Telepocalypse

OPINION://Power to the people

Via digg I came across this design for a better wall plug. (I won't steal the image - go follow the link.) It angles the US socket upwards to make it easier to plug in. We're about to have a wee lesson in network effects and economics.

Now, I won't go into the merits of this particular design other than to note some of the comments:

Yes, but can you childproof them?

... Laura, since I now have a very inquisitive 6 month old kid, this is indeed a good question. I can't imagine why you couldn't childproof them -- those little plastic blocks fit pretty much any three-pronged socket, no?

So we've obviously got a secondary use case that requires compatibility between sockets and "kiddie-proof" socket covers. (My elder kid saw their removal as an irresistable challenge that attracted her siren-like to otherwise utterly uninteresting power sockets.)

In the UK, we don't have this problem, because we settled upon a "mk 2" standard decades ago. (The "mk 1" is still in use in India.) Here's one I made earlier:

The earth pin is the slightly longer one at the top. There are several improvements over the US design:

• The polarity is obvious from the triangular layout with earth at the top, no subtle flanging.
• The earth pin, being longer, pushes away a shutter system that protects the live and neutral sockets from small people with drooly fingers and paper clips.
• That means every appliance has an earth pin. Some "wall wart" converters just use a plastic dummy.
• The bases of the live and neutral pins are plastic coated, so if you pull the plug out a small bit and slide something in to touch the pins, you still don't get fried.
• There's a fuse incorporated into the plug. (We have central RCD circuit breakers too.) It's really, really hard to electrocute yourself here.
• The cord goes downwards, which stops you tugging on the cord to pull the plug out (deliberately or accidentally) and over-straining the wires causing internal derangement and a short circuit.
• A strain relief is there in case of cord trips; and most cords plug into appliances using an IEC connector which easily rips out at the other end.

Now, could the US ever co-ordinate on a better design? To some degree, yes -- but it would take decades. You could mandate all new appliances have earth pins, and all new sockets have internal shutters -- but the current design makes that a challenge because the pin lengths are equal. Everything is path-dependent; there may be no route from "here" to "there".

Now we come to see some parallels and lessons in telecom economics.

Firstly, we have two completely different situations between fixed and mobile networks. Ethernet is distributed via ubiquitous RJ45 connectors and and WiFi (a.k.a. "wireless Ethernet"). We've settled on two co-ordination standards that are in the public domain. A municipal network that chooses a particular underlying distribution system (e.g. PON, WiBro, VDSL, whatever) doesn't bleed that decision into the final interface for the users of the network.

On mobile networks, we're still in a state of flux, with a whole bunch of legacy (e.g. GSM, EDGE, early CDMA), current (EV-DO, HSDPA, UMTS) and emerging (FLASH-OFDM, xG) air interface standards. We want competition between these, as there's uncertainty around the technical and economic merits of each. Duplicative infrastructure has benefit from making these ecosystems compete. Furthermore, the nature of mobilty means that the standard needs to be contiguously deployed. You can't have town A choosing a different solution to town B next door.

Attempts to bridge these two worlds by having "ubiquitous mobile WiFi" are likely to end in tears, since it just isn't designed for that job.

Which really takes me to my public confession.

"Mum, Dad -- I've got something important to tell you. I've been keeping it to myself for years, but I just can't go on living like this."

"Go on son, tell us."

"Errr, um, I'm... I'm... I'm not a believer in for-profit deployment of metropolitan fixed telecom networks."

"Don't worry son, we'll still love you, no matter how strange your peccadillos."

OK, so there it is. I'm slowly coming to the opinion that fixed networks are common public infrastructure like roads. It's time to stop pretending that it's a for-profit industry in developed countries, because all that does is drives players to seek political favours to make their networks viable.

There's still some important unanswered questions.

• How far up the stack to transition from "municipal" to "for profit"? Brough Turner says "layer 0" at the rights of way. Bob Frankston says "layer 2", at the IP interface. Me? I think it depends on local conditions, with a pragmatic cost/benefit analysis. In older suburban USA with poles in the air you might get a different answer to central Edinburgh with everything buried under centuries of accumulated crap.

• Wholesale or retail? Or "free" like the roads? I simply don't know, but I don't really think I want another bill from my council, who will be woefully poor at billing and customer service.

• How to avoid congestion, much like the roads? How to price so as to create appropriate incentives to build capacity and ensure the right people are charged for it? We're facing a tough economics challenge here. Time to go re-read up all those books on public goods theory... but...

• "Municipal" or "collective"? Remember those lighthouses that were taught as canonical examples of public goods? It's bunk, many were privately built by shippers and insurers. It's far too easy to immediately jump to the conclusion that this is a problem that government should solve.

• Beware of politicians bearing the power to tax. It could be more about creating new rules on depreciation, taxation, and liability to enable semi-private collective middlemen to emerge. Do we need a new corporate structure like "Ltd", "Inc", "LLC", "PLC" for these new entities? Shame ".net" is already taken... I wonder if it ends up like the charity sector, where successful managers get to take home a good wage, which incentivises continued success, but don't have to reward shareholders -- just please an oversight board.

• Is trans- and inter-continental traffic really part of this model? I suspect not because the rights of way are free and geographically unconstricted. Governments make fewer mistakes connecting folk up inside existing population centres because it's obvious where the people and demand is. In the n^2 problem of long-haul traffic we may want competing ideas, and just buy in the right capacity from completely private companies.

• How to ensure that telecom shareholders are not subject to asset forfeiture? There's a terrible precedent in the UK with the establishment of the National Health Service. The voluntary hospital movement basically had its assets confiscated and turned over to the governement. We're still struggling 60 years later with the "wrong" municipal model that doesn't harness voluntary exchange and human compassion. Will "voluntary" network collectives have the fical and political muscle to avoid capture and continue to serve the public?

The "compatibility interface" described earlier also underlines the severe dangers of vertical integration of services and connectivity. It isn't nearly so hard to gather monopoly or dominant positions in the service and content space (think: ESPN, Google, Microsoft). Allowing the same entities to own the core distribution system is just plain too dangerous. Although I'm on record as being anti network neutrality rules, by far the strongest card in the "pro" deck is freedom of speech. I'd rather play that joker in a different way. I want to see government as an enabler, not a primary actor or a deliverer of favours to selected players.

It's also probably good for shareholders of telcos to ultimately break up the assets. Are BT's shareholders being given a good deal with the copper network being bolted on to their 21CN IMS application architecture? Better to unbundle the risks and let the market send its signals of scepticism or enthusiasm for the component parts. If we're to have telecom conglomerates, I'd prefer to see them "horizonal" than "vertical". Deutsche Telekom as "Europipe GMBH" anyone? What incentives do we need to create to make DT, BT, FT and so on voluntarily want to accept the buy-out deal?

Back to the power sockets. I have another hunch, but I don't have any data to back it up. Remember the path-dependence issue? How did the technological might of the USA end up with the crappiest of all the plug designs?

I suspect that there's a timing issue at play. The USA was one of the first countries to electrify. Some early municipal/collective efforts are likewise coming from the USA with fairly verticalised models. You can be too early to take advantage of first-entrant failures and lessons.

The other danger is being too late. Once a particular model spreads, an ecosystem is created around it. Shifting to another track becomes impossible.

So my hunch is that we're in a critical transition period. We're on the rebound from the failed 1990s experiment in pure private networks. They clearly failed to capure the value of the bits: the public came out tops, Google et al second, and the telco shareholders took a cold bath.

There's a window of maybe 5-10 years where each city and country gets to decide how much of the infrastructure is "common". Regulations around spectrum allocation that mandated single, monolithic networks with 95%+ population coverage may get re-written to enable more liquidity in the connectivity market. City blocks and upwards may get an opportunity to buy-back the copper and fibre distribution into collective ownership at a price fair to the telco and cableco shareholders. Franchising rules and endorsed private monopolies get swept away. There's a ton of progressive rule-making in prospect.

But within a decade or so midnight will come, the carriage becomes a pumpkin again, the mice scatter. You'll be able to dream of Price Charming, but take the wrong path and timing and Cinders is forever stuck with lousy power sockets and feeble, overpriced connectivity.

UPDATE: The Economist has a good parallel story on the development of the US Interstate highway system. Some of the take-aways are: national security and emergency use are important (remember Katrina's telecom disaster but evacuation success?), and watch out for the taxman and politicians.

Posted by Martin Geddes at 12:03 PM

Trackback Pings

TrackBack URL for this entry:
http://www.telepocalypse.net/cgi-sys/cgiwrap/mgeddes/MT/mt-tb.cgi/753

Listed below are links to weblogs that reference Power to the people:

» Public and private ownership in from Communications
In a long and interesting essay on standard interfaces, Martin Geddes makes the following comment about broadband, or more generally, fixed telecom networks:There’s still some important unanswered questions. How far up the stack to transition from â€... [Read more]

Tracked on July 5, 2006 2:40 AM