hutchison 3 USB modem with 1GB/month cap, and just put the SIM card in a 'vanilla' smartphone instead.... but maybe I can?)

Anyway - frustation has led me to ponder alternatives - and I came up with this:

We all know it's possible to use a phone as a 3G modem for a PC, connected via Bluetooth or USB.... but what about the other way around? Can I get a cheap PC-based 3G card and somehow use it (while the PC is turned off or in standby/hibernate mode) as a modem for my phone? Or better still, could I just have a standalone 3G modem dongle incorporating a battery and Bluetooth and a basic OS so it can be accessed via any device you have?

Just musing, but it increasingly seems sensible to separate your 3G data access device from your 3G data application device, so you can benefit from the best tariffs, Internet-access QoS and policies and so on.

Carnival of the Mobilists, Dongles and 3G

My post from last week about 3G-embedded laptops is listed in this week's Carnival of the Mobilists at Xellular Identity - many thanks. There's also a link to another similar article talking about 3G dongles and their possible impact on WiFi hotspots, from Andrew Grill of Seeker Wireless.

I certainly agree with him that the sort of cheap deals available on 3G modems in some European countries could well impact hotspots - coupled with the sheer rapacious idiocy of some hotspot WISPs over the last four years, who have managed to squander an almost unassailable lead in providing data connectivity. There are still plenty of hotels that charge €25 for WiFi, or ludicrous sums for conference organisers.

I'd single out Swisscom as one of the worst offenders I tend to come across - they've prompted me to go in search of 'free WiFi' cafes near hotels on numerous occasions. The other particularly crass example of WiFi pricing that sticks in my mind is that of Kubi Wireless at the Fira in Barcelona, although it wouldn't surprise me if the GSMA quietly has a word with them to encourage high prices during 3GSM (actually, it wouldn't surprise me if the GSMA tries to take a cut as well, given their general philosophy of applying "resentment-based pricing" to anything they can control, such as analyst registration for 3GSM.... but that's another post)

I just wish Hutchison 3 had a network in Spain, so I could benefit from their zero-roaming data charges during this year's 3GSM....

On the other hand, some other countries and WISPs just seem to "get" pricing for WiFi. I've noticed that on recent trips to the US I've been able to completely avoid cellular data roaming charges, as there has been free/cheap/easy WiFi almost everywhere. And some bits of Europe get it right too.

On a related note, I see that O2 is cutting its data roaming pricing. Well, it's a start, but £3 per MB is still a couple of zeros too high to get the European Commission off their backs. (And I hadn't realised that prepay roaming was previously £15 per meg. Ouch, talk about bill shock).

Friday, January 11, 2008

FDD WiMAX... that changes a few things....

In 3GPP standards, there is an expectation that the 2.5GHz band (which is actually from 2.50-2.69GHz) would be split into a mix of paired spectrum (FDD) - in which you have separate chunks of frequency for uplink and downlink channels - and unpaired spectrum (TDD). In TDD you use a single chunk, but have up- and down- timeslots. (Hence either frequency- or time-division duplex).

The general consensus expectation is that there will be a band-plan which goes 70-50-70MHz, with paired 70MHz FDD chunks sandwiching 50MHz of TDD in the middle.

Historically in Europe, FDD has generally meant the established GSM/UMTS family of technologies. In 3G bands, especially 2.1GHz, chunks of unpaired spectrum have been rather under-utilised but theoretically dedicated to the TDD flavour of UMTS 3G, long but unsuccessfully championed by companies like IPWireless.

The advent of WiMAX (another TDD technology) has changed that perception, as unlike UMTS-TDD, it appears to stand a chance of being commercially practical on a broad scale. Unsurprisingly, the 3G FDD advocates aren't too keen on the competition, which is why 2007 was full of legal and regulatory wrangling about whether or not WiMAX counted as a 3G / IMT-2000 technology & could therefore be used in what were previously thought to be the sole preserve of UMTS-TDD.

There have been various other angles on this around technology neutrality, legal arguments about the semantics of whether "designated" means "exclusively designated" and so on. But in the end, the ITU defined WiMAX as a species of 3G anyway, so it's good-to-go in newly-available chunks of TDD spectrum like that in the middle of the 2.5GHz band.

Some regulators - especially Ofcom - have appeared keen to see WiMAX evolve as a competitor for established UMTS FDD 3G. Indeed, Ofcom has been working very hard over the last year on an innovative and controversial structure for the upcoming 2.5GHz auction. Basically, this involved a complex way of 'flexing' the aforementioned 70-50-70 bandplan for FDD & TDD in 2.5GHz, to allow WiMAX operators to get more TDD spectrum if they bid more money. So depending on a horribly complicated set of auction procedures, there might be 60-(50+10)-60-(10) allocation, or 50-(50+20)-50-20, with more WiMAX-suitable spectrum in the middle & top of the band.

This has cause lots of controversy in terms of auction design, assessing whether there would be more interference, the need for guard bands and whether existing handset & base station silicon designs could deal with the filtering requirements and so on. Ofcom appears to be pushing ahead regardless.

But in the background was something that I'd been wondering about for a while - and which the WiMAX industry has (mostly) kept a lid on..... basically an FDD flavour of WiMAX. Well, that cat is now truly out of the bag & is now frolicking amongst the pigeons, courtesy of some commentary from Airpan's CEO. I'd had some hints about this before, but I'd thought the main aim was to get WiMAX working in paired-spectrum 700MHz bands in the upcoming US auction.

Looks like actually they're pitching it head-on against HSPA, EVDO, LTE etc in the 2.5Ghz band. I wonder if that means that Ofcom & other regulators need to go back to the drawing boards and re-work their interference assumptions for a possible cellular/WiMAX mix across the whole band.....

Thursday, January 10, 2008

The only way that 3G-embedded PCs might become popular

I've written at length before about howI believe that 3G-embedded PCs, although seeming at first glance like an 'elegant' solution to mobile broadband, are in fact currently a poor proposition for most users. And indeed the market has borne this out - the vast majority of new signups for mobile broadband are with USB 'dongle' modems or various forms of aftermarket PC card. Operators like 3 in the UK are pitching 3G dongles at consumers (even in black, pink, leopard-print and so on), and Vodafone has just announced "stick" USB modems this morning.

However, some suppliers such as Ericsson continue to bang the gong about embedded modules. I've seen it issue rather optimistic predictions about 50% penetration of embedded 3G in laptops in 3-4 years' time. Now arguably, if the price comes down far enough (ie close to zero), then yes, you could have 3G in most PCs. But that doesn't mean that people would actually use it - after all, there's plenty of unused capabilities in a typical computer (how many of you use FireWire?). But if it's just a tick-box item that's left dormant, the PC manufacturers will nail your margins to the floor. For Ericsson (and anyone else) to make money from 3G modules for PC and other consumer products, they'll actually need to be used by a decent proportion of customers.

But here's the problem:

Most people will not want to buy a laptop that is tied to a monthly subscription, especially not to a single specific operator for the entire life of the laptop.

Fundamentally, a laptop is a computing device. People are used to buying computing devices and owning them outright, not having them tied to a service provider that might try and get in the way of your use of it. Sure, some enterprises might lease them or get HP or IBM to do a "desktop outsourcing" contract, but generally there's no service model embedded in the PC ownership and usage experience. You get your broadband separate from your PC. Part of the reason is that the lifespan of a PC is >> length of a typically-desired contract. Consumers aren't stupid - they know there's a reasonable chance they'll want to churn over the course of 36 or 48 months, as they know how fast capabilities and prices evolve. While the operators would love them to stay loyal, they'll really have to bribe them heavily to achieve this.

Yes, there are always going to be exceptions. High-end business users with well-known typical usage patterns. Students for whom you subsidise the laptop down to zero, in return for a per-month subscription (essentially acting as a laptop consumer finance provider).

But for most people and enterprises, mobile broadband will have to fit the model of separating the OPEX and the CAPEX. Buy the laptop, and then conduct other transcations to provide connectivity. This is why the standalone USB modems make so much sense - consumers can make a separate decision about computing and connectivity, just the same as they do with their home broadband, or their business IP-VPN. Recent trends towards openness in the smartphone space would seem to suggest that some currently locked parts of the handset market are moving towards device/access separation there as well.

So... what would it to actually convince a lot of people to use 3G via embedded laptop modules? What needs to change? What pressure do the laptop vendors need to exert on the operators?

The following list is what needs to happen if big numbers of people are really going to start using 3G functions embedded in PCs. Some will deeply irritate mobile operators, so traditional cellular operator-centric companies like Ericsson or Sierra Wireless will need to ask themselves if they're truly prepared to be hardcore about this, or whether they should just focus on separate USB modems rather than embedded. (Edit: I realised the original wording made it sound like Sierra is embedded-focused - it's not, it sells mostly USB & card modems)

- Make the laptops totally operator-agnostic. People buy laptops for a lot of reasons - power, screen, keyboard feel, brand, aesthetics, corporate purchasing rules and so on. Let them select the one they want, not one of a handful supported by a specific operator. Don't have it shipped with pre-loaded operator-specific software. Let them choose an operator after they've bought the PC, and make it easy for then to switch operators at a later date if they want.

- Put the SIM slot in an easily accessible place, and make sure it is fully "hot-swappable"

- Beat up the operators until they supply SIM-only data plans, ideally both prepay and subscription. Make sure that these are true 'plug and play' SIMs that contain all the necessary driver software, special applications and so on. Ideally, when the SIM is removed, it will remove all traces of operator-specific software from the PC, with the possible exception of some harmless preferences.

- If operators are insisting on subscription plans, beat them up until they stop charging an international roaming premium (or make sure it's really small). And beat them up some more to make sure they don't have ridiculous application-layer policies enforced in the networks (obviously with exceptions for security reasons, possibly content filtering etc etc). People don't take kindly to others telling them what apps they can & can't run on their PCs.

- Advocate the sale of prepaid 3G SIM cards at airports, hotels and other places that travellers visit. The nice thing about laptops is that they don't need a consistent phone number as people are unlikely to want to accept inbound calls (or can do it after signing in via the web on a separate software client anyway). This is why roaming fees are ridiculous - you should be able to buy 'local' access. If I use WiFi, or go to an Internet cafe, it doesn't route all the traffic back via my UK ISP and charge me roaming rates. There's no reason why you shouldn't be able to access your Vodafone or Orange account (and voicemails, SMS etc) securely, even when connected via a 3rd-party's access network.

- Make sure the embedded connection manager in this PC is user-friendly and user-centric. Allow people to set up their own connection policies in sensible ways. If you think the standard Windows CM isn't good enough, get a custom one pre-installed. Beat up the network operators to use this embedded CM as a core (with temporary branding while their SIM is in the PC, perhaps), rather than trying to install a separate one. When their SIM is removed, make sure that the embedded CM is left in a pristine 'vanilla' condition.

- Disabuse any operators of the notion that they have to test each individual model of laptop in the labs. Get the module pre-certified and usable in any PC. If there are other bits like antennas that are external to the module, either get them pre-tested too, or tell the operators they need to relax about them. Or set up some centralised industry testing facilities. This is the prime reason why USB modems are so popular - they can be tested just once, and then work on all laptops. Putting 200 sorts of laptops per year through 200 operators' testing labs is clearly nonsensical.

- Make sure the modem module is easily upgradeable (software and/or hardware). If people buy a laptop now and expect it to last for 4 years, they'll be upset when they can't get HSPA+ or LTE (or WiMAX!) or whatever else is the best/fastest/cheapest/coolest in 2011. If you don't, you can bet they'll just stop using the embedded module and get a cheap external modem instead.

- Consider making it dual-mode HSPA and CDMA-capable. And maybe TD-SCDMA and/or WiMAX too.

- Put as many frequency bands in it as possible

- Make sure that the HSPA bit plays very, very nicely with WiFi in the laptop. Don't be stupid about this - in certain circumstances, people will ALWAYS prefer WiFi. When a business user is sitting at his desk in his office building, connecting to his Oracle database in his basement, surrounded by his high-spec 802.11n WiFi (and fixed ethernet for that matter) tromboning his data connection via the cellular network is very silly for all concerned. Make sure that the connection manager enables WiFi (even unsecured) to pre-empt HSPA if that's what the customer chooses. And don't even think about putting in unreasonable default settings for HSPA/WiFi preference when it ships.

- Make the laptop conscious of when it's connected to a femtocell rather than the macro network. It may need to behave a bit differently.

If all of these occur, then maybe just maybe usage of embedded 3G in laptops might take off. There's still time to exploit some of the more egregious pricing strategies of the WiFi hotspot firms and substitute usage. But don't focus on WiFi as a competitor.... the main one is the 3G USB modem. At the moment, the business model is much more attractive to the end user as it decouples the laptop purchase from their access expenditure. That's how people buy their computers & broadband - I'd suggest you work with this rather than against it. It's probably easier to change the attitudes of a 100 operators than a 100 million consumers.

Tuesday, January 08, 2008

VoWLAN for the enterprise - role of the service provider?

It's come to my attention that a particular vendor has sent out a press release taking shots at my recent VoIPo3G report and some of my other commentary on VoWLAN. So far, it seems to have got precisely zero coverage, so I'm not going to embarrass them further by naming them. (One journalist has been a consummate professional and asked me for my response to the comments, though).

I've got no particular problems with either different opinions, or people criticising my conclusions. I dish out my own views with a fairly pointed tone; it would be hypocritical not to accept back counter-arguments.

That said, I do take exception to exaggeration and outright misrepresentation of my views. Given I have previously had a good working relationship & several healthy debates with the anonymous firm, I'll blame the tone & wording on the hapless PR-droid, who may still be hungover from the excesses of the New Year, and who I guess may now be looking for another job.

I'm particularly irked by the false suggestion that I've predicted that VoIPo3G is "set to end" VoWLAN, especially in the enterprise. Even more bizarrely, the release infers that the alternative I suggest in the report is "a dedicated 3G femtocell for the enterprise".

The reason it's an inference rather than stated as a fact, is because neither the company nor the PR agency has bought or read the complete report. They've seen a brief summary and a conference presentation. Neither of those mentioned enterprise femtos as the way forward, as I've long held the view that's an extremely unlikely proposition in the near term. Given that I've closely followed the excruciatingly-slow evolution of enterprise picocells since 2001, I like to think I'm on fairly safe ground here knowing what I'm talking about.

Conversely, I've long been moderately bullish on enterprise use of VoWLAN, especially when it's provided in an "enterprise-centric" fashion, integrated with an IP-PBX or corporate mobility controller. The sort of approach taken by Cisco, Avaya, Divitas, Siemens, Nortel/Microsoft & others resonates well with enterprise users who wish to operate a blend of desk phones, cellphones, dual-modes and PC softphones and manage the system themselves. No, it's certainly not easy, either in creating a good on-handset experience with access to applications like SMS, nor in providing good-quality voice-optimised WiFi coverage on enterprise sites. Integrating with existing systems - especially enterprise network security systems - is also a challenge.

But if anyone can solve these issues, it's the people with lots of LAN expertise, existing PBX/telephony sales & support relationships with the customer organisations, the ability to install complex WiFi installations and integrate them with all the appropriate security widgets, and integrate the whole thing with other IT and communications platforms ranging from SAP to Exchange to investment dealers' trading systems. Frankly, the mobile bit is easy in comparison to much of this.

Some of the best channels for this sort of thing are actually the systems integration arms of the large fixed (or fixed-mobile) operators. Often, they have long-standing expertise in selling PBXs, loads of established client relationships and engineers in vans, and they actually have some customer-facing personnel who understand things like ethernet, firewalls, contact centres, calling groups and so on. They also understand all the annoying minutiae of fixed-line telephony in businesses - things like fax machines, alarm systems and so on. Also for large companies VoWLAN involves some serious work doing RF site surveys, installing and monitoring complex WiFi installations from the likes of Cisco, Aruba, Trapeze or Meru. And because the fixed operators are shifting towards IT services and consulting models (plus things like international VPNs), they're not going to get too hung up on precious minutes of mobile traffic being diverted into the IP domain. And they have a healthy dislike of excessive mobile termination costs.

As well as fixed operators, some other channel partners at the vanguard of the IP-PBX industry are also well-placed to fulfill this type of project and provide adequate consultancy, integration and support.

But the average mobile operator does not possess these skillsets. Sure, they might have recently started selling ADSL lines, on which they could (in theory) hang a WiFi AP and connect to some dual-mode phones (or a pico/femto for that . Or they may have been brave and attempted dual-mode services on another operator's broadband, hoping for benign traffic management conditions. They might even have sold a few UMA or pre-VCC phones and plans to consumers. Maybe even a handful to some 'small businesses' (generally home workers or perhaps 3-person offices or shops, I'd guess).

But the simple (broadband+one or two WiFi APs+dualmode) operator proposition is not 'enterprise-grade', and nor is selling it going to be necessarily a straightfoward or profitable exercise. Clearly, the odd WiFi AP here & there isn't going to cut it for large campus sites or office blocks. Yes in theory they could put a general-purpose WiFi hotspot in the lobby for visitors, but frankly that's not a hugely profitable business. And although retailers and fastfood companies and other "high street" operations have 1000s of small locations which could get by with smaller WiFi networks, getting them to deploy VoWLAN is a huge exercise in systems integration and rollout logistics. And is usually linked to 100 other in-store retail systems, IT back-end, corporate VPN etc etc which again is well outside of scope for the mobile guys.

Which leaves the semi-mythical SME (small-medium enterprise market). I wrote what was quite possibly the first major analyst research report on selling IT to SMEs, back in 1995ish. Key takeaway - it's not one market, it's hundreds. And the headline numbers are misleading. When you first look at the statistics, you discover there are millions of SMEs (ie companies registered or paying tax). Marketing departments love this. Then they discover that half of them are actually self-employed people working at home, and many of the rest are dormant companies (XYZ Trading Ltd with the same address and sole director as ABC Trading Ltd). Then they realise that some of them aren't exactly ideal for VoWLAN deployment with dual-mode phones (eg farms, factories etc.)

Then they find information/telecoms-rich segments like fast-growing media companies or small software firms, or consulting/accountancy practices. These are indeed the right targets for FMC. Then they realise that they already have PBXs halfway through the depreciation cycle, and buy their kit exclusively through one of 30,000 local VARs and resellers who know their business inside-out. Where the wiring closets are, who's going to be around on Saturday to unlock the building to install stuff, where the sales team sits in the office, and so on. Then you find that they've moved office 3 times in the last 2 years as they're growing so fast, so are reluctant to sign a 24 month contract tied to a specific location.

And so on. That's not to say that certain companies, with the right technology set-up, caught at the right time in the investment cycle, aren't potential targets for operator-led VoWLAN. But identifying them and keeping down the cost-of-sale (and cost-of-support) isn't an easy task.

In addition to this, there's a bunch of other issues to consider - but I'll leave discussion of those for my clients. As a taster - think about numbering, user behaviour with mobile devices, WiFi performance, DSL performance, security, dual sourcing and myriad others. Not to mention the fact that many mobile operators still (often rightly in my view) treat WiFi with disdain or at least caution.

In general, the discussions above are part of the reason I am so dismissive about the prospect for Mobile Centrex, especially when positioned as a "PBX replacement". Most of the issues about customer support apply to single-mode cellular enterprise systems as well. My view is that is very much a 2003-type black and white story, and that in reality, any mobile substitution story in the enterprise will need to go through many steps of interoperability, integration and slow migration. Maybe adding 50 mobile 'seats' onto the existing system, for example. Or rolling out to a couple of retail sites a week over the course of a 2-year project. Any business model which works on the premise of "throw all your old gear in the bin on Friday night, and the new system will be 100% ready on Monday morning" is laughable.

Bottom line on my real views (as actually discussed in my reports, presentations and this blog, as opposed to being guessed or inferred by someone who hasn't asked me):

  • VoWLAN in enterprise is a tough nut to crack, but there's a lot of the hardest work done already, and deployments are now growing. But it's primarily going to be enterprise-centric rather than carrier-centric. Some operators are definitely going to be involved, it's principally going to be as channels to market via their fixed/IT/PBX systems integration and reseller divisions. There are some options for future enterprise VCC-based dualmode with more carrier involvement and management, but it's still likely to be integrated with existing corporate-managed PBXs and LAN/security infrastructure.
  • VoIPo3G in enterprise will be driven firstly by laptop users with mobile broadband, then will be slowly added on to smartphones which may also have VoWLAN capabilities. VoIPo3G is unlikely to gain much near-term traction indoors in large enterprise sites beause of RF coverage limitations.
  • Enterprise femtocells are a very long way off
  • Mobile-operator managed VoWLAN in enterprise is going to be almost trivially small. My predictions for 2008 stated that mobile IP centrex had an opportunity to capture "10-20% of 10-20% of business lines". Well, I'll follow on that theme and estimate that mobile operator-managed DSL+VoWLAN/dualmode mobile IP centrex has an opportunity for 10-20% of the 10-20% of the 10-20%. Total market worldwide is about 400m business extensions - you do the sums.

Monday, January 07, 2008

Two fixed HSUPA routers launched... both using VoIPo3G as a use case

I see that two rival vendors of 3G fixed routers (typically used in corporate networks) have announced HSUPA versions of their products, with commentary about their suitability for VoIP.

Top Global - "Many of our enterprise customers are using the 3G MobileBridge (MB8000) with IP phones which route voice traffic from their remote offices over 3G cellular network work to their corporate IP-PBX"

Sarian (release isn't up on the site yet, but their PR team has sent it to me) which major on the use of the technology for uploading real-time, legal evidence-quality CCTV but also refers to the possible use of the 3G router for emergency backup of fixed-line VoIP "ability to route large quantities of data in real-time also facilitates business tools such as VoIP and video-conferencing, while in the event of fixed-line network crisis"

Obviously, this is something of a niche usage case, compared with handset- or laptop-based VoIPo3G, but it's still interesting to see, especially as it will be a further driver for carriers to relax restrictions on VoIP on their 3G data plans, particularly for their corporate customers.

It's also another thing that packet inspection / policy boxes (and their managers) need to be aware of - having the network squash someone's disaster-recovery traffic is unlikely to drive customer loyalty.....

The question is whether the rest of the QoS & policy infrastructure (and the people managing it) will have the fine-grained control to say that certain customers' VoIP is acceptable, and others might not be. Especially as the VoIP may be added as a use case for the backup system by the enterprise network manager at a later date - you'd hope they would check the T's & C's they'd signed previously, but in reality?

Tuesday, January 01, 2008

Mobile and wireless predictions for 2008

I think that 2008 is going to be a tough year for the mobile industry. With economic clouds gathering, saturation in major markets, competitive and regulatory impacts on margins, and little traction for new services, I'd say that developed-world operators are facing an unpleasant background to the new year.

But the grey outlook for operator services doesn't equate to the whole of the mobility and wireless industry. New business models and products will gain traction. Above everything is the impact of Moore's Law and the power of the handset vendors. Increasingly, the phones have the ability to tell the network(s) how to behave. There is clearly a pent-up demand for access from mobile devices to 'the real Internet'.

I've put together 10 predictions which I'm expecting to evolve into recurring themes over the next 12 months:


1) There is a notable shift towards non-operator unlocked 'vanilla' handsets

Globally, about 50% of phones are sold through operator channels - although it is much higher in operator-controlled markets like the US and Japan, as well as those with an addiction to subsidy, like the UK. Various trends are emerging that will start to reduce this in 2008, although change will be slow in markets where operators retain stiff control over retail outlets.

  • Firstly, the US is starting to wake up to the idea of unlocked phones, courtesy of non-AT&T iPhone angst and the new 'bring your own phone' openness pitch by Verizon.

  • Secondly, Moore's Law & scale economies are bringing down the price of really quite good non-subsidised phones to below important thresholds like €100 or £100.

  • Thirdly, operators' attempts to lock people into 18 or 24 month contracts will mean they get hungry for cool new devices long before they're due for upgrades.

  • And fourthly, some segments like VoIP users or enterprises will recognise that 'there's no such thing as a free lunch' and realise they're better off with vanilla devices rather than subsidised ones locked-down and larded with operator menus.
Expect alternative options for handset financing, a healthy second-hand market for phones, and a slew of new retail, B2B and online options for acquiring devices.


2) The European Commission cracks down on data roaming prices.
You've been repeatedly warned that this is on the watch list. Yet many operators still charge around €10 per megabyte for roaming for mainstream customers. There are a few exceptions - Vodafone's €12 per day cap, and Three's free data while on-net. But as we move to flatter networks, flatter tariffs and split-tunnels breaking out Internet traffic in the radio access network, the current prices start to look more like those of a cartel. Pricing needs to have at least 1 zero, and arguably 2 or 3, chopped off. Expect Ms Reding to take action.

3) Mobile broadand continues its rapid growth - but 3G-embedded laptops lose out even further to USB-based 3G modems
External USB modems are now a massmarket proposition. They're available in pink, even. They work with anyone's existing laptop, they selfload software, they don't require that operators put the PCs themselves through testing labs, they're easily understandable and upgradeable. Sure, an operator would love to tie a PC user to a single access service for the life of the device. But it's not going to happen - unless they fancy subsidising computers down to free. In enterprise, they overcome sales channel issues that mobile operators aren't great at explaining computer features, and IT hardware salespeople aren't experts on HSPA. But unless you're absolutely cast-iron sure that your behaviour (or competitors' prices or coverage) won't shift over the 3 years or so you've got the notebook, you'd be crazy to go with an integrated solution. A possible solution is a good 'universal connection manager' as part of Windows, rather than risking layers of operator-customised software on your PC. Expect the USB approach to accelerate even further away from 3G built-in PCs in 2008.

4) At least one mobile operator will face an investigation over reported numbers
I've posted several times before about operators fudging their reported KPIs, using opaque or arbitrary definitions. One of my more vitriolic blog comment critics has even claimed, as an insider, that carriers regularly overstate subscriber numbers to appeal to investors' superficial analysis - and seemed quite sanguine about this. I've had private discussions with other analysts where we've looked in disbelief at some published data. Sure, internal management have access to more 'real' figure - but that doesn't excuse them from providing it externally as well, as most are public companies. I suspect that a serious financial miss, or a whistleblower, could provoke some more serious scrutiny into irregularities, if obfuscation turns out to actually mean misleading.


5) Android... hmmm, it's just another platform. I really can't see what the fuss is about Android yet, especially outside of open smartphone-starved North America. That said, I'm not going to write it off - clearly the Big G holds a few cards up its sleeve. But one thing I'm pretty certain of - it ain't changing the world in 2008. We'll be back here in 12 months saying that Google's Android might be a big deal in 2009.

6) Technologies for exploiting end-user context and state become the hot topic of the year. At the moment, mobile 'presence' (IMS or proprietary) is pretty lame. The user is registered to the network.... hasn't been 'active' on a call for 79 minutes, has a self-set status they've forgotten to update.... and that's it. How much richer would it be if the network could extract more useful 'state' information about the device and/or user, especially if it is enriched with embedded sensors... "phone on charge", "user is on a Bluetooth headset", "battery low", "at location xyz", "moving in a way that looks like it's on train", "in a darkened room" and so on....



7) Operators realise that knee-jerk attempts to block VoIP are counterproductive.

2007 has seen many operators move towards flatrate data for both PCs and smartphones. But quite a few have used restrictive T's and C's, or in some cases port-blocking or other network means, to try and stop people using VoIP over wireless networks. A few have tried to charge VoIP-able 3G flatrate as a premium service, although a few more enlightened operators have shrugged and adopted an 'anything goes' approach to their data pipes. Expect many more to follow suit in 2008. The possible cannibalisation threat is overstated - and is more than balanced by the benefits of mobile broadband service contracts. In any case, things like mashups, VPNs and non-telephony VoIP will make it much trickier to isolate VoIP as a distinct "service" in future. Net result - pragmatism, and some of the more visionary operators exploiting VoIP rather than fearing it - either launching their own fully mobile VoIPo3G services, or partnering with players like Skype, Truphone or fring.


8) Femtocells have a year of ups and downs. Some niche success, but practicalities will mean it's H2'09 or 2010 before massmarket deployment.
Sure, we'll see some headline "deployments" in 2008, much like we've already seen Sprint's much-ballyhooed Airave launch. But these will mostly be v0.9 soft launches, not full, production-ready offerings that are able to be sold in their millions to Mr Joe Average Punter. Operators will realise that the proposition, while sounding good in paper, has a million niggly little issues around user experience, management, billing, regulation, emergency calling, numbering and so on. The business model will also need to evolve beyond purely coverage-based or macro-offload. As per normal, more attention has been lavished on the network aspects than actually thinking through user experience and requirements. There needs to be more standardisation, a recognition that there could be lots of multi-femto households, guarantees that femtos work nicely when colocated with WiFi and so on. And, whisper it.... but I think that there might need to be some optimisations or modifications to handsets, to make them work nicely in a femto context.



9) OK, this might be wishful thinking, but I'm hoping to see more pragmatism & innovation around the concept of mobile multiplicity.
Lots of people already have multiple phones, SIMs, numbers, identities and so on. While some purists may bemoan the inelegance of this "oh, if they just had the right offer, they'd only want one", they are flat-out wrong. People are happy with complexity. The software in your brain is much more flexible about choosing appropriate modes of communication than the software in the network. People like multiplicity. They want multiple service providers. They often don't care about multiple bills or logins. They are not loyal. So live with it - and exploit it where possible. Offer people ways to federate messages or numbering across multiple carriers. Give them ways to manage the richness. Offer them second SIMs or phones for the same number, or multiple numbers on a single device. Don't try & shoehorn them into your marketing department's polarised view of what a user experience should be. Sure, give them incentives to be stay with & extend their business with a single provider (ie bribe them), but also recognise that coercion or bloody-mindedness isn't conducive to loyalty.


10) No, No, No, No, No

- Mobile search. Still pointless. Haven't you realised by now that people just want normal, full-fat Google on their phones?
- Mobile advertising will definitely grow - but it's not going to get beyond a few % of ARPU in the foreseeable future except for a handful of segments
- Mobile centrex. I stand by what I've been saying for years. Hosted/centrex enterprise telephony has an opportunity for gaining 10-20% of total business lines. Mobile hosted services only have an opportunity for 10-20% of the 10-20%.
- UMA dual-mode services. OK, maybe we'll get to 2-3 million users in 2008. Wow, that'll really set the handset industry on fire....
- Unfortunately, we'll still have mobile industry dinosaurs referring to handsets as 'terminals', as if they were mainframe-style dumb endpoints. Mind you, it's a good filter to see who really doesn't 'get' what happens to mobility when it intersects with Moore's Law. Unfortunately, I predict we'll still have the word 'terminal' at the end of the year. (And 'seamless', too, ugh)
- Unlicenced-spectrum wide area wireless. Some people have been talking up 2.4GHz WiMAX to me recently. And of course there's the ever-delayed xMax as well. Maybe it'll work in a couple of rural areas where there's not much interference, but unfortunately I can't see city-wide unlicenced data working.
- GSMA's IPX. Sorry, but this just won't work in the real world apart from maybe some internal peering arrangements between operators. End-to-end multi-operator IMS-based QoS? Oh, please.