Robert X. Cringely:  I always seem to get good stories around holidays when fewer people are reading. That’s certainly been the case the past two weeks, when I wrote first about the Google shipping container data centers and then about the Google Box (Gbox? Gcube?) interface device. Readers have been throwing in their ideas and only Google remains mum. So let’s review the underlying strategy here and understand the real implications of these devices.

To most of the alternate ideas I am hearing about why Google is building those shipping containers and what those little cubes are for, I say "Yes!" Of course, it is for that and for this, too. The great danger here is in limiting the potential applications, because Google won’t. The whole idea is that this is a PLATFORM, not a product or a service, and a platform becomes what you make of it. So yes, it is for that, of course.
But there are some limitations. The shipping container data center isn’t, for example, the basis for an ISP to compete with cable or DSL or WiMax or whatever. There’s a huge impending battle coming for that last mile, and Google quite specifically doesn’t have a dog in that hunt, nor does is it looking for one, because as I have written before, the risk is too great and the reward too little.

"In our internal discussions we’re debating lately who will win the IPTV war, and triple-play in general (one has to know his future potential customers)," says a reader inside a network equipment manufacturer. "On one hand we see carriers and network operators trying to reduce customer churn and increase their ARPU (Average Return Per User) by providing VoIP, IPTV, HSI (High Speed Internet), and fixed-mobile convergence when possible (and thus paying us a lot of money on routers, switches, DSLAMs, and optical equipment). In this model, the vast majority of the network traffic is terminated at the local/regional carrier’s network and the HSI bandwidth remains for best effort traffic and doesn’t grow so much.

"On the other hand, we see Google, Yahoo, MSFT, etc. providing IPTV over that HSI the customer pays a minimal flat fee for… and which makes the carrier/operator no real profits (and just like you wrote, these Googles will have to build a true video distribution network to provide "toll-quality" TV). In this model, as HSI bandwidth increases (upon customer demand to see higher quality Google TV shows, and competition — "if you don’t double my bandwidth, I’m switching providers") the carrier is just selling larger Internet PIPES on which he makes little to no margins."

Wow, that’s a great perspective, and I am so grateful for his sharing it. These concepts like IPTV and triple play (voice, data, audio-video) are heroic and scary — so scary that the smart money is staying out of that game completely. We’re talking billions and billions of dollars in new infrastructure, at least half of which will be ultimately a waste of money. Yet if you are a cable television company or a phone company in America, this is what you do. You spend tens of billions of dollars on infrastructure to better deliver SpongeBob SquarePants.

Please understand that this discussion is U.S.-centric, but the same tussles are happening in many other countries so the same ideas apply.

This reader wonders who is going to win? I wonder instead who CAN win? The players are broader than most people expect — CATV, telco, WiMax, and Broadband over Powerline (BPL) are going to fight it out. By CATV, I mean Comcast in the U.S. The others are active, some of them, but only Comcast has the vision and the resources to win the day. Telco means Verizon, SBC (now AT&T), and BellSouth (which I expect to be eventually absorbed into either Verizon or SBC — probably Verizon, believe it or not, because of the Cingular partnership, which will then once again become AT&T Wireless). WiMax means Sprint-Nextel with the assistance of Intel. And BPL means Current Communications and its partner utilities. CATV and telcos are fixated on each other and that’s a mistake, because Sprint-Nextel has a complete end-to-end package (national network, backbone capacity, and faster deployment) while Current has the financial power of the electric utilities, which have the most to gain and nothing to lose, making them the most dangerous of all. My guess is there will be two winners, but I can’t yet call who they will be.

Why would Google or, for that matter, ANY company not defending a network infrastructure paid for 30to 60 years ago want to bet on the winner of this war? It makes no sense at all. So it won’t happen. And if you see a Yahoo or MSN as being in that business, they aren’t, except as paid suppliers of content. Even AOL — part of a huge CATV company, remember — is trying to extract itself from that dependency.

And the reason they are all trying to extract themselves is just as the reader described. With an increasing number of broadband alternatives, the advantage is turning more and more to users, which means greater profit pressure on ISPs.

What happens then is simple. The telcos and cable companies first use their political connections to try to legislate and regulate these lower-cost competitors out of their markets. This may work from time to time and place to place, but in the long run it is a stalling action, during which the big companies pray for a miracle.

The step after that is technical obstructionism — traffic shaping and other techniques to disadvantage the Yahoos and the Googles. That’s where those shipping container data centers come in. Parked at the peering point, sitting on the same SONET ring as the local telephone company, Google will have done as much as it possibly can to reduce any network disadvantage. By leveraging its own fiber backbone Google not only further avoids such interference, it has a chance to gain a step or two through better routing or more generous backbone provisioning. What’s stored IN the data centers is important, but how they are CONNECTED is equally important.

The other part of the strategy is the gBox or gCube or — how about this one, the gSpot? — Google’s interface device, which might be Google’s version of the "Home Gateway." Another example would be France Telecom’s Livebox (or the number two French ISP Free’s Freebox, which is even better), integrating video, Internet, and VoIP. And if you check out the latest Xbox or PS/2 releases, you’ll see everyone is heading that same way, from different starting points in the home. But the gSpot strategy is completely different. Where the company is deliberately deciding NOT to compete against the infrastructure builders on the street corner, they plan to overwhelm all players inside homes and businesses.

The Google Box can function as just such a home gateway, but with two twists — extremely low cost and massive deployment. You could have a Google box and get your home on the Internet, but if you had SIX Google boxes you could do so much more. Every TV could be independently on the Net; a couple computers could be running in different rooms; your telephone would be connected; maybe your house, itself, and even the house next door would be networked using the built-in mesh networking. That goes beyond any device I’ve yet seen described by, say, a Cisco.

Microsoft is effectively taping $126 to every xBox 360 as it leaves the factory. Say Google did something similar with the imbedded, unhackable, valueless for anything else gSpot, driving its cost down to $10 (it has to cost something or some people won’t plug it in) from a Cost of Goods of, say, $40. By investing $1 billion, Google could put 33 million gSpots into local service, networking at least five million homes for two percent of the amount per customer Verizon and SBC are throwing into Fiber-to-the-Home.

Who is going to win the triple play? It doesn’t matter. Who is going to win the game? Any player with deep pockets and no particular technological dependency. At this point that could be Yahoo or Microsoft or AOL or some new player altogether, but it probably means Google.

More here.