Tuesday, October 30, 2007
- PCs are the new mobile phones, at least from a VoIP, IMS & unified comms perspective. Huge amount of interest in porting softphones & other clients to laptops, for both FMC operators and various previously-mobile-only Internet players. This fits in with my own belief that early VoIPo3G rollouts are being catalysed by PCs with flatrate data plans.
- Everyone and his dog has some sort of social-networking application / plug-in / enhancement to their products. Think FaceBook-to-mobile apps, uploads of pictures or speech or video from handsets and assorted other permutations.
- IMS handsets (or even standardised IMS clients) are still very thin on the ground. Quite a few people are talking about "IMS proxies" in the network, which ordinary handsets can communicate with. That means not terminating the IMS session on the device itself, but on a server and then essentially using a 'thin client' or browser on the phone to display the results.
- Quite impressed with some of the work that Ericsson's doing to encourage developers to create apps on IMS platforms. Definitely some of the right thinking - I heard the phrase "two men & a dog in a garage" which is a bit of a paradigm shift coming from the telecoms establishment. I do have a slight worry that any startup wanting to get exposure to the global audience via carriers ends up being 2 men + dog + 30-person legal team. Although that said there are supposedly some options for easier routes to market via catalogues and quick 'vetting' of new services without detailed contracts.
Monday, October 29, 2007
As far as I can see, this is technically very similar to the earlier deal with Skype on the X-series products about 10 months ago. In other words, a software client on the phone acts as a 'dialler' for Skype, but doesn't actually use VoIP over the radio network. It uses a circuit call into a Skype gateway in 3's network. The earlier X-series one used iSkoot on the Symbian platform - whereas this one has an application built-in to Qualcomm's BREW framework.
I'm expecting quite a lot of this type of hybrid 'dial through' option before we move towards 'proper' mobile VoIP over the next 2-3 years. It's a lot easier to guarantee decent QoS using this approach in the short term - especially as this deal is on a cheap phone which I expect uses a basic WCDMA 3G chipset. 'Ordinary' 3G doesn't support VoIP very well - there's too much latency and things like header compression are absent.
I'm expecting operator-led VoIPo3G (probably with partners like Skype) to start to become more important from the point at which full HSPA-capable (HSDPA+HSUPA) devices become common. This is already realistic for laptop modems, and will start filtering into more high-end phones in 2008
Yes, I was right about my guesses on this - just got conformation from the 3 PR folk. It's a low-cost WCDMA-only phone based on a relatively low-end Qualcomm 6245 'value platform' chipset.
Thursday, October 25, 2007
Regular readers will recognise this as a theme I've touched upon numerous times in the past year or so. This year has seen the emergence of disruptive companies like Fring and Yeigo enabling wide-area VoIP on smartphones, and the increasing use of Skype and other softphones on 3G-connected laptops. At the same time, operators and their suppliers have been talking up the future generations of all-IP technologies like UMB and LTE, which make true mobile VoIP inevitable in the longer term for the established carriers.
I'm not going to give away any conclusions yet - but suffice to say that any company with a VoWLAN strategy had better start looking very seriously at VoIPo3G as a complement to it, if they haven't already done so. The comparative forecasts are eye-catching.
I believe that this will be the first, dedicated, comprehensive report on VoIPo3G, covering:
- Both operators' and competing independent VoIP players' business models for VoIPo3G
- CDMA/UMB and HSPA/LTE suitability for VoIP, and the likely timelines for adoption
- Different flavours and usage cases for VoIPo3G - from PTT to Multimedia Telephony to Enterprise communications
- Scrutiny of the opportunity for 'non telephony' VoIP and Voice 2,0 services
- Prerequisites and enablers, such as suitable open handsets and data tariffs
- Inclusion of phone- and laptop-based VoIP
- Analysis of likely competitive tactics of incumbents responding to VoIP rivals' efforts
- VoIP partnership options for UMTS carriers while they're waiting for LTE
- Full quantitative forecasts, with extremely detailed treatment of assumptions & variables
More details will be available shortly. But if you'd like a sneak preview, I'm moderating a panel on VoIPo3G at VON in Boston next week.
And if you'd like to register your details, and express interest in obtaining the report as soon as it is published, please email information AT disruptive-analysis.com .
Many thanks for reading this unashamedly commercial post.
Wednesday, October 24, 2007
Although they still seem to be denying it, the web is a-splutter with accusations against US broadband provider Comcast. It's apparently doing something to block uploaded files - a process intended to counter things like filesharing, especially via BitTorrent.
Well, there's a great lesson to be learnt here about false positives. Apparently, the allegedly-non-existent traffic-shaping minions at Comcast forgot to think through what else might be affected. Like IBM Lotus Notes for example. Not to mention anyone using BitTorrent for legal purposes. Or assorted internally-developed corporate applications that might stop working.
Apparently there's now some sort of fix for this, I'd guess done behind the scenes between IBM and Comcast. While IBM will have the clout (and contacts) to do this, I'll bet that corporate developers won't.
I can also envisage some fun conversations....
"Hello, is that Mr Comcast Corporate Account Manager?..... yes, this is Mr Smith from XYZ Investment Bank. Well, we've got this application we developed... it uses a modified version of BitTorrent.... it distributes equity derivative pricing formulae to our traders.... it seems to have stopped working..... we've isolated the problem to your network. You started blocking it with no warning..... Yes, it's causing us some trouble...... Yes, quite a bit. About $300m so far...... well, if you could just hang on for a minute, maybe you'd like to speak to Mr SueEm-to-Bits our chief counsel? And Messrs Sarbanes & Oxley would like a word with you afterwards".
Meanwhile, I've noticed a distinct upswing in people talking recently about putting broadband and mobile traffic inside VPNs, or encrypting it.
My view is that a lot of the more onerous terms-of-service for broadband (fixed or wireless), or heavy-handed traffic shaping, are going to evaporate under commercial & PR pressure over the next year or so. And I'm expecting quite a lot of Comcast-type attempts to be outed by people monitoring traffic flows and reverse-engineering what's causing problems.
As I've said before - I can't get especially exercised about Net Neutrality. Because anyone stupid enough to block benign traffic (even if they don't like it for commercial reasons) is going to heavily burnt. I think the market will understand providers wanting to stop network integrity being threatened - but it will show zero tolerance for vindictive or accidental 'collateral damage'.
Just remembered a little snippet I heard last week. Apparently, P2P traffic has actually been dropping as a % of the total on the Internet over the last year or so. Can't remember the exact numbers, but it was something like 50% a year ago, but only 35% now. Not because it's being throttled...
....but because something else has grown faster: social networking. Blame MySpace, FaceBook, YouTube et al.
And while certain elements are easily trackable in DPI (facebook.com's IP address etc), the proliferating 1000's of plug-in are not. Are you going to traffic shape bits of Little Johnny in Iowa's MySpace page, but not others? Who are you going to bill when you decide that the Vibrating Hamster widget on my FaceBook page is consuming too much bandwidth?
Personally I'd never opt in for SMS advertising, and I'd certainly never buy an product or service marketed in that way. I've even chased down culprits and informed them of my punitive fees for deleting future unwanted messages. I prefer any marketing, news updates etc to come through via email, where it's much easier to scan, filter & prioritise messages - and read them as & when I have time.
Posting to ForumOxford recently, I thought about what it would take to persuade me to change my mind on this. The answer was an upfront payment (in cash or credit to my mobile account) of £3-£5 per received message.
The other option would be to develop SMS clients that are similar to today's email clients, so that rather than getting a "You have mail!" alert each time a message come in, you can just look at your inbox. However the problem with this is that it would mix up texts I want to receive immediately and would allow to interrupt me (friends, work contacts), and ones which have no realtime importance (advertising). And I don't think I can be bothered to set up a filter for who I'm happy to be interrupted by, plus occasionally I get urgent texts from unknown numbers ("Running late, in a taxi, there in 15 mins").
The other thing is that SMS marketing is roughly now where email marketing was in 1996 - "you have mail!"... cool!! But once you get to 'You have 94 unread SMS' you need to be a bit harsher in your attitudes.
Hopefully handsets will soon ship with preloaded SMS antispam software, and your SMS client will have Inbox / Sent / Spam folders in the same fashion as today's webmail & desktop clients. Then you could just sift through the rubbish once a week to look for anything that's fallen through the gaps - and you could mark the updates from your bank / hairdresser / favourite bar as 'not spam' to whitelist them if you really wanted.
Friday, October 19, 2007
"Once certain interference problems have been resolved, WiMax could become as ubiquitous as mobile phones and conventional broadband, Bubley said. "The real kick comes between two and five years from now," he said, when consumers will start seeing the first mobile phone-style devices using WiMax come on the market"
Well no, that's not what I actually said, but I guess being quoted incorrectly is better than not being quoted at all.
For the record, what I actually meant about two-to-five years was that is when I expect to see large-scale WiMAX networks operational in a decent number of countries - ie after 2.5GHz spectrum auctions, network build-out and increasing device maturity.
I see WiMAX as important, yes, but I'm certainly not expecting it to scale to levels comparable with 3bn cellular subscribers or 500m DSL/cable homes.
I can't see WiMAX devices becoming as small/power-optimised as cellphones for a considerable time. I see WiMAX' main role in the next 5 years being for non-phone devices like laptops, media players, in-car systems, PDA-type products etc.
Can't say I'm overly surprised by this, especially given the recent ramp-up in announcements & perceived credibility that the technology has garnered. Cisco's reported sniffing around Navini was also a bit of a signal that it thinks WiMAX is about to get its mainstream credentials bolstered.
It'll be interesting to see if this means there will be some back-pedalling on the other parallel regulatory approach to getting WiMAX accepted - technology neutrality. Now that WiMAX is a member of the 3G club, perhaps it'll decide that exclusivity isn't so bad after all.....
Wednesday, October 17, 2007
Get a PC with WiFi and pay $39 a month for the same thing.
Presumably they're assuming that PC users either (a) have more money or (b) use more data
Seems like a great way to ensure lots more people start using their phone as a modem, to connect their PCs....
Either way, it makes a bit of a mockery of Nokia's claim that an N-series 'multimedia computer' is a replacement for a full PC. Although to be fair, nobody believed that anyway.
Tuesday, October 16, 2007
So while I'm expecting to see a continued growth of Nokia & assorted Windows devices with WiFi, plus the iPhone, the odd UMA device and perhaps VCC feature phone (and yes, probably a handful of Motos & S-Es), I really don't see massmarket devices adopting WiFi - especially when aimed at the global marketplace where not that many people have broadband+WiFi at home.
I reckon a reasonable target for 'attach rate' of WiFi into the global mobile phone market is 10-15% by 2011 in terms of shipments, so probably <10% into the installed base for at least 5-6 years. And I'd punt at an active-usage rate of probably around 3-4% in the same timeframes.
- I'm feeling smug about my longstanding predictions that mobile applications will become 'bearer aware' not 'bearer agnostic'. Symbian's new Freeway architecture is essentially a bearer-awareness layer in the IP networking part of the device
- I'm also feeling quite smug about my ongoing belief in motion (and other) sensors being integrated into handsets. Nokia's keynote presentation highlighted this, with some cool usage cases like sending a busy tone & diverting a call to voicemail by just flipping over the phone while it's lying on a table.
- Apparently there will be upcoming Symbian WiMAX devices - my bet is for Nokia to look at ones for Sprint as it's still hoping to make headway in the US. On the other hand, SonyEricsson may be looking at Japanese deployments. One to watch.
- The spec sheet for Symbian v9.5 mentions that it 'supports UMA', although I still haven't seen a UMA Symbian phone in the wild.
- Symbian still doesn't seem to identify prepay subscribers as a specific end-customer segment. Most of these users don't get subsidised phones, and don't have data plans suitable for regular low-cost (ie daily) usage of data services.
- S60 is still essentially a Nokia-only game, although Samsung now has a range of 4 new devices. Not much sign of LG, or any of the ODMs and other tier-2's.
- Less FMC and wVoIP than I'd been expecting, although there were a couple of interesting things hidden away, like an encrypted voice app from Atelier and Phil Zimmerman of PGP fame
- Apple was hanging over the event like a cloud, the word 'iPhone' unspoken but essentially meaning that everyone's scrabbling to divert game/multimedia graphics efforts over to the UI instead
My thought is that the smartphone community is slowly coming to uncomfortable realisations. Firstly, normal people don't download applications to handsets. No, it's not about 'educating ' them, they just don't/won't. Get used to it, and use the browser instead, or get vendors/operators/retailers to preload them if you're really, really good. And secondly, the ultimate addressable market for smartphones as a percentage of overall handsets is more limited than they'd hoped.
Separately - I met two people brandishing iPhones who've been using them (with European eyes) for a month or more. They both said that, against their initial expectations, the SMS user experience is surprisingly good - something I'd thought might be an impediment to its uptake. Apparently it's not.
Monday, October 15, 2007
Motorola's polyamorous relationship with most members of the mobile software ecosystem is pretty confusing. It's got its recently-renamed Motomagx Linux/Java platform with a new partnership with Trolltech, it's got the former TTPCom's Ajar featurephone platform, it's got various Windows Mobile devices (especially in its Q range and its enterprise-device division Symbol), plus various legacy proprietary RTOS platforms. It's got of Flash Lite here and there, and it's Linux platform is supposed to be pretty Web 2.0-friendly.....
This is one of the reasons I'm skeptical that mobile operators outside Japan will ever be able to restrict themselves to just a couple of supported OS's in a few years' time, despite their hopes. S60, UIQ, Windows Mobile, UIQ, Motomagx, Ajar, Nokia S40, BlackBerry OS, midrange S-E & LG & Samsung OS's, Apple OS and a bunch of lower-end platforms or proprietary/closed implementations of Linux.
I wonder what OS Motorola will put in its future WiMAX devices? Fortran, probably. Or a hamster on a wheel. Or it'll partner with Google for yet another handset software platform....
That's finally crossed the threshold of usability / value versus a typical London Internet cafe's £2 per hour, and could therefore appeal to ad-hoc users who've got access to a laptop but are loathe to sign up for a monthly contract. It also makes a lot of the current pay-per-use WiFi services look very expensive - for example T-Mobile's own £5 / hour at UK branches of Starbucks.
The interesting question is whether this will be available to travelling business users as well, enabling them to just buy a suitable SIM card when they land at Heathrow, enabling a massive saving against 3G roaming fees.
As a side note, I'm definitely seeing HSDPA becoming a viable alternative to both WiFi in the UK at the moment - although this largely reflects stupidity among WiFi hotspot operators on reaching sensible pricing levels. They've squandered a 4-year lead on the cellular industry.
HSDPA is also making inroads into home broadband in some niches in the UK - I've had a couple of friends recently mention they're looking at it rather than ADSL or cable. It's especially attractive to young urban professionals, immigrant workers or students who have a high chance of moving home more than once a year, and who therefore don't want a 12-month broadband contract. In some cases they'd also have to pay £££ to have a BT line reactivated & reconnected at a new place, which definitely makes the process more drawn-out and expensive.
But anyway, I just had a look at the Jaiku FAQ.
20MB a month is quite a lot, especially for people who don't have flatrate data. Which, outside the US, Japan and Finland, is most people - particularly teenagers on prepay tariffs. Yes, you can sign up for the capped £1 / €1 per day deals, but I can't see the average student wearing a £30 a month extra bill for this. Not to mention the fact that Nokia S60 phones' penetration into the (largely unsubsidised) prepay market is fairly low. Last time I asked Symbian, they didn't even have a strategy for targeting prepay users, or identify it as a specific target segment with unique needs.
On the other hand, maybe Google just wants the core software engines and mobile-savviness of the founders to plug into its own future social networking and/or mobile platforms. We shall see....
...so UMA can work over any generic broadband IP connection..... DSL, cable, apparently WiMAX....
...so could you run multiple UMA connections over a good-enough mobile broadband connection? Especially HSUPA which has sufficient upstream bandwidth.
There must be some usage cases for, say, 4x GSM connections being multiplexed over a single (portable) HSPA pipe. Answers on a postcard, please.
Friday, October 12, 2007
"Home Zone Service: a mobile phone in home phone’s disguise Subscribe to Home Zone Service and you would not need traditional fixed-line phone service ever again. Home Zone Service offers the convineince of mobile phones at a rate comparable to fixed-line service, at home or in your office.
Home Zone Service combines the benefits of the convenience of mobile phones with the thriftiness of fixed-line service in a single mobile handset. Home Zone Service is as easy as installing Access Point, the Home Zone Service device, in a place of choice. Whenever Home Zone Service subscriber makes a call within 30m from the Access Point, the call automatically registers at a rate comparable to fixed-line service. Up to seven mobile handsets can be operated via a single Access Point."
What's not clear to me is exactly how this works. I haven't been able to find much via Google, and unfortunately I don't speak Korean, so I have no way to interpret the FAQ on the more-detailed domestic part of LG's website.
There's a mysterious access-point widget with a 30m range that plugs into a power socket.... but which doesn't seem to have any hookup to a fixed broadband line that you'd need for either WiFi or a femtocell. I'm guessing that it's either some sort of repeater that signals to the network to flip to a lower tariff when it's in use, or perhaps it's got a CDMA modem in plus a Bluetooth antenna with cordless telephony profile (CTP). It needs a dedicated phone (there's a range of 6 according to the website)
Anyone able to shed any light?
Wednesday, October 10, 2007
In the last few weeks, I've had several 'work' contacts add me as a friend on Facebook, either as individuals or in one case as a company. Some are people I've had a drink with after a conference, so they're not 100% 'professional' contacts I've met solely in a dry work context.
After thinking pretty hard about this, I've turned them down and directed them to LinkedIn.
Personally, I like to keep my social life & my work life pretty separate. I won't take business phone calls when I'm 'off duty', and often restrict my email access my business email/web device at home. I almost never talk mobile stuff with my friends (although my father is a huge Apple fan, so iPhone discussions are inevitable).
I'm trying to come up with a reasoned criterion for someone being 'Facebookable'. If you read this blog, you're probably on the wrong side of the line. Ditto if I met you when I was wearing a suit, or if you only have the email address on my business cards.
Once again, this comes back to my theme of multiplicity. I don't want to have one 'identity' or one message inbox. You might differ... but for me, divergence is much more important than convergence in maintaining my work-life balance.
EDIT I've realised that actually there is one place where I happily blend social & business contacts.... Skype, probably because you can keep everyone you know segregated. Maybe there's something in that for eBay to ponder....
(For my US readers - that sort of fashion faux pas would probably get you turned away from a trendy bar in London or Milan.... it's the European style-mistake equivalent of what I'm told are called pocket protectors)
There are a few reasons I can think of for this. And despite the swarms of tourists in my part of London I'll refrain from commenting on any sweeping or offensive stereotypes.
- Firstly, the US embraced PDAs much more vigorously than Europe, so the idea of some sort of handheld computer is much more engrained and accepted.
- Secondly, the preference for email and IM, and the later/slower adoption of SMS has reduced the number of people mentally-tuned into writing with multitap vs. Qwerty
- Thirdly, the BlackBerry phenomenon has run much longer, and wider, than it has in Europe. I see many more non-corporate 'civilians' with them in the US.
- Fourthly, the US historically had quite a high adoption of pagers relative to Europe.
But perhaps most interestingly.... geeky devices in the US have cool names. Hiptop. Dash. Curve. Q. Mogul. Blackjack. iPhone.
....compare those with MDA Vario III, XDA Executive, HTC S620, iPaq hw6910, Nokia E61i, Samsung i320, M3000 and so on.
Can you imagine a stranger ever coming up to you in a bar and saying "Oh wow, is that really a v1605?".
Tuesday, October 09, 2007
"Nokia will customize the new Multimedia Menu to provide easy access to Telefónica services.. Similarly, Telefonica customers will be able to access all Nokia services through the Multimedia Menu"
I presume this opens the way for things like Ovi, and maybe Nokia-optimised versions of popular web apps like Flickr. Using Nokia as an intermediary could be a way around the current problem with the mobile/web collaboration model - web services companies really don't want to do 100 separate operator-specific optimisations, and 100 separate negotiations, to get their application deployed widely (and hopefully virally).
"Our customers can take advantage of Telefónica's high quality mobile broadband network to access both Telefónica and Nokia services, as well as other third-party applications"
What's not immediately obvious is whether this just applies to Nokia phones (and if so, whether it includes S40 featurephones), or if it relates to S60, in which case it could apply to Symbian-powered devices from Samsung, LG and others.
Also, looking out a couple of years & flying a kite here.... I wonder if we might eventually see a Nokia VoIP service as well, given how good Nokia is at integrating VoIP clients onto devices - especially if Nokia positioned it as a 'non-primary telephony' VoIP service (eg for voice-enhanced IM, conferencing, gaming etc etc) and enabled developers to access it via convenient APIs.
Monday, October 08, 2007
a) Having a mobile phone, with the convenience of an integral phonebook, camera etc
b) Having the ability to make & receive personal calls from wherever you are, send SMS and emails, browse the web etc
c) Having the ability to do (b) while you're actually moving - for example in a car or bus
I'll hazard a guess that the answer would reveal perhaps a 50/40/10 split of perceived value. Yes, there's probably better ways of phrasing the question, but you get my point. The actual 'moving about' part of mobility isn't that relevant unless you regularly make calls while driving - and ironically, the move towards data services exacerbates this.
This is all a bit of a problem, as mobile operators have traditionally priced calls on the basis of a premium which essentially blends all of the above. That customers value the handsets themselves as much as the service isn't news - it's at the core of the ongoing tug-of-war between phone manufacturers and operators, as to who 'owns' the customer relationship.
More relevant is the distinction between nomadicity and full mobility. The advent of VoWLAN in all its guises, cellular homezones and assorted other location-based mechanisms for differentiating call prices has been apparent for the last few years, in the guise of FMC.
But with moves like BT's tie-up with FON last week, as well as the nonsensical hype around metro-WiFi, it's clear that nomadicity (and cheap/free calls on mobile devices) is taking a step further to achieving 'B' above.
I wonder, though, if the push for geo-specific pricing is actually a special case of what should really be in place: velocity-specific pricing. Maybe we should be pricing mobile communications based on how mobile you are. A call when you're stationary is cheaper than a call when you're walking, which is cheaper than one made on the bus/in a car, and so on to trains & planes. If you're on a femtocell in your bedroom, or connected via WiFi/FON in your neighbours house, you shouldn't be subsidising the calls of the salesman in his Mercedes driving 80mph down the M4.
Now obviously there's a world of pain in communicating all this - especially as even though customers might only value high-speed mobility at 10%, they've grown up to expect it, on the odd occasions they need it. But with the onward march of GPS in handsets, motion sensors and so forth - even just working out a given customers' average number of cells-per-call - it ought to be possible to measure who's really deserving of paying the mobility premium, and who should be on the 'nomad' tariff.
Wednesday, October 03, 2007
In common with most of these, it also has the usual backside-covering 'fair use' terms and conditions.
"Permitted Uses of theO2 Web Bolt On are uses of your SIM Card within a handheld mobile device for the purposes of Internet Browsing and email only. Any other use of the O2 Web Bolt On will not be a Permitted Use, including but not limited to:
Use with Data Cards or Modems;
Point 2 Point file sharing and file transfer,
VoIP (e.g. Skype™),
Video and TV streaming,
Use in conjunction with routers."
And I fear that using the Internet without somehow being 'in conjunction with routers' may be a little tricky. Especially if O2 launches a femtocell / home gateway.
You get my point. 'Hard' definitions don't work very well with traditional Internet applications, and are even worse with AJAX, Widgets and Mashups. Yes, of course I know what the T's and C's mean in spirit, even if it is mean-spirited.
Put simply, I don't think any of this stacks up, apart from instilling 'fear, uncertainty and doubt'. I'm not picking on O2 specifically here, most of the other operators, with the notable exception of 3's X-series, do something pretty similar. Essentially they're writing themselves a 'get out of jail free' card.
"Fair" in 'fair use' seems to mean "uncompetitive with anything else we do already or might do in the future". "That is, if it is used by enough people that we decide to take some action." So "Unfair" = "Competing effectively".
The strange thing is that they haven't applied equivalent rules to ordinary voice calls. Surely, taking this to its logical conclusion, it's not 'fair use' of my O2 voice minutes if I use them to call Vodafone's sales department, or Orange's analyst relations team?
Personally, I don't think these onerous but vague and sometimes threatening data plan T's & C's will stand up for much longer - except in the case of really bandwidth-heavy or malicious apps which can threaten network integrity. I expect most operators to have quietly moved away from these positions by the end of 2008.
Tuesday, October 02, 2007
.....Nokia N95 with GPS and 8GB of memory and a memory card slot..... Navteq's giant mapping databases.... Nokia developing its non-carrier channels to market.....
Look at TomTom or Garmin or maybe even Tracker: mobile location is mostly about standalone products. Services are mostly limited to incremental add-ons: updates, or critical realtime traffic info, plus a tiny and sexy but mostly non-revenue slice of geolocation-meets-social-networking mashups. Not only is downloading maps or directions OTA to a handset slow and expensive, it also consumes valuable power.
Bottom line: mapping on mobile doesn't need an operator or very much over-the-air location services. Maybe in the future as upsell, but not as a starting point. The pricing/performance curve on flash memory is much steeper than that for wireless broadband.
Interesting, this is totally different from things like Google Earth. For that, a Web 2.0-type model works fine. Mobility is less important, bandwidth is abundant always-on and essentially free, more granular data is becoming available continually and there are no power constraints.