Monday, March 30, 2009

Control of subsidised FTTH will bring structural separation to Telecom NZ

New Zealand is getting closer to its ultra-fast broadband dream. The government issued a 'draft proposal for comment', basically to set up a PPP for constructing the network.

As promised earlier, the government will contribute half of the cost (NZD 1.5bn) of building out an open access FTTH network to 75% of the people (in 25 towns - note that 29.5% of all homes are in Auckland). During the first six years, focus is on schools, businesses, the health industry and greenfields. However, FTTH must be deployed within 10 years.

The big issue of course is: what will Telecom NZ do? Will they overplay their hand the way Telstra did in Australia? The document clearly demands structural separation of Telecom NZ, should it want to invest in the passive infrastructure; unless it would hold a minority stake (comparable to KPN holding 41% in the new Reggefiber Group), in which case functional separation (which it already implemented) would suffice. In other words: if Telecom NZ steps in with a minority share, it wouldn't have to change its structure, but (unlike KPN) it would need to be structurally separated once it would gain a majority stake in any LFC.

Here are the Key Principles from the document:
  • making a significant contribution to economic growth;
  • neither discouraging, nor substituting for, private sector investment;
  • avoiding entrenching the position, or ‘lining the pockets’, of existing broadband network providers;
  • avoiding excessive infrastructure duplication;
  • focussing on building new infrastructure, and not unduly preserving the ‘legacy assets’ of the past;
  • ensuring affordable broadband services.
Here are some conditions:
  • The vehicle for investing the subsidy will be crown-owned: Crown Fibre Investment Co (CFIC). It will invest, alongside co-investors, in Local Fibre Cos (LFCs). CFIC will hold up to 50% of the shares of the LFCs.
  • "Selection criteria will be focused on several aspects – the amount of additional fibre coverage being proposed, the proposed capital structure (including the parties’ relative capital contribution requirements), the commercial viability of the proposal, consistency with government objectives, and the track-record of the partner."
  • "The government’s shareholding may be concessionary, and in particular may be subject to a lower rate of return than the partner for an initial period (for example, up to ten years). These provisions will be negotiable."
  • "LFCs will not provide retail services. However, the government will not exclude partners that own or operate telecommunications retail operations, but such partners may not have the majority of voting control on the board of LFC (unless they divest themselves of any retail business). Telecom, and other telecommunications operators with retail operations, will therefore be able to participate in the contestable selection process, subject to the above requirement."
And here are some more details:

Time line
Comments due end of April, report back to the government end of May, appointmnet of the vehicle mid June, RfP to be released mid August, proposals due mid October, initial decisions due January 2010. All submissions, due April 27, will be published at www.med.govt.nz/broadband.

Wholesale
The network owner will primarily sell dark fiber, and "potentially other approved wholesale broadband services. (...) LFC may: provide a wholesale bitstream service; and enable the provision of interim solutions by wholesale customers, such as wireless last mile or ADSL2+ or VDSL2 solutions, provided that this is consistent with the LFCs achieving the government’s objective of FTTH within ten years; and subject to the CFIC’s approval, provide any other wholesale broadband service."
"The government investment will be in fibre networks that will operate only at the wholesale level, selling dark fibre based services enabling telecommunications providers to design and specify their own downstream services. This approach will ensure that all decisions regarding active network technology options are left to private sector investors."
"By keeping the new fibre business out of retailing, it will have no incentives to act anti-competitively, and there will be little need for regulation of its prices. In fact, there will be considerable initial incentives for it to keep the fibre rental prices low to facilitate use by downstream providers."
"The new network will provide dark fibre services to any ISP or telecommunications service provider, and will operate as an infrastructure ‘utility’ at the passive level of the market. The aim is to provide a new fibre platform upon which service providers can develop their own services and create unique, innovative offerings."

Services
The usual suspects are there, including: "There is also a strong likelihood of new applications being developed in the future that will require residential users to have fibre broadband connections to operate them effectively, particularly as increasing numbers of services are delivered digitally."

Thursday, March 26, 2009

Shopping time for eBay

At its recent analyst day, eBay laid out plans for all its units: MarketPlaces, PayPal and Skype.
PayPal and Skype combined are targeted to equal the MarketPlaces business in terms of revenues by 2011 (up from 35% of eBay in 2008). Skype should double its revenues almost to $1bn in 2011 (from $551 in 2008). Skype's margin must be grown to 18-20% (from 11% in 2008). PayPal too has has for doubling revenues and restoring margins to the 18-20% range.

An interesting part of the statement was that eBay now targets the secondary market, estimated to be worth $500bn worldwide, including liquidation inventory.

Now, coincidentally I ran into just the site that eBay could single out for an offer: Troostwijk. The company was founded in 1930 and they are just the place to go if you are looking to fill your newly acquired real estate. I have no financials, but somebody registering and tracking what's going on for a while, could get a pretty clear picture of the value of the merchandise passing hands on the site. And the interesting thing is: such auctions deliver the auctioneer a hefty 16% fee, with not too much risk and a self-explanatory web site.

Sunday, March 22, 2009

ISPs meet MLTA

Here is a link between two recent posts.
  • MLTA (multi-level TV advertising): This is not like the Red Button (interactive commercials). Who watches TV to see commercials? Right, nobody! People watch TV to see TV shows. The box (Daily Media) and the software (Dynacast) combined create a split screen reality, offering viewers a much more immersive TV experience, with information, sharing, communication and, last but not least, highly targeted advertising. And all of that via the remote control. Hence the term: Zapping 2.0.
  • ISP differentiation: Video may come to the rescue of ISPs trying to differentiate (the flat fee business model is killing them, remember?). They could strike a deal with content providers (YouTube, catch-up TV) to provide subscribers with enhanced picture quality as part of a premium subscription.
The link: ISPs could team up with Daily Media for a host of applications, plus MLTA.

The situation is becoming worse for ISPs offering their services over FTTH: how to justify a premium price when all you have to offer is a triple play? If the ISP teams up with Daily Media/Dynacast, interactive TV could very well become a driver of FTTH.

Clusters of drivers of FTTH

This post is for future reference, summing up the drivers of FTTH (or NGA in general), brought together in clusters.


A. DEMAND
First, applications as drivers of FTTH. There is not one killer app for FTTH - there are many. Convergent applications (see #3) demand special attention - what's next?

1. Video
After data and voice comes video. Except, the internet wasn't designed for it. The rise of 3-D, HD and holography makes matters worse. These applications are run in traditional locations, except the latter that is about public screens.

Applications:
  • Broadcast TV
  • VoD
  • Catch-up TV
  • Embedded video (tele- and e- apps)
  • City displays
  • Computers with embedded 10 Mp cameras for video calls/conferences
2. Everywhere, everybody
We do not want to be limited by our home or office.

Applications:
  • File sharing, community efforts
  • Sling box (place shifting)
  • Cloud computing
  • Online storage, back-up
3. Mobile backhaul
More stuff that the internet wasn't originally intended for. Essentially, the broadband connection can be seen as a form of backhaul for new services.

Applications:
  • Fixed/mobile convergence: femtocell (backhaul of mobile traffic)
  • Fixed/mobile convergence: as wireless networks go 4G, more base stations will be built and hence more backhaul is needed. This will go on to the point where mobile backhaul can be seen as an overlay of fiber backbones.
4. Connected devices
The internet increasingly reaches beyond PCs, laptops and even smartphones (which may connect to fiber via WiFi).
  • TV/internet convergence: interactivity (BB/internet content added to the TV experience on a split screen: the Daily Media box with Dynacast technology - see next post).
  • Widgets make the internet accessible on a new range of devices, even low-end handsets.
  • Digital cameras, e-books and e-readers (such as Amazon's Kindle), navigation gear, digital TVs, STBs: a broadband connection is often supplied. This has consequences for the network, esp. in the upstream.
5. Networked hours

Due to multitasking, the nu mber of networking hours rises from 36 in 2008 to 48 in 2013 (Cisco, June 2009)
  • Active multitasking: e.g. reading mail while gaiming
  • Passive multitasking: e.g. watching one TV channel and simultaneously recording another


B. SUPPLY
Second, operator centric reasons for migrating to FTTH.

1. Intrinsic
Fiber beats copper and HFC.
  • Quantitative criteria: bandwidth, symmetry, latency, loop length
  • Redundancy (we hate waiting - "too much is just about right")
  • Aesthetics (no sat dish needed).
  • HFC (and copper) is not (always) unbundled, so innovation at the active layer (for community services e.g.) is not possible.
2. Timing
Why now?
  • Growth is high
  • Roll-out takes a long time
  • Every home needs to be connected because people move all the time
  • Resources (funds, labour) are scarce.
3. Financial
Capex may be high, but there is more.
  • Capex is continually falling
  • Opex is very low
  • Green (save on travel, carbon emissions)
  • GDP grab (attract business)

Friday, March 20, 2009

NGA news: Australia, Virgin Media and 3-D

Hot news:
  • Australia NBN: Acacia is the front runner, rumour has it. It will bring along Leighton Holdings, the owner of both a backbone and a maintenance unit. Acacia is backed by wealthy individuals, an important asset today, including: Doug Campbell, Solomon Lew, Doug Shears, Paul Bassat, Andrew Bassat, Matthew Rockman, Steven Skala, Leon Kempler, Lawrence Paratz.
  • Open acces to cable broadband: Sky is trying to open up rival Virgin Media. That would introduce unbundling in the cable domain (as it will at TDC Cable), i.e.: regulatory symmetry. Virgin Media doesn't have nationwide coverage (as BT does), but I fail to see what that has to do with it. It's about consumer choice and regulatory symmetry.
  • 3-D: A nice overview in this Wall Street Journal article. Jeffrey Katzenberg estimates that 3-D will enable movie owners to raise ticket prices by $2-5.

Wednesday, March 18, 2009

FTTH snippets from around the world

Several FTTH related news items caught the eye over the past few days. More about the Dutch market in the next few days/weeks, rumour has it. For your diary: Singapore is set to decide on the OpCo of its NGNBN within the next few days; Australia is to pick the winners of its NBN RfP.
  • Green: Verizon FiOS uses just 38% of the energy that its copper network uses.
  • SMP. The proposed KPN/Reggefiber joint venture in the Netherlands is drawing criticism. Both cable operators and unbundlers are preparing legal action. One can only guess what it's about, but I suspect the wholesale tariffs for access to the passive infrastructure are way too high and will be contested.
  • Open access. Bell Canada is fighting it, Ziggo is denying the business case (he prefers a duopoly) for it and TDC Cable will have to allow unbundlers onto its network. The latter case will be interesting to follow, to see if this technically and commercially works.
  • Cable. Kabel Noord, a small MSO (23k subs) in the Netherlands is embracing FTTH for new builds and fully acknowledges the benefits (bandwidth, symmetrical, not just VoD but TVoD as well).
  • Munifiber. Alas, the Palo Alto project fell through. First, Axia's financier quit, and then (when Axia asked the city to pony up the money) Axia itself hung up.
  • M&A. Telefonica supposedly is close to a deal to buy both Hansenet and United Internet. It will be interesting to see how they will fight an incumbent that is equally conservative as the Spanish mothership.
  • In-building. Ericsson has demo'd 500 Mb/s over 500 meters, using 6 bonded pairs. Sounds a lot like solving the in-building problem (for a much faster roll-out).
  • Services. Anderson (Indiana) is contributing a service, as part of the answer to the question: do we need all that bandwidth. It's called storage.
I will be meeting a bunch of fiber people over the next few days, from OBR (munifiber Rotterdam), Reggefiber and PacketFront, so keep posted.

Monday, March 16, 2009

MLTA: strengthening the marriage of TV and the Internet

My friend John Goedegebuure is taking the company and product that he works for, Daily Media, to the next level. I have written about the mediacenter/set-top box several times before, but the new direction is definitely worth taking a look at. You can read an article from him here, explaining the new direction.

First a few words recapitulating the box and the services.

The Daily Media box brings internet content to the TV, much like comparable offerings from Apple (Apple TV), Microsoft, Sony, or start-ups like Sezmi and MyBroadbandTV. Internet-based content, available through a broadband connection, is streamed to the TV and services are added. Daily Media however takes a different approach when it comes to marketing the product: the business model is such that the box is free for consumers and a third party is engaged to carry the cost (some EUR 150 per box). Further, the product is entirely plug-and-play.

Just about any company could join the partner program, but they will typically share a few traits:
  • They will have a subscriber (or user) base of some sort, or are trying to build one,
  • They will be looking to raise customer loyalty (or reduce churn),
  • They will be trying to differentiate in a tough (commoditised) market,
  • They will be looking for new revenue streams (from targeted ads delivered through the box and shared revenues from services such as VoD).
Partner companies could be telcos, cablecos, ISPs, TV networks, insurance companies, pizza delivery chains, newspaper companies, you name it.

The new direction involves Multi Layer TV Advertising (MLTA), which takes the marriage of internet and television (including existing red button applications) to the next level. The user experience is still entirely TV-focused. Instead of TV shows directing viewers to their computers ("find out more at www.dot.dot"), the new technology empowers the viewer to remain seated on his sofa and control both the TV and related internet services using just the remote control (or a keyboard). In other words, it's an integrated multimedia single (but split) screen experience.

As such, it could even be compared to the amazing 'sixth sense' device shown here (thanks Vincent), which one day will be the entirely mobilised version: both devices enhance our reality with internet-based content. In John's case, the technology ('Dynacast') is from Harris Broadcast. Read more about it here.

Obviously, the model involves advertising, which will be much better targeted than traditional TV ads and hence carry a much higher ROI and CPM (or CPA). But, make no mistake: if MLTA is sufficiently successful, ordinary TV ads could even be reduced, enhancing the viewer experience even further.

One last word: if this takes of, there's yet another reason to build NGA networks.

Tuesday, March 10, 2009

ISPs can differentiate in their video offerings

One question that keeps being asked is: why don't ISPs move away from the flat-rate pricing model? Good question of course, and not so easy to answer.

ISPs started this offering when always-on broadband made its appearance, most notably at a time when video wasn't part of the internet yet. It kickstarted the adoption of broadband, especially so because consumers don't have a sense of what a MB or a GB is.

Now, here are two options to do something about it.

1. Catastrophe. It was calculated that YouTube runs up a $1m bill for transit each day, whereas it only makes $0.1m in advertising. Just suppose Google would pull the plug, or threaten to do so and only give ISPs access who are willing to pay - reversing today's battle, because currently ISPs are trying to charge Google for 'giving it access to their pipes'.

2. Differentiation by providing HD video, VoD or BB video to the TV. I wrote about the Daily Media box several times before, including this option for differentiation. They were also covered recently by Gerry Kaufhold. Now, there is yet another box: Zillion. Again, a different business model from Daily Media's. Let's wait and see who scores the first ISP deal ...

Structural separation revisited

Apparently, the regulator nor many a competitor thinks structurally separating the incumbent is needed. In the Netherlands, the situation could be the same. Still, I believe there could be a point in moving the issue up the political agenda. Yes, it's a huge issue, but precisely that may require huge steps.
  • Half of the employees of this regulator are former KPN employees. Most munifiber networks have recently become KPN partners. Are the representatives still objective?
  • Structural separation is seen as a remedy, whereas I would see it as good for business. As long as this doesn't change, structural separation will likely not happen. Unless the incumbent totally screws up (Australia), or regulation suddenly changes (US). KPN is well respected (perhaps even too much so), also among competitors. It has a good wholesale portfolio. Further, competitors are not asking for separation. And any competitors still around seem to have doubtful commitment to the Dutch market (BBned was put up for sale by Telecom Italia; Tele2 is facing large investments in mobile and broadband but hasn't made any commitments yet; Online doesn't sit very well in T-Mobile's portfolio, but there are simply no buyers).
  • Separation is unlikely as long as regulators think that competition is at a decent level. In the broadband market, KPN has a 45% share, cable has 40% and unbundlers have the other 15%. The EC is rightfully worried over ongoing incumbent dominance. It looks like the 15% share of altnets will be going down, especially when the market moves toward FTTN/VDSL.
  • The advantages of structural separation are not about pricing only. Incumbents like to drag their feet; let's not forget that Openreach's P&L still is included in BT's. Also, separation opens the way to attract third-party funding, or even nationalisation.
  • Why do incumbents object to structural separation? Not just because of the disruption and the one-off costs. Surely, it must be because they fear the loss of any synergy benefits of being vertically integrated. And that is precisely why they must be separated: this inequality of enjoying these benefits will only go away once the incumbent is separated. In other words, if incumbents object, they implicitly say they have advantages over competitors. This is not good for true, long-term competition.
  • If you think that proper wholesale prices are the way to avoid structural separation, look again. BBned (active operator in Amsterdam's Phase 1) pays only about half of what is proposed now by her employer (14.50-17.50 EUR/mo/line for ODF access to passive FTTH lines). Going forward, active operators and service providers will have to recoup about 10 EUR/mo more from their customers than currently is the case in Amsterdam. Does FTTH have this kind of pricing power? It certainly paves the way for low margins at service provider businesses. KPN will be the only service provider that can afford this kind of pricing.
  • KPN will not ony be a service proviser, but it also is co-owner of the passive layer (with a call option to a majority stake), and possibly the monopolist of the active layer (which is not regulated), which could lead to serious re-monopolisation. I would like to call upon the regulator to make sure that the active layer doesn't turn into a monopoly (as it will in Singapore). If it does, the wholesale tariffs mentioned above should fall. Further, an active operator monopoly is not good for competition, because true service differentiation arises on the active level. Otherwise, we are stuck with WBA only.

Telstra gears up HFC against the NBN

Telstra is launching the roll-out of DOCSIS 3.0 on its five city HFC network. They are targeting 100 Mb/s to 1m homes in Melbourne by Christmas and claim it is upgradeable to 200 Mb/s. Total spend in 2009: AUD 300m.

Some remarks:
  • Apparently, this is the way forward for Telstra competing the NBN, which will have a winner by the end of the month. Looks like Telstra is definitely not going to be it. And if it will get involved after all, horizontal separation (spinning off of BigPond) should be next.
  • Let's not forget about the inherent limitations of HFC networks: highly asymmetrical (limited upstream spectrum and no gear available); unbundling is ruled out (from a technical perspective); not quite as much bandwidth (compared to FTTH), to be shared by quite a few more homes (compared to FTTH).
  • The HFC network and the future NBN (FTTN + VDSL) will be on a par, especially since a VDSL-network is tough to unbundle too (but from an economic perspective). However, the next logical step (FTTH) is pushed into the distant future by this development (unless Axia or Acacia wins the NBN tender). That will not be good for Australia on , say, a 5 year timeframe.
  • Interesting: "... infrastructure that complements our world-leading Next G mobile broadband network." (Not the other way around.)
  • No vendors mentioned (Cisco? Motorola?).

Wednesday, March 04, 2009

Twitter is like 'rich man's SMS system'

Google's Eric Schmidt described Twitter as 'poor man's email system'. It is usually described as microblogging, but pulling it into the realm of communications can be enlightning.

I think of communications in a three dimensional matrix:
  • Along one axis is the type of message (text, voice or video).
  • Along the other is the level of immediacy, ranging from immediate via somewhat delayed to almost time-insensitive.
  • A third axis could have the type of device by screen size: inch (mobile device), foot (computer) or yard (TV).
Since I don't have any 3-D skills, this what would describe the matrix:

Text:
  • The immediate form is chat (inch or foot).
  • SMS (inch) is somewhat delayed.
  • Email (foot or inch) is even more time-insensitive.
Voice:
  • Traditional voice calls (inch, foot) are immediate.
  • Voice-to-SMS conversion (inch) is the delayed version.
  • Voicemail (inch) is the least time-sensitive.
Video:
  • Videocalls (inch, foot or yard) will be immediate.
  • Recorded videocalls could be alerted for via SMS for a somewhat delayed experience.
  • Recorded videocalls could also be vodcasts or YouTube clips (least time-sensitive).
Now the question is: where does Twitter fit in? To me, it feels much like SMS. What it adds to the matrix is that it isn't just an inch app, but a foot app too.

Finally, what is missing in the matrix is mainly yard-versions, such as reading email on TV, but obviously that may not develop into the mainstream. SMS delivered to the foot-screen (computer) is probably a good idea (examples, anybody?). Voicemail to the inch-screen could be done by podcasting.

Tuesday, March 03, 2009

KPN's case for FTTH

The really cool promo that KPN's Ed Vermeulen ran during the Rotterdam Fiber Lab, back in January, landed on YouTube.

Monday, March 02, 2009

FTTH beats DOCSIS beats VDSL

Comparisons of FTTH and DOCSIS usually aren't very thorough. So I got some industry intel to help me better understand the differences. Here's a first shot at providing a better comparison.

The conclusion could be that FTTH is superior, but upgraded HFC networks will go a long way. Telco incumbents have a choice: sweat the copper plant (some running into a huge problem because VDSL doesn't handle analogue TV very well); or leapfrog cable and swallow the capex pill.

Shared bandwidth
Fiber guys usually point to the fact that cable is a shared medium. Well, alas, all networks are shared. The real question is: how much bandwidth is shared among how many subscribers? Verizon answered that question recently for its FiOS network. And that's 'just' a PON network! In other words, fiber beats cable, but it is shared as well.

Symmetrical speeds
Fiber guys also tout the fact that FTTH enables symmetrical connections, and the cable guy usually retorts that DOCSIS can do that too. I'm not sure how it is in other countries, but in the Netherlands the fiber guy is 'more right' than the cable guy. Coaxial networks offer a full spectrum of some 862 MHz, which would deliver several Gb/s if it was all used for broadband (DOCSIS). Alas, most is reserved for cable TV (see below: cable TV v. IPTV). Moreover, only the bottom 60 MHz is reserved for upstream, and so upload speeds are constricted.
Even if cable operators decided to go symmetrical, it would require a large capex and opex bill. Moreover, symmetrical gear is hardly available.
Further, a quick calculation, assuming 30 TV channels (occupying 8 MHz each) leaves the equivalent of some 45 channels for DOCSIS, or around 1.8 Gb/s (shared!).

Cable TV v. IPTV
Cable TV is like pushing all TV channels all the time onto the subscriber's premises. Copper could never handle anything like that, so IPTV was invented, delivering only the channel that is being watched. It is a complicated technology and certainly in the Netherlands, analogue TV (the standard line-up of some 30 channels) sells particularly well because it doesn't require any STBs. The signal can easily be split by the subscriber, usually even without an amplifier, to service multiple TV sets in the home. Digital TV needs an STB, which is a major inhibitor in box-fobic countries such as the Netherlands. (It must be said however that people are getting used to them, because cable, DTT (Digitenne) and sat TV push them to consumers. Still, each TV set needs one, and usualy they are not all free.)
Further, cable has SDV technology to mimic IPTV, and free-up spectrum, but it looks like analogue TV is a must-have for cable, even if it were just for the profit: the total rights bill is some 2 EUR/mo/sub, whereas the retail price is around 15 EUR/mo/sub. Which is interesting for several reasons: 1. cable operators are saying that they want to switch it off; 2. Opta is preparing regulation for open access to cable TV; and 3. FTTH-networks have learned that they need to offer analogue TV (over a separate fiber!) to stand a chance of winning subscribers.

Latency
Apparently, both FTTH and DOCSIS score verry well, normally, even if FTTH is a little better (c. 1 ms).

Loop length
FTTH will work over 10 km without needing any equipment, whereas a cable network requires an amplifier no further than just 100 meters away from the subscriber. However, cable networks have had their VDSL-like upgrade long ago, pushing fiber quite deep into the networks. In the Netherlands: from Headend to Distribution Nodes and further down to the Fiber Nodes. From there coaxial cable takes over, first to a Group Amplifier, and (via a handful of amplifiers) on to a Final Amplifier and then to the subscriber's homes. On average, some 1500 homes are connected to a Fiber Node (over coax and 5-7 amplifiers).

Optical/electrical
The farther fiber is pushed, the farther weigh the benefits. FTTH is superior in that sense (see before: Loop length). The optical signal is only converted to an electrical signal in the subscriber's home. That's nice for FTTH operators, reducing their electric bill, but naturally it adds to the subscriber's electricity bill. Also, the overall electric bill is higher in FTTH networks for being less centralised.

Opex
Despite the electricity savings, opex differences apparently are small for FTTH and cable/DOCSIS. Take a look at this document.

Sunday, March 01, 2009

Mobile broadband is threatening Israeli ISPs

An ancient question is gaining relevance: will mobile broadband be able to replace fixed line broadband?

At first glance, the answer could be no. The bandwidth difference between fixed and mobile networks remains large. In fixed networks, 1 Gb/s is a reality now in Hong Kong and Lattelecom is working toward half of that by the end of the year. In mobile, Telstra is leading the pack with 21 Mb/s and upgrading to 42 Mb/s later this year. That's still a large difference and it doesn't look like it will be getting any smaller soon, in a relative sense (Lattelecom already talks about 10 Gb/s, and LTE is some years off).
However, without a video component, mobile broadband looks more and more like a fixed line replacement. HSDPA USB modems have come to the market and gateways are avaliable for connecting up to 4 PCs to the mobile network. The addressable market could be limited (probably business travelers) as broadband is becomig pretty universal and heavy video consumption or P2P file-sharing is not going away (so, giving up on your fixed BB connection doesn't seem realistic). Also, mobile broadband carries much higher opex than fixed broadband, I would suppose.

Now, in Israel operators are asking the government to prevent mobile operators from providing internet services. According to this article, Partner and Cellcom are trying to sell mobile broadband services through a gateway. The operators crying foul are asking for 'structural separation' (I think horizontal separation would be a better term) of mobile operators and ISPs. Interesting development, underpinning the view that mobile broadband is indeed becoming a fixed line replacement.