In-Home Networking Gets More Respect - and More Dollars

The investments in the next generation of in-home networking are starting to heat up. That's the best way to know it's real.

Belkin announced a new product today for gigabit-rate home networking using powerline technology (i.e., using the electrical outlets and cables already in your house) and based on chips from Gigle Semiconductor.

Powerline networking is an alternative to in-home WiFi networks, also offered by Belkin as well as by Cisco's Linksys division. Just last week, Cisco announced a big $1B program together with Duke Energy for an in-home Smart Grid deployment.

Besides powerline and WiFi, there are alternative networking formats like ZigBee and, in the near future, white space radio (WSR) technology that rides over the unused TV spectrum. These formats will eventually battle it out for dominance, although I think that WSR is perhaps the most promising due to its longer range. In any case, these billion-dollar bets by big consumer companies make it clear that the time for in-home networking to mature beyond just WiFi Internet routers is at hand.

I think it's important to distinguish between the Smart Grid inside the home (the Home Area Network), as opposed to the external portion of the Smart Grid, which is in the power generation and transmission network.

Initially, the demarcation point will be the power meter at the side of the home. That demarcation blur rapidly, however, since power companies have already begun to make deals with consumers that allow them to reach into the house to control appliances. But just at this moment, I'm refering to the Smart Grid inside the house.

Singapore Advances Toward 1 Gbps Residential Service

Singapore already has more residential broadband subscriptions than it has households (11.7% more subs than households, as of last April). That doesn't seem to be enough to keep them happy. Now they want to move those subscribers to fiber.

According to Light Reading report from CommunicAsia 2009, Singapore is targeting 95% FTTH coverage three years from now, and 100% a half year after that. They plan to offer a 100 Mbps and a 1 Gbps dowload-speed service tier (businesses get the same upload speeds as download, while residences get half of that).

Singapore isn't the only Asian country with aggressive broadband plans. Consumers in Korea, Japan, and Hong Kong alsoenjoy some of the highest bandwidth rates in the world, and that's going to position these Asian regions well in the knowledge economy and cloud-based computing world.


The rationale for some of this need for excellent connectivity is geographically based. Singapore is a small island nation located a long distance from many major trading partners; South Korea is for all intents and purposes in a similar situation, and the parallels extend to a lesser extent to Hong Kong and Japan. That means these countries need both the fat pipes and something to deal with the latencies that would otherwise slow things down. That's what RocketStream is for. :)

It's a mix of entertainment and business that really drives this need to be connected, and so these countries are investing heavily in their national IT infrastructures. Also, these regions are no longer low-labor-cost centers. So if they don't have low labor costs, they're moving to a new low-cost competitive advantage: low IT costs. We in the U.S. might learn something from this.

To be fair, though, we need to recognize that it's much easier and cheaper to achieve high broadband penetration rates in a tiny island city-state than in a large, distributed nation like the U.S. That's just a function of geography. It's just much more expensive to connect all the ranches in Montana than the apartment high-rises in downtown Singapore.

.

Be that as it may, this super-broadband connectivity going to put the average Singaporean at a huge advantage compared to a rural - or even and urban - American who just got ADSL at 3 Mbps. By then FiOS and DOCSIS customers in the U.S. will be the only ones with a prayer of keeping up.

.

That's why I think that the BTOP portion of ARRA (a.k.a. the Stimulus Plan) can only be the start of a sound national broadband policy. You're heard me say this before: a nation that isn't well connected is also not economically competitive.

Britain, for example, is also not bery ambitious. They are targeting 90% of the population to have 50 Mbps service, but not until 2017.

According to the BBC: " Antony Walker, chief executive of the Broadband Stakeholders' Group, believes the tax will mean that 90% of the UK will be able to benefit from broadband of up to 50Mbps by 2017."

That will make Britain an online backwater, as pointed out by The Guardian.

The Aussies, too, have launched their own FTTH program, with 100 Mbps planned for 90% of the population. This, too, will take about 7 or 8 years to come to fruition. (Australian geography has parallels to U.S. geography in terms of wide open spaces, but there's less population in those open spaces.)

And You Thought I was Kidding About Internet to the Toaster…

This week’s announcement about Intel buying Wind River was pretty big news. This is Intel making a software play in embedded systems and mobile devices, both for Wind River’s software and to push Intel’s mobile Atom processor.

Yeah, so embedded software isn’t always the sexiest thing around, but I was interested by Stacey Higginbotham’s observation on GigaOM that one of the driving forces behind big, beefy embedded software for consumer devices is the complexity caused by rich connectivity between such devices.

As all kinds of devices interconnect – mostly wirelessly – to each other and to the outside world, an abundance of machine-to-machine (M2M) networks will form. This collection of networks is sometimes called “The Internet of Things.” Intel hopes that the Atom processor will become the default chip to power this network.

The choice of processor is only one of the battles around M2M. Wireless data format is another. White space radio (WSR) will be going up against ZigBee, WiMAX, LTE, and WiFi. The advantage that WSR has is that is goes far, travels through walls, can handle high data rates, and doesn’t require licensed spectrum. None of the other formats can boast all four of those characteristics.

M2M communication is only one of many applications for WSR, but I think it’s a potentially huge one. From a technology perspective, it’s very elegant. From a business perspective, it’s an extremely attractive play for Voyant.

The Hunt for Tech Bargains Heats Up

Voyant isn't the only company that thinks this is a good time to pick up undervalued technology assets. Anupreeta Das just wrote an interesting little article noting some companies intested in M&A in the tech sector. Her examples of companies on the hunt for bargains include IBM, Oracle (buying Sun), NetApp (buying Data Domain), Broadcom (although their hostile bid to acquire Emulex is turning nasty), Sybase, Yahoo, Corning, Symantec, and a few more.

While I don't know anyone who thinks the economy is going to fly high any time soon, at least the OECD is confident enough to say that they think the global economy is "out of free fall."

This feels to me like a sort of "bottoming-out" sentiment. That sort of mood is now getting a somewhat broader acceptance than it was last February, when I posited that it was now time to think about tomorrow's technology.

Finding technology diamonds in the rough is at the core of what Voyant does, and the good new is that there are bargains out there now. Just ask the biz dev guys at IBM. Or Oracle. Or NetApp. Or ... (you get the idea).

Fortunately, there are still more bargains than buyers.

Verizon Sells Some GTE Landlines to Frontier

Verizon announced earlier this week that it was selling off some of the landline subscriber access lines that it acquired when Bell Atlantic bought GTE to become Verizon in 2000.

I might have thought that most of these subscribers that Verizon is ditching are rural customers that aren't terribly profitable for Verizon to support until I saw this story from Telephony Online, where Craig Moffett of Bernstein Research is quoted as saying that these lines "are materially more profitable than average for Verizon's wireless business."

Still, I wonder whether these lines will continue to remain highly profitable in the future, given rural operations costs. Also, this deal allows Verizon to consolidate in regions where it is a dominant service providers, and getting $8.6 B in return for selling them off to Frontier is a pretty nice coup.


If we assume that Verizon concluded that these customers are indeed unattractive in the long run, this would be an example of the same economic reality that has led to poor Internet coverage in rural America, and hence the $7.2B portion of ARRA (a.k.a., the Stimulus Plan) dedicated to extending broadband throughout the U.S.

One might object that I'm mixing two different services that are only partly linked: basic telephony and broadband. Verizon is ditching their telephony services, while ARRA is concerned with broadband, but the economics arguments are similar, and of course, broadband often rides over telephony infrastructure and vice versa.

It's interesting to note that, while white space radio is often discussed as a medium for rural broadband coverage (among many, many other applications), maybe we should be thinking about white space radio for basic telephony service, too.

6000 Airline Passengers Should Not Have to Share a Tiny 3 Mb/s Pipe

Om Malik wrote a short piece this morning about in-flight connectivity on GigaOM. He focuses entirely on the U.S. market, without much discussion of any overseas markets. I can understand that, given the list of announcements made to date, but there's a whole set of overseas markets that shouldn't be ignored.

Several U.S. carriers are trialing or building out in-flight WiFi Internet access in the U.S. Overseas, there's only been one announcement with a tiny airline about in-flight Wi-Fi so far. There have, of course, been a whole raft of overseas announcements about narrowband cellular services, but the overseas in-flight broadband markets are not yet really served.

In the GigaOM blog responses, several people have commented on data rate available to the passenger. However, most of them have been focusing on the fact that all of the passengers on the plane share the same pipe. That's true, but the problem with existing solutions is even worse than that.

The real problem is that, with existing solutions being deployed today, all of the planes in one cell or region have to share the same data pipe. Imagine 10 people on each plane, all watching Hulu or content from their home Slingbox.

Now imagine 50 airplanes in the same cell - quite realistic for 1000 planes over the U.S. spread (unevenly) over 300 ground stations. That's 500 video streams. In the case of Aircell, these 500 streams are all sharing a roughly 3 Mb/s pipe. In the case of Row44, they share 4 Mb/s. There's no way that's going to provide a satisfactory user experience. And that's why I submit that these are fine narrowband technologies, but once you get more than a few trial flights in the air simultaneously, this won't scale. That's why true broadband service will be needed.

Voyant Aviation Broadband: Up to 35 Mb/s to each airplane.

Now, I'm not arguing that you'll never statistically multiplex users together onto shared pipes. Of course not every passenger is streaming at full speed at the same instant. But a conservative estimate of 50 narrowbody planes in a cell or region, each with 120 passengers, means 6000 people sharing that bandwidth. Even assuming a modest take rate, you can see that a couple Mb/s for the whole gang just isn't enough.

Just think of a Starbucks hotspot today. You've got typically about 20 people there, and they get a about 3 Mb/s allocated to them. If that's the level of service a consumer expects, then why should we think that going from 20 users to 6000 users without a corresponding increase in bandwidth will satisfy anybody today, much less over the system's 10-year design lifetime?

The reason that these services seem to provide something like broadband is that, in trial flights, there are only a handful of WiFi-equiped aircraft in the air at any one time. Just wait till there's a full fleet up there. It'll feel worse than dialup. That is, until true broadband services are launched.

GE's Latest Optical Storage Advance

As longtime readers of this blog may remember from last year, I'm very interested in the relative changes in the cost of bandwidth vs. the cost of storage. Those relative costs affect so much of how communication networks will be architected. The financial consequences are consequently huge.

Lower storage costs mean that architectures like CDNs will prove in more often. Conversely, lower bandwidth costs help make the case for central data storage and related architectures, notably cloud computing (where it's important to access the cloud often and cheaply). Of course, low bandwidth costs also help pave the way for applications that lend themselves well to the transport technologies behind RocketStream and RocketConnect, so Voyant has a vested interest here.

So it was with interest that I noted GE's announcement about advances in their holographic storage research. To put things in perspective, this announcement is a significant step in the evolution ofholographic storage techniques. Such techniques have been used for many years in military applications, but now they're one step closer to commercial application.

What does this mean?:

• From a technical perspective, the significance here is the establishment of high-reflectivity holographic images.
• From an implementation perspective, the key is that they can write 500 GB of storage onto one CD.
• From a business perspective, the NY Times summarized nicely with Brian Lawrence's* quote: “The price of storage per gigabyte is going to drop precipitously.”

This is still far from commercialization, but it is a signficant step and, more importantly, part of a trend. In the horse race between low bandwidth and low storage costs, we all win.

* Just a little disclaimer here, or perhaps a personal note of interest. GE's lead for this program is Brian Lawrence, who I've known for many years - since our grad student and post doc days.