Ring! Ring! Hot News, 25th March 2008

Posted on

In Today's Issue: 37% of Ultra-Mobile PCs to get WiMAX; Virtual PBXs could eat your business customers; low-cost telepresence like low-cost spaceflight, i.e. not very; MSFT buys callcentreco; Don Price on managed services; topology aware P2P; variable speed limits for the Net; price war rages; i-mode fails in Europe; huge telcos win huge telco auction; epic Aussie brawl over WiMAX; Sprint's new core network - platform perfection or IMS infection?; Vodafone & MTN; French FTTH; Deutsche Telekom disaster; sickening "human skin" phones.

37% of ultra-mobile devices to fit WiMAX. So says Intel -- but then again, how big will the market for ultra-mobile PCs really be? Time will tell...

Virtual PBX providers are building the platform; developing Web-based click-to-call APIs and virtual call centre capability. This is a crucial area of interest for Voice & Messaging 2.0; they're explicitly targeting businesses that are too small to afford one of the Big Vendors' enterprise VoIP systems, but who need functions that are more sophisticated than the ones the telcos offer (which is to say, not necessarily that sophisticated at all). There's another of those undemarcated frontiers here, between Virtual PBX/IP CENTREX, hardware vendors, VoIP software developers, and mobile operators' business products.

Similarly, Hewlett-Packard just launched a new, low-cost telepresence suite targeted at smaller enterprises. Telepresence - it's a staple of everyone's vision of the future. Basically, it's souped-up teleconferencing, verging on virtual reality; a typical suite involves a wall of very high quality video screens, surround sound speakers, and cameras, wired into a device that uses SIP to control a videoconferencing session. But low-cost is a very relative term here; the list price is $120,000, which is a sizeable dollop of CAPEX for any smallish business, or of course any smallish division of a bigger organisation. With a few well-known exceptions, we're sceptical of telepresence in general, and we suspect Cisco and HP probably have it sewn up; rather, the opportunity is in making the context of telephony visible. Microsoft thinks so: they just bought a call-centre software firm which gives them the assets to integrate the consumer and enterprise environments better.

The increasing co-opetition between hardware, telecoms, and services also applies to network operators themselves; Informa Telecoms.com has a fascinating interview with Don Price, CTO of Bharti Airtel on developing a major emerging market network with managed services (in their case from Nokia Siemens Networks), and how to manage the critical relationship with the services provider, who after all owns the heart of the business. Counterintuitively, he reckons it's better to buy the kit from the services provider's parent company than go for diversity; that way, in the event of any conflict with services, you can threaten to chuck the X billion in hardware to Ericsson...

There's more here on Verizon's effort to make nice with the P2P world. The Register makes the interesting point that there's a divergence of interests between the telcos - defined as the RBOCs and friends - and the cable operators here. Cable operators have TV genetics, and their revenue cash cow is TV, or increasingly, video content delivered like TV. Their key business relationships are with the studios and TV networks; no wonder they really hate P2P. Telcos don't have this vested interest (yet); perhaps topology awareness will make P2P a competitive differentiator for telecoms? See also this story on the Guardian Online blog about a proposal from BT Research on this. Since 1987, the Internet's Transmission Control Protocol, TCP, has an inherent mechanism to deal with congestion; really simply, the TCP stack in a user device tries to send as fast as possible, until it either reaches maximum speed or starts to see lost packets. When packets go missing, it backs off, slowing down until the loss stops. The effect is that all the TCP users on a given link slow down until utilisation gets back under 100 per cent, sharing bandwidth equally between all the TCP connections involved. Fair's fair.

But this is dependent on the assumption that all users have roughly similar numbers of open connections; if one user opens a lot of them, their total bandwidth is their share multiplied by the number of connections they open, so they can grab more than their fair share. This tends to happen with peer-to-peer applications, which by definition need at least one connection per peer. Bob Briscoe of BT Research suggests that, instead of sharing bandwidth between connections, it should be shared between users. (We blogged about this back in August.)

Failing that, there's always the price war. Virgin Mobile USA announces a new tariff with no standing charge, so somewhere between PAYG and contract service. But how far are you going to go? Motorola just fired half the staff at its UK devices R&D operation (and Carl Ichan is suing); Vodafone gets rid of 10% of head office head count, but claims it's simultaneously hiring customer facing staff. BT and 3UK, meanwhile, take OFCOM to court over termination fees. Perhaps we need...a new business model?

But it's probably not i-mode: E-Plus just canned its i-mode service, with the result that the only i-mode network in Europe is now in Romania. The users have been flipped over to a flatrate data tariff instead. The Mobile Web is dead. Long live the Web on mobiles!

To no-one's great surprise, AT&T and Verizon won the Block C 700MHz spectrum auction; it will be interesting to see how the Open Access provisions play out. Google, however, has shifted target to the so-called white space frequencies in the analogue TV bands; we reckon it's all about keeping the option of a Google wireless access loop open as a bargaining chip.

That's if the technology even makes it possible; there's a brawl going on in Australia (should that be a donnybrook?) regarding WISP Buzz Broadband, WiMAX, and British WiMAX vendor Airspan Networks. Briefly, Buzz's CEO recently announced they were abandoning a WiMAX deployment because the technology didn't work. As the same guy had been one of the most aggressive proponents of it last year, this raised eyebrows. Now, Airspan claims it was Buzz's engineering at fault; they claim that rather than using their recommended equipment for wide-area deployments, they attempted to get away with cheaper micro-cells (presumably by cranking up the TX power). Anyway, the problem may really have been in the core network rather than the radio air interface; Airspan claims Buzz didn't provision enough backhaul capacity.

It's also worth noting that Buzz was using the earlier 802.16d-2004 (so-called "fixed WiMAX") standard in less-than-ideal 3.5GHz spectrum bands. Most WiMAX operators chose to hold the roll-out until 802.16e ("Mobile WiMAX", though it's not mobile-specific) was available, or else updated their equipment with 802.16e firmware.

Meanwhile, in the Telco USSR, Sprint-Nextel looks at building a single core infrastructure for all its networks; can anyone spell "platform"? After all, if the WiMAX network is meant to be an open development environment covered in APIs, this does sound interesting. But only as long as they resist the temptation to build an IMS monster, and the project doesn't get dragged down in the general chaos. This isn't encouraging though.

Vodafone , meanwhile, was knocked back from buying the rest of Vodacom by the South African government; now they're being rumoured as possible buyers of MTN Investcom. It has a kind of grandiose logic to it, but Vodafone acquisitions are still a little radioactive in the City... More concretely, Vodafone is apparently planning to float 20% of Vodafone Qatar this year, thus getting value out of one of its many, many emerging market buys.

France Telecom is reported to be looking at 800,000 FTTH hookups; meanwhile, the horrible figures at Deutsche Telekom's fixed-line operation just keep coming. Revenues fell 8% in 2007, and EBITDA 14%; this year they reckon with 4-6% and 5-8% drops respectively. No wonder the stock market treated this as a profits warning.

Perhaps they could try appealing to the fetish market? LG this week announced a phone designed to "feel as if you were texting on someone's skin". It's got HSDPA. And who knows, maybe even acne? Now, we finally realise why Google called their mobile development platform Android.