DISH for Sprint, Google Fiber economics, Rackspace licensed cloud, PC sales sink: Telco 2.0 News Review
Telco 2.0 News Review
- Carriers: DISH jumps in on Sprint-Softbank
- Broadband: Open access to civil works makes Google Fibre cheap
- Voice 2.0: WebRTC's secret weapon: not voice
- Smartphone Roundup: Samsung to ship 80 million GS IVs, 320m smartphones
- Cloud Computing: Amazon strategy update, Rackspace licensed cloud for telcos
- Facebook: Facebook Home reactions, "Partner Categories" on the pad
- Security: Security externalities - the heart of the menace
- Strategy & Finance: PC sales go "thud" in Q1
- Technology Disruptions: HP, IBM, Bechtolsheim, Intel - microservers everywhere, shades of the Transputer
[Ed: Don't forget to book now for the Telco 2.0 EMEA Executive Brainstorm, in London on the 5th-6th June]
In the US, we've had the Sprint-Softbank acquisition. We've had the T-Mobile/MetroPCS deal and T-Mobile's price disruption. Now for the DISH Network bid for Sprint. Yes, really. DISH is offering $25.5bn, of which $17.3bn is cash and the rest, shares. If you're wondering about the cash, $8bn of it comes from smashing the DISH piggybank and the rest is borrowed from Barclays Bank.
The big idea is selling satellite TV along with Sprint mobile, and using the huge Clearwire spectrum block to get fixed-wireless out in the outback - rather like DISH's "cantenna" partnership with Verizon Wireless, which sounds cheaper.
The GSMA has published a report on universal service, arguing that emerging markets' universal service funds tend to be ineffective and are very often sitting on large amounts of money paid in by operators but not used. India, for example, may have as much as $4.1bn.
This isn't going to make anyone any friends, of course; the hugely bubbly 3G auction brought in so much money that the Indian political scene was immediately convinced that mobile operators had ripped them off over 2G, while the operators have been struggling to pay back the licence fees and the cost of their investments in a low-income market crowded with players.
Fitch points out that Indian operators are investing much less in their networks than comparable players elsewhere in Asia. Chinese MNOs typically invest about 30% of their service revenues, those in the Philippines about 20%, while Indian operators invest about 16%. It's much worse if you look at investment or spectrum per subscriber. And the figure is rising in China, while it's falling in India.
Fitch argues that the problem is that the Indian MNOs borrowed too much money to pay for the licences, and now the repayments are eating the economic surplus. With data revenues beginning to grow, this is going to be a serious problem.
The US Department of Justice wants to make sure smaller operators have a fair crack at the wall-punching low frequency bands, after all, and Benoit Felten is trying to get Susan Crawford picked as FCC chairman. Here's her pitch:
Meanwhile, EverythingEverywhere's secret product launch turned out to be bigger LTE channels.
In the fixed world, Google's announcement of fibre in Austin, Texas called forth a press release from AT&T promising "an advanced fibre-optic infrastructure capable of delivering speeds up to 1 gigabit". Teresa, The Voice of Broadband points out that Google Fiber happened in Kansas City because the city let them use the city-owned power poles and other civil works, for zero dollars, not a penny more, not a penny less. AT&T wants the same terms Google gets from Austin.
Business Insider quotes Bernstein's estimated costs for Google Fibre. The main take-home messages: adding TV to broadband ups the cost-to-connect, but dramatically ups profitability and take-rates. Also, access to the civil works makes deployment much cheaper. Ars Technica picks up on the same work but highlights the possibilities to scale up, estimating that Google Fiber could reach 15% of US households at a cost of $11bn over five years.
None of this will be new to Telco 2.0 readers.
In Australia, meanwhile, more political storms swirl around the industry. NBN Co is steadily digging away, blowing open-access fibre through Telstra's ducts, while the political opposition want to shut it down and have fabulous FTTC, keeping the Telstra copper, just like we do in Britain. Apparently, the British market is considered a great success in Australia, proving once and for all that it really is upside-down down under.
Hilariously, though, the opposition spokesman on the issue turns out to have bought into France's national FTTP plan in the most literal way possible, buying a boatload of FTel stock. Not, say, BT stock, or any other operator that's going with the FTTC-and-no-unbundling strategy (aka "facilities based competition") that he recommends for Australia. FTel, the one that's building a lot of wholesale FTTP.
Further, his estimates for the cost don't seem to account for the copper OPEX, which is usually an important economic driver.
Meanwhile, a good point: "on-demand FTTP" is like an "on-demand street", especially in a GPON system like NBN.
In the voice world this week, Chris Kranky points out that "the secret weapon of WebRTC is the data channel" - that is, the ability to transfer structured data between the parties to a conversation, before, during, and after it - and it's a great opportunity to make IVRs and call centres much less hellish.
Twilio's blog has a great HOWTO on making two-factor authentication for a basic website, and another on working with node.js and their browser client.
Here's an interview with Voxeo Labs and Alan Quayle.
Vodafone sells a very traditional unified comms deployment to British local governments.
Pain over WebRTC codecs.
JPMorgan reckons Samsung will move 320 million smartphones this year, and 80 million of those (or precisely 25%) will be Galaxy S IV supershinies.
They've also got a Note III coming at some point, perhaps at IFA. It's suggested here that they prototyped the GS IV with a metal body but didn't launch it because it was difficult to manufacture at the time, but now the tooling is in place and they're ready.
Meanwhile, Huawei is worried that it's being cannibalised by cheap Chinese Android ODMs, and is therefore in a hurry to move up the value chain. The question is whether they can get more like Samsung (or Apple) before they end up like HTC.
Interestingly, a survey of north American mobile users suggests that there is surprisingly little platform loyalty or lock-in. However, is the industry just competing to offer "moar", rather than anything genuinely interesting?
Blackberry 10 is meant to be that. This week, a US stockbroking firm suggested that returns were very high. BlackBerry disputes this and the lawyers are out.
Jeff Bezos gave some insight into Amazon strategy this week, in a letter to shareholders. He claims that the company has a policy of always releasing new features and cutting prices first, in order to avoid building up technical debt and to keep disrupting the competition.
The company also handed out a wedge of stock options to its executives. More importantly, this move came as it added Apple to the peer-group it uses as a performance benchmark. Amazon now considers Apple a key competitor.
The AWS Blog marks a year of CloudSearch.
Rackspace is selling what is, in essence, a licensed version of its OpenStack cloud. To some extent, this competes with Rackspace's own main-line cloud product, but the point is to scale up faster in order to compete with Amazon Web Services.
Interestingly, the product is targeted at service providers, especially regional majors where Rackspace is unlikely to build their own infrastructure and the AWS Death Star hasn't been finished yet. If this reminds of you of something, you probably want to refer to our Cloud 2.0 Strategy Report, and if it doesn't, why not read it for the first time?
So you're a telco and you're worried that you've bought too many VMWare licences, but you're not convinced you can build your own cloud. This may be the product for you.
Meanwhile, here's a photo tour of Microsoft's Dublin data centre, concentrating on the fresh-air cooling system. You know you want to. And Google is spending $390m on an expansion of its Belgian data centre, the home of GMail in Europe.
And Salesforce is offering its own mobile app framework to wrap around the APIs to its enormous cloud of CRM. There's something ironic about a company whose logo is a NO SOFTWARE sign advertising...software.
Facebook Home rolled out this week, and the first reactions are in, from the reviews on Google Play. They are not necessarily encouraging:
You have to be a complete facebook junkie to use this for more than 5 minutes," reads another. "It takes away the ease of your phone actually being a phone and even adds complication to using facebook. Uninstalled
The theme is that Facebook obsessives tend to 5-star it and everyone else hates it, which is rather what we argue in our Facebook Home note, available for download to Telco 2.0 subscribers. For a long time, Facebook had one website to rule them all, that was meant to be all things to all their users. Facebook Home seems to represent a move to focusing on the company's fans, and risks disappointing the others. The problem is whether there are enough "true fans" out there - Facebook was certainly a business driven by them in 2009, but you might question that now.
As a two-sided business, Facebook needs to delight one customer, the advertisers, without alienating the other, the users. Home is a user-facing play, and here's the advertiser-facing play: Partner Categories. (The Facebook announcement and screenshots are here.)
This application pulls data from a variety of other retail data miners into the Facebook Studio ad-planning tool, so you can (for example) target the friends of people who buy a lot of washing powder in given geographies but who don't fly United Airlines much. Evidently, this could be the very definition of "creepy", and advertisers would be well advised to use it with care.
Perhaps they could stop serving ads into pages with names like "drop kicking sluts in the teeth"?
Meanwhile, Eric Schmidt is ironically concerned about the privacy threat from privately-owned surveillance drones, although it seems a remote concern compared to, say, Google's cars taking photos of everyone's house and logging your WiFi traffic. Perhaps someone's been flying one over his pool in Palo Alto.
After all, Google has just offered to settle with the European Commission over search issues, accepting to mark more content as Google-owned, show more competing results, and to sell ads to competitors in the all-paid sections of their sites.
Google also did a clean-up of the Google Play app store this week, killing off 60,000 apps for spamming, data-grabbing, spreading viruses, etc. The problem is whether this sort of periodic spring-cleaning is enough, or whether an Apple-like model with more stringent approval might be better.
Freedom to Tinker discusses the other week's major DDOS attack on Spamhaus, and points out that the worst security threats are the ones that make use of economic externalities. For example, BCP 38 implementation benefits everyone, but individual networks have relatively little incentive to fix their egress filters. Bruce Schneier's comments thread is as good as usual.
Microsoft pushes out a security fix and then reverses it because it's not secure.
Mobile phones, unlike computers, have a nationality - the country code in their phone number. Who is it in Hong Kong that's targeting South Korean Androids?
Here's a fascinating piece on how the FBI investigates mobile devices. There's an IMSI catcher, but there's also so much more.
The headlines were exciting: "Plane hijacked with Android phone!" The story was a bit less sexy.
Arrgh. IDC's PC sales estimates for Q1 are out, and they show a sickening thud, with shipments off 14% on an annual basis, or the worst ever since IDC started counting in 1991. Gartner's slightly different metric suggests they fell 11%. Either way, it's a nasty surprise for the industry.
IDC reckons that the launch of Windows 8 hurt, rather than helped, PC sales. However, this is possibly an arteface of Win8 being more touchscreen-oriented and shipping on a lot of hybrid tablet/notebook machines, which aren't necessarily counted as PCs. IDC, for example, doesn't consider anything with a detachable keyboard to be a PC. (It wasn't so long ago that pretty much everything that was considered a PC had a detachable keyboard, although you couldn't do much without it.)
The shrinkage breaks down very unevenly across the major manufacturers. HP's shipments fell 24% in Q1, while Lenovo held steady. They sell heavily into Asia and into emerging markets more broadly, and they also sell into enterprises well - all three are sectors that are still growing.
ZDNet's Simon Bisson argues that hardware innovation has slowed down, while Microsoft has put much more effort into improving performance in software, and a software upgrade is cheaper than buying a new computer. Further, we'd add, Apple holds a lead in the hardware beyond the CPU, like displays and graphics. So if you want a better computer, you'll either change the software (whether to Win8 or Linux), or you'll get a Mac. Apple is still gaining share in the US.
Meanwhile, another set of forecasts suggests that the netbook may actually go extinct by 2015.
The company that must be really worried by all this is of course Intel. It should come as no surprise that they have decided to make a touchscreen part of the basic Ultrabook specification.
Here's something interesting. HP may be suffering in the PC market, but for all that, they're putting their formidable engineering resources into the cloud and the servers and networking equipment that underlies it. Project Moonshot is a new line of servers that are small, cartridge-like devices that plug into a box of 10 to 40 devices. Each one's exact configuration is entirely customer-specified.
The big draw is less power draw, with the devices typically consisting of a chunk of Flash storage and either an Intel Atom or ARM/Qualcomm CPU. The other major draw is their ability to add more flexibility to cloud deployments; with 40 machines per rack, it's easier to offer different architectures and hardware options to a cloud.
Meanwhile, IBM is spending a billion on Flash storage R&D. The other week, we read about Andy von Bechtolsheim, and an all-star cast of Sun veterans, , and their chip startup that wants to "move the computer to the data" by integrating a really tiny Linux VM on each chunk of RAM or Flash storage. And of course there's Intel's Next Unit of Computing. As this was also the week Margaret Thatcher died, we couldn't help thinking of the Inmos Transputer. Everything old is new again.
On the network side, a whole new SDN standards effort has emerged, called "OpenDaylight", which looks like it's an alternative to OpenFlow and perhaps also to the emerging carrier world NFV project. It includes most of the router and switch vendors you've ever heard of, but neither Google nor Huawei.
And the Apple/Google patent feud has finally exhausted the court's patience.
"The parties have no interest in efficiently and expeditiously resolving this dispute; they instead are using this and similar litigation worldwide as a business strategy that appears to have no end," U.S. District Judge Robert Scola in Miami said in an order dated yesterday. "That is not a proper use of this court."...
The case in Florida involves more than 180 claims related to 12 patents and disputes over the meaning of more than 100 terms, Scola said in his order. Apple and Google were unable to streamline the case, he said, calling the companies' actions "obstreperous and cantankerous conduct."
"Without a hint of irony, the parties now ask the court to mop up a mess they made by holding a hearing to reduce the size and complexity of the case," he wrote. "The court declines this invitation."
And here's a ticker counting how often bar codes get scanned, as we count down to the 40th anniversary of the original augmented reality application.