The idea of an energy crunch seems to be getting some more exposure. This has an interesting long-range repercussion on telecommunications. The obvious stuff is that you have a lot more demand for telecom since moving bits is a substitute activity for moving atoms. The less obvious impact is the type of telecom you get.
Smart networks require energy to run. In the absence of commercialized reversible computation technology, every time you turn a 1 into a 0 inside a silicon chip you need to discharge some energy as heat. The Internet has a pretty hefty energy requirement. The energy bill will start to matter more and more, since it scales linearly with the number of bits transmitted through an electrical or optoelectrical system.
This points us towards the Gilderean future of all-optical networks. Optical switching is the end destination in terms of dumb pipes. (Unless you have some insight into superstring theory and data transmission via parallel universes we haven’t seen yet — but the evidence points away from this at the moment.) The energy needs are much lower — one photon per bit.
For wireless, it also means mesh networks offer more than convenience and low cost. Whilst directional antennas help, ultimately you’re radiating energy over a roughly R^2 area with any wireless transmitter. A sequence of mesh nodes radiate your bits over a smaller area. (Think of a sequence of overlapping lily pads in a bigger pond. The pond is a traditional cell tower. The lily pads are mesh nodes. A path from A to B covers only a few lily pads with energy, not the whole pond.) The question is, though, does the increased energy cost of routing the meshed bits nullify the lowered energy cost of transmitting them?
If energy gets really expensive and the number of bits to be sent increases dramatically, we’re all going to get a lot more familiar with the information theory work of Claude Shannon over the next few years.
Interesting aside: quantum computing also appears to be fundamentally limited by the energy budget needs.
I promise to write some original and witty content of my own real soon, honest. But in the meantime, here’s a great quote from Russell Beattie:
I think Nokia can [dominate the handset market again]. I think their R&D, the investment in their own chipsets, focus on the enterprise and multimedia and general concentration on the long term is going to pull them through. Motorola and Samsung are kissing carrier asses and producing a million different handsets to see what sticks, but that’s not what’s going to win in the long run. What’s going to win is developing an ecosystem around your brand and platform, and that’s what Nokia (to me) is doing. The threat to this, of course, is that the carriers don’t like it because it devalues their brand. But these guys are going to be nothing more than “cellular-based ISPs” in a few years anyway.
No, it’s the platform, not the pipe, in the long run that’s going to rule the day. That’s why I’d invest in Nokia now.
All together now! [Gregorian style.] A-m-e-n.
UPDATE: The Economist takes the opposite view (subscription required):
The company with the most to lose [to original design manufacturers like Sendo and BenQ] is Nokia, which has become so powerful that operators and rival handset-makers are keen to take it down a peg or two. “For Nokia to stay on top of the game it will have to adjust,” says Brian Modoff, an analyst at Deutsche Bank. Unlike Microsoft and Intel in the PC business, Nokia is not protected by ownership of proprietary standards. To maintain margins and stay ahead of the industry’s ever-faster product cycle, says Mr Modoff, it will have to stop doing everything itself. “They will be more of a brand, a design shop, rather than building everything,” he says.
Nokia is doing its best to diversify, notably into mobile gaming with its N-Gage handset. At the same time, as handset technology is progressively commoditised, a strong brand will be increasingly important, and nobody has a stronger brand than Nokia. Its recent troubles may turn out to be a blip. But given the seismic shifts now under way in the industry, observes Mr Dean, with Nokia’s market share so large “there’s only really one way to go.”
At the risk of exposing my ignorance of spectrum regulation, I’d like to propose a way of dealing with the hogging of UHF frequencies by the broadcast TV industry. Two simple regulatory changes with low political impact could make a difference.
TV is undergoing a fundamental shift in economics. As noted today by C|Net News.com, the mass market broadcast 30 second advert is an animal in serious decline. The replacement is Google-style personalized adverts. The Internet is a two-way medium, which makes this possible. All very Cluetrainy.
So change #1 is you insist the current spectrum can only be used for one-way TV broadcast. There is method in the madness of preventing progress. It means nobody has to change. Just keep doing what you’re doing now. No squeals of protest. Of course, you’re becoming increasingly irrelevant to your true market (advertisers).
Second, insist that you only keep the spectrum as long as you actually use it to broadcast TV at the full power allowed by your license. Sounds very reasonable. But a TV station pumps out hundreds of kilowatts of output 24 hours a day. You need a big stack of coins next to the meter to keep it going. And, as noted earlier, energy is likely to get a lot pricier over the next few years. So this eliminates the danger of squatting on unused frequencies. It simply costs too much.
The alternative to the above is to try to remove sitting spectrum tenants through force. So you can either spend a decade of campaigning and lawsuits, or a decade of waiting for the invisible hand of the market to do its stuff. I’d chose the latter.
I was reading this old article at Tech Central Station, and noodling about what it was saying. In a nutshell, it is arguing that the 1996 Telecom Act is fine as it stands, just that it was badly implemented because incumbent telcos are allowed to charge excessive wholesale rates.
I completely disagree. Most regulation seems to be about trying to directly fix outcomes rather than let the market take its course. A monopoly supplier of telecom that charges high prices should be sending a price signal that encourages new entrants into the connectivity business. Telecom is not a natural monopoly. Cable, municipal fiber, MMDS, satellite, meshes — they are all viable alternatives, and more are on the way.
What the regulators need to do is prevent discriminatory pricing against newcomers. I would therefore propose that most of the existing regulations be scrapped and replaced with a much simpler framework. The goal would be to prevent predatory behaviour. The ability of incumbents to price differently according to geography would be restricted - a price drop when a new entrant arrives would have to be given to all. Additionally, I would consider restricting the ability of incumbents to drop prices following the arrival of a new entrant, for maybe a period of two years. That would encourage price moderation at the outset, and make a stable business environment for new entrants to form a business case against.
One thing I missed on my post on relying on Vonage is our other single point of failure: Kansas City Power and Light. We’re about to enter storm season here in Tornado Alley, which is frequently spectacular if somewhat unnerving. We don’t get many outages, but they do happen. Last Christmas, the moment I headed back to Europe we had one. My home server got stuck on rebooting at a screen asking if I wanted to repair filesystems that might have been corrupted due to incorrect shutdown. So I couldn’t access a ton of my email and files.
A traditional criticism of VoIP solutions is that the analog PSTN is powered at the edges, and smart VoIP endpoints depend on being powered from the unreliable mains power. (Don’t lick your fingers and stick them in those phone sockets kids! Someone might ring you with a ring tone you won’t want to feel again…)
But I’ve had several experiences that have shown me that this isn’t a very strong argument. I’ve been in Scottish winter storms where both power and telephone service has been knocked out to rural areas. And the 2002 ice storm in Kansas City had telephone and power lines down over half the city. (The blue and purple arcing of failing power lines was quite a scene from our balcony.) Clearly there is a significant correlation between failure of the telephony network and the power network. It might be interesting to have some facts on the relationship. Anyone out there know what the correlation coefficient might be?
Via Corante comes news that Comcast cable are creating their own set top box. This is interesting because it continues an ongoing trend. Imagine you’re the network operator or some other middleman in danger of disintermediation. You don’t care about being cut out of the picture if you also control the end points of the network. Think subsidized Analog Telephone Adapters locked into Vonage service. iPods locked into iTunes. Cellphones locked to their network operator. Even PCs locked into trusted computing architectures.
My wife’s parents are over in the US from Europe. They finished their two-week stay with us this morning. We’ve equipped them with a car, a credit card and a cell phone, and pointed them westwards. They’re driving down towards Dodge City today and on to a rendezvous in Vegas in two weeks’ time. (Maybe the credit card wasn’t such a good idea after all…)
Anyway, by the principle of conservation of cell phones, my wife is now cellphoneless. Decellularized. Unmobilated. (In case any of you think that’s lacking in chivalry not relinquishing my own phone, let me tell you I’m now carless. In sprawling suburban Overland Park, that is like being blind and limbless in most of the rest of the world. Serious personal impairment.) So no cell phone means she’s sat at home, baby on lap, with just a Vonage phone and a Yahoo instant messenger client to get hold of me. Both of which have a single point of failure, our Roadrunner cable modem. And her parents are 100% relying on Vonage to get hold of her in an emergency.
Now, I’m damned pleased with Roadrunner. Reliable, fast and good value. But their DNS server can be pretty wobbly at times — they’ve been struggling to return the address of one of their own web sites recently. (I have a home web server that places my current IP address on my ISP’s FTP server once an hour, in case the dynamic DNS service update fails when I’m away. I can then SSH back home to fix any problems. When it fails, I get an email.) The service has been known to go down from time to time, too. We’ve had a few difficulties with Vonage too, but nothing that a quick hard reset couldn’t cure. And before now all it did was delay an intercontinental social call back home.
It’s only when you really rely on VoIP and VoIP alone do you feel that feeling of being a bit exposed. You know, like when you’ve decided to pass another car and get alongside and wish you’d put your foot down a few seconds earlier as the next corner looms. I’m at risk, and either this damned machine works perfectly, or I’m in a lot of doo doo. Not as bad as when I did a bungy jump, but a definite moderate nervous edge.
There’s a nice comment by Steve Saenz on this old article of mine on Vonage. He documents his difficulties in getting it to work reliably. He can’t be alone.
I just wonder if we’re set up for a backlash. Some day, someone will have a serious mishap because a VoIP call failed when a POTS one clearly wouldn’t. They might even die. Even if as a society we’d be safer on average with an IP communications solution, this will make the TV news and the press. And they’ll make a BIG thing of it, because nobody tunes in to hear that there were no sensational events today.
How long will it be before the speeding truck comes round the corner? I don’t know, but it will give the incumbents potent political ammunition when it happens.
The IP-MoU initiative seeks to propose a set of “freedoms” for IP communications, thus:
Ensure that consumers worldwide are assured basic rights as users of IP communications.
* Freedom to Access Content: Consumers should have access to their choice of legal content;
* Freedom to Use Applications: Consumers should be able to run applications of their choice;
* Freedom to Attach Personal Devices: Consumers should be permitted to attach any devices they choose to the connection in their homes;
* Freedom to Obtain Service Plan Information: Consumers should receive meaningful information regarding their service plans; and
* Protection of Consumer Privacy: Consumers should know that their personal information is safeguarded, except to the extent necessary to abide by law enforcement obligations.
This is, of course, extremely similar to the software libre freedoms propounded by the GNU Foundation:
* The freedom to run the program, for any purpose (freedom 0).
* The freedom to study how the program works, and adapt it to your needs (freedom 1). Access to the source code is a precondition for this.
* The freedom to redistribute copies so you can help your neighbor (freedom 2).
* The freedom to improve the program, and release your improvements to the public, so that the whole community benefits (freedom 3). Access to the source code is a precondition for this.
I suspect the coincidence is intentional.
Users traded availability and reliability for mobility and ubiquity when they fell in love with cellular. Looks like they might not have to trade in quality and audibility when moving to VoIP according to Stuart Henshall:
What’s more this user has learned that Wi-FI Skyping from HotSpots is better than a Mobile phone when available. Thus the paradigm that threatens the landline system may have more impact on mobility than current projections suggest. Some of you may have seen the recent releases of mobile phones like the Nokia Communicator 9500 that provides the traditional cellphone features along with Wi-FI. So now consider the user experience. When they are in a hotspot sound quality goes way up. When they get home their cellphone automatically becomes the home phone and the cellphone and the quality is way up. It’s just possible that the mobile providers are entering a sound spiral as well.
Om Malik recently found that VoIP on today’s cellular data was quite good enough thank you.
Hmm, better and cheaper. Not a hard sell, methinks.
Don’t usually write about work, but I don’t think I’ll be giving any deep corporate secrets away with this one. It’s just a little parable of the cell phone and the smart network.
People who have both cellular and local telephone service need to be put through to a different call center from people with just cellular service. This is because normal cellular care agents can’t deal with local telephone service issues.
But the project for combined service means upgrading the intelligent routing software in customer care to send the calls to the right place. Every inbound call needs intelligent routing. This will take time and money.
The easy way round this would be to re-provision selected customers’ cellphones so that the short code for calling cellular customer care pointed at the combined care center. Except you can’t do that because the end points are dumb and the routing of short codes is done in the network, which in turn has no idea whether you have local service as well as cellular.
So the smart network and smart call center routing system conspire to prevent what would be a trivial one-off routing change if done at the edge of the network.
(There is also currently no way of remotely re-provisioning handsets this way, although there are standards in the offing to do it. But the existence of legacy dumb devices that need a smart network leaves you stuck in the high-cost smart network world forever and a day.)
The moral: start with smart edges and a dumb center and you’re better off. And if you start the other wrong way round, you’ll never make the transition because the embedded base of dumb edge devices imposes an impossible switching cost (if you’ll excuse the pun).
In case you had forgotten, smart networks are incompatible with personal privacy.
Cellular carriers are struggling to see how to integrate WiFi into their mobile data product offering. Is it a complementary or substitute product?
The thing about cellular is that it has asymmetric value. Incoming calls are more valuable than outbound ones. Why? Because there are plenty of substitute ways of making outbound calls most of the time — desk phone at work, home phone, e-mail, call box in the street, borrow someone else’s cell phone. But there is no substitute product for “send this voice message to Martin wherever he may be right now”. (For an illustration of this, see the failure of Hutchinson’s Rabbit network in the UK in pre-cellular days: it only allowed outbound calling.)
And that “find the user” capability is wired into the lower layers of the network architecture through a home agent switching service.
Now WiFi isn’t a very mobile technology. In end-to-end fashion, any “find this user” capability is built into the end points via things like instant messaging presence servers, dynamic DNS lookups or mobile IP. There’s nothing smart in the network itself to support mobility. So it’s philosophically very different in how the value is dished up. The core “find the user” function is freely available independent of the connectivity, so there’s less to charge for.
Even worse from the cellular operator’s point of view, the WiFi connections are not under carrier control. So you can’t filter and meter, unless you lock down the device. But consumers may be unwilling to accept locked-down devices that run on their own connectivity. (Although DVD players and iPods suggest that rejection of lock-down isn’t always the case.) And if you can’t filter and meter, you can’t charge for the true value the customer receives. Cue usual dirge on business model collapse for telcos.
No surprises that the cablecos are trying to gang up on the telcos, as documented by the ITU Strategy and Policy Unit. NeuStar are going to be the ENUM directory provider. The idea is that the calls between cablecos are kept 100% VoIP and don’t go anywhere near the PSTN.
Eventually the public are going to discover that the sole “voice” function they’ve bought is the mapping of a phone number to an IP address. This isn’t any harder than DNS, which comes for free with any Internet connection. The Internet is perfectly capable of doing the haulage.
Repetez apres moi: The correct price for point-to-point voice service is zero. There is no such thing as the “market for VoIP calls”. It’s a mirage. The cost structure is lower than (zero marginal cost) email. There isn’t even anything to store. This is a race to nowhere.
So, The Euro carriers are ganging up to try to stop open operating systems controlled by Microsoft and Symbian running on their nice closed networks. To do this they are agreeing to their own common specification, and sucking in a puppet vendor to build it for them.
From dictionary.com
car·tel n.
1. A combination of independent business organizations formed to regulate production, pricing, and marketing of goods by the members.
Hmmm… production. Debating the meaning of that word will keep some competition lawyers well supplied in Mercedes convertibles and pate au fois gras for a few years.
Of course the whole thing is ultimately doomed. Competition will occur from open rivals. Smaller local rivals and municipal networks will pop up regardless. (Particularly if we have an oil crunch.) And just like people pay a big premium to live near physical communication links, the economic pressure to be near good and cheap data communication links will be equally pressing. Mesh networking isn’t the technological equivalent of hot nuclear fusion. We have it today in a rudimentary commercial form. The only question is whether the traditional carriers have the collective will and power to hold things back for a few years.
UPDATE: More here.
UPDATE: Similar thoughts over at El Reg.
I’ve just been doing some reading on the usual telecom business model debacles, and something popped into my mind. We’re using the wrong words to talk about the participants.
The limits of my language are the limits of my mind. All I know is what I have words for.— Ludwig Wittgenstein, Philosophical Investigations, 1953
Today I’ve read articles about “cablecos”, “cellcos”, and of course “telcos”. This is all horribly misleading. Each of these businesses has varying degrees of forced bundling of connectivity and service. But we really should make the distinction between connectivity providers and service providers truly stark.
For instance, over at Forbes they say:
In a world of Net phones, local monopolies and duopolies will no longer exist; Internet consumers will have every telco in the country competing to win their business.
The word “telco” has clearly lost all of its traditional meaning in the above sentence. And if we can’t consistently use the word to say something, it has no meaning. In the wise words of our philosopher friend:
For a large class of cases—though not for all—in which we employ the word “meaning” it can be defined thus: the meaning of a word is its use in the language. — Philosophical Investigations
(Incidentally, there are some great quotes in the Forbes article: The rest of the telecom industry, however, seems to be in denial of the coming cataclysm. Fears that phone traffic will migrate to cheap networks are grossly overblown, according to the Telecommunications Industry Association, a leading trade group. It forecasts that revenue has hit bottom in the U.S. in every single category of the phone business, from local to toll calls to wireless. Hey, didn’t Christensen say you had your best years just before you fall off the cliff of disruptive innovation?)
Anyway, back to linguistics. First, connectivity. For these, “netcos”. We recognize that Internet connectivity is the key for most public networks. For the rare occasions when you need a truly private pipe, you go to a “pipeco”.
At the other end of the spectrum are pure service providers like Vonage or Amazon. Their applications let you buy connectivity service from anyone. These are “servicecos”.
In the middle you have the legacy business model. Hmm. Something mildly derogatory needed that highlights the limited lifespan of service and connectivity bundling. Must encompass telcos, cablecos and cellcoes. How about “Nogocos”? Too obscure. “Controlcos”? Still not right. Aargh! We don’t have a word to describe the business model we argue is going away!
Any better ideas out there?
There’s a nice quote at the end of this America’s Network article on bland branding of US celular compared to lifestyle branding of Orange, Vodafone and Nokia in Europe.
Ultimately all will benefit from a broader set of higher quality products inherent in the shift by the carriers from voice corner store owner to multi-service, multi-segment department store purveyor.
Of course, the hollowed-out downtowns of many US cities bear the sun-beaten shadows of the names of almost-forgotten department stores from the past. The department store made sense because transport was the constraint to shopping. It wa just too darned hard to run a multi-point distribution network. And it was too hard to shop at multiple locations when you had to hop on a trolleybus (or horse and cart) to do it.
Once users had total and flexible control over their own transport with the arrival of the affordable motor car, everything changed. Of course at first there was a boom in department stores. More customers could make it into town more often from a greater area. Wal-Mart and out-of-town malls didn’t happen overnight. Indeed, history often turns full cycle with the Wal-Mart Supercenter or mall becoming a single shopping destination. But in the meantime most of the old institutions disappear or become shadows of their former glory.
Spot the similarity? Packet networks give flexible control over data transport to users. All the same principles apply. The same future applies to these all-things-to-all-men carriers. Things will look really great as pervasive broadband hots up. Boom times will return. But the business model of control will become an albatross that can’t be shaken off.
How often do you see this same drivel appearing in general business articles on VoIP:
Because VOIP services use public networks and software rather than a fixed circuit, the service is far less expensive to build and maintain than traditional phone systems.
Wrong, wrong, wrong and incorrect. It just ain’t so. It’s nothing to do with transmission cost or efficiency. Indeed, the software inside a traditional telephony switch is not much different in complexity than that inside a modern large router. Most of the effort goes into security and management, not packet or circuit routing. And on wireless, IP is hideously inefficient unless your network goes to great pains to strip down all the IP and MAC headers.
Nope, it’s about bypassing the levies and tolls. It’s about people and their control of economic resources. Packetization technology is a side show that happens to drive the change by naturally moving control to the edges. Ignore the efficiency canard. It’s wrong and irrelevant.
So, Nextel has gone commercial with its Flarion deployment, as previously described here. Here’s the pricing. Similar to normal wired broadband.
This is going to cause an oh f*ck moment at some point in the next two years at the traditional cellular carriers. (I’m assuming the wireline carriers are oh so f*cked already.) The question is will Nextel have the balls to carry the open IP network stance all the way through, and chuck in VoIP PSTN calling for a nominal cost? Will they try to lock down the devices? Can they manage peak usage and contention the right way?
My approach if I ran Nextel would be do an exclusive deal with Flarion, and slowly kill all my regional competitors, one market at a time. No big bang nationwide launch. Buy up their bankrupt spectrum. Repeat until false. Then eat the big guys.
I can see the marketing campaign now. TV ad with woman in shower, lathering hair. Partially draw back curtain, pokes head around. Says “DSL and cellular bills? No way! I just surf and phone with Nextel Wireless Broadband”. (With apologies to Procter and Gamble and those who never saw the infamous original 1980s TV advert.)
Two very interesting stories over at Slashdot on how scammers are using relay services for the deaf to anonymize their telephone transactions.
The Arizona Daily Star is reporting on how 419 scammers and credit card thieves are abusing the US’ TTY service which enables hearing-impaired citizens to make phone calls with the help of an intermediary operator. ‘The callers try to use stolen credit-card numbers to make big purchases of merchandise from American companies. The operators often suspect fraud, but they can’t just hang up. Federal rules require them to make the calls and keep the contents strictly confidential.’
This will only be solved by some for of digital identity. Identity is going to be the hot story of the decade. Just like in 1994 few people could see how cellular would explode into the general population, a decade later the need for better digital ID is equally pressing and under-recognized.
My only hope is that we find a happy medium on privacy. You don’t need to be personally identified in order to use this sort of service. That makes deaf people second class citizens in the privacy stakes. But you do need to have membership of the “verified deaf American people” set to get through. Deaf people should be able to launch scams through this service if we’re to preserve the right to privacy and anonymous communications. Then we’re all equals. It’s a funny world.
UPDATE: More here.
Someone’s got some good ganja cooking over at Always On:
The core technologies to make personal broadband a reality are already in place—Wi-Fi, for example, is a personal broadband technology. However, Wi-Fi does not represent a sustainable service model because it uses public spectrum, because it does not support end-to-end class of service, and because it represents an addition to the existing cost of broadband access. The upshot is that we’re still missing a clear global market vision and understanding of the personal broadband opportunity and its scope.
Since when did a technology need money-making middlemen with a “sustainable service model” to become blessed as a valid concern? Just because nobody can engage in rent-seeking behaviour and make globs of profit doesn’t lessen the concept one iota.
“I’m sorry sir, you can’t marry this woman as you do not appear to have been introduced by a licensed introduction agency to whom you have paid the appropriate fee.”
Having been reading about Session Border Control systems, I’ve been pondering over what is really meant by “the edge of the network”.
The classic end-to-end argument is that a dumber pipe is the better pipe because it is can be repurposed to applications never conceived when it was created. The consequence is that all the application smarts should be “at the edge”. This usually implies a single terminating device.
But we’re moving to a world where functionality is increasingly spread among multiple devices that co-operate to complete tasks. Is your “TV application” inside your TiVo-style home media server or the smart screen that decodes the video stream? Or in the neighborhood content cache?
We live in a very end-to-end impure environment chez Geddes. We have a NAT router and a caching web proxy server at home. So I don’t have an unmediated path from my PC to the Internet. But I don’t care, because it is all under my control. The fact that a bit of my network is “smart” doesn’t trouble me one bit.
So perhaps the classic end-to-end argument is too techno-centric. The real issue isn’t one of dumb pipe vs. smart pipe, and pushing functionality out, but one of control. It’s the politics, not the technology, that matters. If I control what’s going on, then I don’t care how dumb or smart the network is. If the smarts aren’t what I need, then I can replace them.
So if the PSTN were to ever evolve (picture: Hades, snowdrifts, river Styx, ice flows), here’s what it would have to do to survive. All those switches would need to become open and flexible. They would enable me to run my own agent software. I could use the circuit reservation capabilties to create new functionality that a non-QoS IP network might find harder. I could write better presence management software to deal with wireless devices that go out of range (a classic failure scenario of end-to-end: a disconnected smart device can’t do jack). I could create triggers that would operate even when my power is out at home disconnecting my smart home voicemail server.
We already have the technology to do this. Virtual servers, virtual machines, process protection. Stuff IBM and Sun have worked on in the labs for decades. We have the identity technology to know who is running what and where is came from and who to whack if it misbehaves. It’s just a question of making it happen. Which will never occur, because telcos believe loss of control is their nemesis, not their savior.
From Russel Beattie comes this gem
Okay, about the Motorola i730: It’s a really cool phone, but… you can’t put your own apps on the phone even with a cable unless you get a special application approved by NEXTEL! Dammit! Moto has this program called the Java Application Loader (JAL) but only “JAL Lite” is readily downloadable and it won’t load any application that uses the network classes. So even though I bought the cable and ran that app, I still couldn’t test our map-based J2ME client here at work on the phone (which was the whole reason to get the phone in the first place!). I’ve submitted an application to NEXTEL to get the full-featured JAL and I’ve been Googling for workarounds. Stay tuned on this subject.
The PC is an open architecture, and Windows is open in the sense that Microsoft doesn’t control what apps can run on it. (Although if you make too much money doing it, watch your back!) These PCs are all wired up via an open network, the Internet. The result was a stunning success.
Cell phones have been locked up from the get go. The architecture is closed. The carriers control which applications go in their portal, and how applications are signed and what APIs they can access as a result. The result has been mediocrity.
The one big mobile applications success has been iMode and its imitators, which created a pool of “official” branded content whilst also positively encouraging a primordial soup of innovative new applications. The technology was made as un-proprietary as possible within the constraints of the device.
Prediction: open systems will continue to crush closed ones. Why? The obvious reason is that open ones are fitter in an evolutionary sense. An application that almost hits the mark is easily superceded by one that scores a bullseye. In many ways, that’s the core concept of end-to-end or the stupid network.
There’s a less obvious reason, too. It’s what I would call the openness ratchet effect. The easy, zero cost way of differentiating yourself from your competitors without incurring product development costs is to be a bit more open. Someone will always make the jump.
Funnily enough, in the music business they’re busy trying to lock down the entire system. But the harder they fight to lock it down, the more boring and corporate the music has become, and the lower the sales. People have been substituting their entertainment dollar for other less restrictive and more interactive products.
The music industry relies on legally granted power over its customers to perform price discrimination. For the music industry, they want to charge a vastly increased amount to play a 30-second clip in a TV advert compared to you playing it on you iPod. Copyright is the mechanism to do that.
For telecom, the price discrimination control may be:
In many ways the primary purpose of an old-fashioned telco isn’t pipe provision but price discrimination. Smart networks to classify, filter and meter traffic. Complex pricing plans and megamillion dollar billing systems for contractual confusion. And an army of lawyers. Price discrimination of service use is the core competence.
The trick to survival in the long term is to forget about piracy or disintermediation, and think only about profit. Get a smaller slice of a much bigger pie, rather than try to grab the whole pie for yourself. Unfortunately, the music business and communications corps keep backing themselves into a corner. New value creation will ultimately route around your toll gate, like it or not.
A quick extract on a recent Pulver email on VoIP:
* If IP voice services are to be regulated, will it be at
the European or national level? Does it matter?
* When should IP voice services be classified as PATS [publicly available telephone services], and why?
* Can public services — such as access to emergency services on
VoIP — be assured in the absence of regulation?
Oh my. We’re stuck in a situation where the circuit-switched mentality still controls our vocabulary and thoughts. Every part of the application value chain is separable. There’s more than just application service and bit haulage. You have to look at all the individual parts and decide whether to regulate.
In this case, the concept of regulating VoIP is meaningless. Are you going to regulate voice chat embedded in multiplayer games? I don’t think so. Are you even going to be able to define VoIP? Detect VoIP traffic? Nope. So the bit that’s left to regulate is the namespace. The digital identity made from telephone number digits.
You can control the issuance of telephone numbers. By definition, you need a central authority to state whether a particular number is already issued and to whom. Centralized things are amenable to definition, detection and regulation. If IP-based voice services want to hook into telephone numbers, then regulate that interface. But don’t regulate a service you can’t even define.
Even better, moving the regulatory focus to the identity sphere, you aren’t inhibiting competition and innovation. Anyone can invent a new namespace and stick up a directory. Namespaces are not a scarce resource being artificially rationed.
You can regulate PSTN-alike VoIP services where there is an identifiable service provider. But regulating VoIP as a class is like regulating anger or success. It’s an ether you’ll never put back into a bottle.
UPDATE: For those interested on how language regulates thought, read the works of Wittgenstein. Then please report back here, as I’m too lazy to read it all myself…
This rather nice summary of the impact of VoIP on call centers reminds me of something I’ve been meaning to get off my chest.
I’d just like to say some rude things about the concept of an “IP PBX”. (For the acronymically challenged, this is a packet-based teeny weeny local phone exchange you stick in your enterprise and lets you dial Sharon in accounts by just entering her extension number.)
What an utterly cretinous concept. Does anyone believe this market has any future whatsoever? At worst it will be crushed as SIP call routing becomes a minor feature of Microsoft’s Real-Time Communications Server (at a tenth of the cost). At best, you’re left with a routing directory with the profit potential of a domain name server (i.e. none).
There must be some good money to be made shorting the stock of suppliers of dead-end technologies like this.
The International Monetary Fund (IMF) has had a rather controversial history. The IMF is sort of an economic policeman. And, as the song goes, a policeman’s lot is not a happy one. Picture an indebted and impoverished country that has been pillaged by its leaders enough to collapse into insolvency. That’s when the IMF’s bankers and economists get to fill a few extra pages of their passports.
Any medicine for such sick patients is unlikely to be pleasant. And potentially unwelcome, too, when not every previous patient has gotten better. Even if they did a splendid job and made everyone a relative rags-to-riches success like South Korea or Costa Rica (not IMF rescue subjets), the IMF probably wouldn’t have crowds on cheering admirers throwing garlands of flowers in their path as they emerge from passport control.
I believe it was in Tunisia that an interesting phenomenon was observed. The IMF’s aim was to gradually increase the price of bread to reflect the world price for wheat. Being sensible chappies, they staggered the price rises. The first went into effect, the market price rose, people still bought bread. The second went into effect, up went prices. The third happened, and there were riots on the street.
What this illustrates is the breakdown of free market economics. Lawlessness undermines the foundations of freedom of contract, price discovery and open trading.
Which brings us to the most excellent talk from Professor Deffeyes at WTF!?! last weekend. In a nutshell, the Professor demonstrated that world oil production is almost certain to peak next year, and then begin a steady decline over decades to come. Which is a bit of a disappointment if you’re a newly minted member of the Indian or Chinese middle class hoping to buy and run your first car. Because richer Americans, Japanese and Europeans are going to be sucking up what’s left at ever increasing prices.
The invitation that David Isenberg extended to Prof. Deffeyes was particularly cunning. At first sight, you have to ask youself why we’re looking at oil production at a telecom conference. Then the talk starts to get interesting and you think, “OK, this is interesting enough I’m glad he came, it was still worth the money.” But what we were really getting was a lesson in supply and demand. And what happens when supply goes keeerunch. The obvious answer is that prices rise. But here’s the non-obvious bit from my notes:
One possibility that economists love is that we will ration by price. Nixon govt fixed the price of oil. Ration by inconvenience – lines for gasoline. End of WW2, Roosevelt had ration coupons. Will be some form of rationing. What happens when gap between supply and demand opens.
Aha! So last time this happened the price mechanism was only allowed to go so far. Before long, the clamour to “do something” by those disadvantaged by high energy prices became too much. And rationing was de facto introduced. (The same thing happens with the National Health Servce in the UK. Extended waiting times are an alternative means of rationing without the political pain.)
So what we see is a transition from a market economy to a political economy. In the former, production and prices depend on what you offer in return. In the latter, it’s down to who knows whom and has influence. So oil, even in the richest countries, can behave like Tunisian bread. At some point the assumptions of a well-behaved market go wobbly and you’re left with a mess.
And there’s more. The oil industry clearly has lessons to teach telecom. The current strategy of most incumbent telcos is to restrict supply, much as any good monopolist will do. But oil and telecom don’t exist in different worlds. They are, to some degree, substitute goods. Instead of shipping goods around the world at each stage of production, you can ship the knowledge of how to process them. People can telecommute and teleconference. An oil crunch means we’re going to need a lot more telecom.
Which in turn means we may see an intensification of the political economy of telecom. In many ways, the core competence of most large telcos and cablecos is lobbying. In the US there is the Telecom Act, sundry FCC rules, the allocation of broadcast spectrum, franchising rules, and so on. Each acts to make entrance of new competition harder. Even if the label on the tin says the opposite. I’ll say that again. Many telcos only make money from their copper landline operations, and that only happens because the political system makes it that way.
This doesn’t paint a happy picture. On one hand you’ve got a potential explosion in demand. And on the other you have incumbents dearly set on restricting supply (Where’s The Fiber!), and the political leverage to make it happen. And if Tunisian bread is a guideline, the result is going to be a riot.
UPDATE: More comment here from the Angry Economist.
UPDATE: New York Times article here
UPDATE: The “peak oil” hypothesis is just that — a disprovable theory, not a fact, as this article (pdf) describes.
The talk at WTF by George Gilder on Korea has set me thinking about structural causes of broadband uptake. The south of Kansas City when I am sat now is one vast suburban sprawl. Fifteen years ago, 95th Street was the edge of town. Now when we take our daughter swimming we head down to 159th Street. The few fields left down there have “For Sale” signs on them awaiting development.
I’m just wondering if the US appetite for consuming land is a leading cause of the relatively low uptake of broadband. The US is clearly no broadband backwater, but given its riches and economic dynamisn, you would expect it to be doing better. All that sprawl just increases the mileage the cables and fibers need to be laid over. And there’s an R^2 effect. Make those houses 50% bigger in each direction, and you’re only getting 45% of the density of properties per square mile. That makes broadband wireless more expensive to deploy.
Even worse, VDSL can’t deliver high speeds over long distances of copper wire without lots of local repeaters. And the uplinks of wireless connections are pretty sensitive to distance from the central tower.
So perhaps the future of American suburbia will remain two kids, an SUV and an expensive trickle of data for the forseeable future.
At least they won’t have the problems seen in the UK where cable TV laying got into trouble because so many urban pavements [=sidewalks] have trees embedded in them, and the cable TV folks where cutting their roots. Here there aren’t many sidewalks. Only freaks and deadbeats walk.
At there’s no danger of a hill blocking the line-of-sight view of a cell tower in Kansas…
The usual suspects have been bleating to the FCC that wicked Internet access providers are trying to turn a profit.
Tech heavyweights explain how to destroy the Internet
… The best speech came from University of Virginia Law School Associate Professor Tim Wu, who cited actual examples of industry abuse worth regulating against. … He mentioned as well that broadband providers, Comcast in particular, have restricted or banned the use of virtual private networks (VPNs). The idea here is to charge the customer as a business user, rather than a home user, and extort extra money. Servers and VoIP have also been banned in places, to protect other services that the provider offers.
I’m afraid that whilst such activities may raise the ire of customers and create a lifelong loathing of their connectivity provider, they aren’t worthy of regulation. Price discrimination is a GOOD THING because on average it tends to reduce prices, even if some people suffer.
In the days of PTTs and Ma Bell, it was reasonable to restrain the mandated monopoly from exclusionary or discriminatory tactics. But if you’re looking for an explosion of wired and wireless connectivity possibilities, then those new entrants need a carrot of juicy profit to attract them to take a risk.
The role of government is firstly to ensure the customer has a clear understanding of what they are buying into. Only contracts freely entered into with clear terms and should be enforceable. Words like “broadband” and “Internet” in the advert should imply certain minimum functionality. Secondly, new entrants should be protected from predatory pricing of incumbents. And that’s it. Consumer protection and anti-trust.
If we’d had that rigorously enforced thirty years ago, there would have been no need to break up AT&T. The market would have done it for you. Of course, as we’ve seen with Microsoft, that isn’t easy to implement in practice. (Plus the astute observation at WTF!?! that the first imperative of regulators is to perpetuate the need for regulation.)
If the “tech heavyweights” don’t like capitalism, I’d rather they said it directly than try to regulate around it. Also, focusing energy on resolving symptoms rather than causes just diffuses the forces of communications freedom. And holds you up to ridicule for economic illiteracy, too.
One comment someone made at WTF!?! was whether people who work for monopoly telcos know they are doing evil [in holding back the tide of human progress]. There were many references (including my own) to the big, bad telcos. I’d just like to put my amateur economist hat on for a moment, and pass comment.
I’m in essence in agreement with Professor Milton Friedman of the University of Chicago, who believes that business’s sole duty is to make profit. I think we all acknowledge that there are ethical bounds that surround that mission. You can’t bump off your competitors too literally. But declining to extract the maximum from your customers isn’t an ethical issue. You MUST do it. And it’s a good thing.
Why? Because that’s how the price mechanism works. It sends a signal that attracts competitors with better, faster, cheaper products. You can argue that the price mechanism is ineffective because the telcos work to avoid competition by rigging the rules. But that doesn’t mean eliminating the monopoly price signal is wrong. Regulated prices just entrench the status quo forever.
So, no, telcos aren’t evil. They have no moral scale you can pin them on. They just are. If you don’t like it, go set up some competition, or agitate to change the political rules that keep them artificially in business.
The masters of serendipitous eclectica over at Boing Boing are pondering how to deal with the bandwidth bills resulting from their inorexible success. Indeed, I occasionally live in fear of a sudden overage bill if this site suddenly got a spike of visitors (I should be so lucky) or suffered a denial of service attack.
The answer is simple in theory if somewhat hard to implement in reality. Don’t centralize what doesn’t need to be centralized. Boing Boing is a “bag of bits” website, where the only interative feature is the search function. We need to start thinking of how to re-wire the web into a peer-to-peer architecture. Don’t store static data centrally. Just store the pointers who who might have a list of where it is (a-la Napster mk1 or BitTorrent).
Of course, Boing Boing can’t alone be expected to change the world this much. But you can’t escape the cold clutch of economics. If the only way of getting your message out is to reinvent the web, so be it. I can imagine a day in the not too distant future where your visit to a website returns a lightweight page like this:
We’re sorry, but we have reached our monthly bandwidth charge cap and your user agent “Microsoft Internet Explorer 7.0” does not support distributed web page delivery. May we suggest you upgrade to one of the following products…
The interesting thing about BitTorrent is the algorithm by which download speed is coupled to the uploads you’ve made available. No free riding, folks. Applied more broadly, this would greatly discourage the use of dynamic IP addresses (which make acting as a server awkward), NAT devices (hard), and firewalls (impossible). Worried about the balkanization of the Internet? Then just set up the technology so the economics are in your favor.
I bet someone is already working on peer-to-peer locally cached web delivery. (If not, I want credit on your patent application.) Anyone want to enlighten me?
Due to overwhelming demand, here’s an index to the transcripts.
WTF Sunday Afternoon Sessions - Talks [on Wireless]
WTF Sunday Afternoon Sessions - Developing World Telecom
WTF Sunday Morning Sessions - Sifry [Politics and Internet]
WTF Sunday Morning Sessions - Operations Culture [Peopleware]
WTF Sunday Morning Sessions - Corporate Culture
WTF Sunday Morning Sessions - Short Talks
WTF Saturday Evening Sessions - Deffeyes [Oil, Supply and Demand Thereof]
WTF Saturday Afternoon Sessions - Wireless Wonderland
WTF Saturday Afternoon Sessions - McGarty [Municipal Fiber]
WTF Saturday Sessions - Short Talks
WTF Saturday Morning Session - Noam [Infopocalyse Mispredictions]
WTF Saturday Morning Session - Netches [Putting Change in Perspective]
Before I collapse into bed, just a few words on WTF while it’s all fresh. Wow. An orgasm for the intellect. I must thank David Isenberg for pulling together an amazing crowd, excellent presenters, diverse agenda, and top-notch organization.
My biggest take-away is the historical significance to the species of getting things connected. It’s a mission, not a business. Trashy telcos aren’t the central issue. We’re too close to it to see the implications. I want my grandchildren to know what it was like when it happened.
OK, the last session. What a panel! Tim Shepherd, who modestly disclaims inventing mesh networking. Gordon Cook, the sharpest quill in telecom. And Col. David Hughes, and extraordinary gentleman and soldier who has been wirelessly connecting the ends of the Earth.
Tim Shepherd
PhD thesis [Theoretical basis for mesh networking] misused. Used to whack FCC commissioners on the head. Things we think are settled in wireless are not. Want to teach a few technical things.
Imagine 100,000 people go to a football stadium. Acoustic spectrum for speech only a few KhZ wide. If everyone talked at once nobody can talk. Wrong! But you can talk to neighbour if everyone else talking at once. PA system still works. Can use intuition of football stadium acoustics as metaphor, with limits.
Where do we need more regulation? Acoustic or electromagnetic. Need more acoustic regulation and less electromagnetic.
Electromagnetic waves do not interfere. Slit experiement is inferference. Economists and lawyers use it in a different sense. Analog of two cars going to a single spot.
Interference occurs happens in the receiving apparatus.
On open spectrum list someone said there are times everyone in football stadium needs to shut up, for example during anthem. Life isn’t like a football stadium. 3am music – call police. In theater, kicked out if I annoy my neighbours. Response – football stadium analogy not perfect. Need less refulation than acoustics.
Hearing different from electromagnetic. Can’t redesign ears or brains. If neighbour transmits 100W of RF only a problem if you have a receiver. No receiver, no problem.
Second way it is different that there are 5-7 orders of magnitude more bandwidth. Like being able to listen to 50,000 at once. Also small wavelengths mean you can have directional signal.
So yes there are times when we expect people to be quiet in the real world. What does it mean to be quiet on the Internet? Maybe on a mailing list, but can’t expect people to keep it down and stop writing so much on their blog.
Military should be able to do what they want. Robust civilian systems like GPS need to be robust against people with malevolent intentions. Then we should have a free-for-all.
Gordon Cook
Three things. If I’ve learned anything in the last 6 months and the last 2 days. Former: wireless everywhere, ubiquitous, gigabits to kilobits. WISPs to RFID to tiny remote sensor networks. All interlinked in intriguing ways. To the extent that there is room for innovation left in ICT field it is in building this global, ubiquitous wireless network.
In last 48 hours have seen things get clear. It’s boundary crossing. Dinosaurs getting weak at the knees and keeling over. The only opportunities for creative existence not shackled by constraints of corporate business model. Need to understand how all the pieces build on relationships. Interesting business propositions.
Chaos, but can see enough of the big picture. Three years ago was losing subs. Asked Dave [next speaker] for advice.
Went from supercomputer center to office of technology assessment. Didn’t understand what it was all about. Start a newsletter. How to do that? Just start and keep on going. One foot after another. 3 years later meeting Steve Goldstein in Moscow and talking on NSF and international connections system. My devil’s tail not so long, so introduces to person running BBNS presentation. Head out to Colorado in 1994, meet Dave who is hooking up school in mountains at 56kbps.
Mitchell – passion is getting K12s onto the Internet. Hate the RBOCs. Whoa! Do I have the person for you [Dave]. Dave been doing amazing wireless stuff ever since.
Went to Tibet and Nepal in 98/99, Kashmir in 01, Nepal in 02. Had a revelation. Guy who runs largest ISP in Kathmandu. 10 years earlier was a student in Montana, student of a student of Dave’s. Said there’s an interesting Sherpa putting a telecom on the side of a mountain. Met about 4 Nov 02, talked to him for 45 mins. In 3rd world, Maoists had blown up the repeater tower. Get a satellite dish, PBX. Needed to explain whole copper and PBX stuff. How much is a box and can I stick it into a plane and fly it into Namche? Sure. Explaining how he did all this. Suddenly, sentence that changed everything with my relationship with Nepal and Colonel.
But that’s not the only reason I’m doing this. I want to preserve Sherpa culture. Want to put Tibetan characters on computer. Nobody under 35 can write Tibetan. Have Tibetan culture, not Nepali. Now using telecom to enable local indigenous cultures to survive. Another human being has come to same conclusion independently. From straight back to Toyko, New York, imedially call Colonel [Dave]. Back from Mount Everest – news for you. You’ve been reincarnated on Mt Everest. Met this crazy Sherpa. Spent 02-03 getting cybercafe to Everest base camp. Dream to go there. Did it.
Colonel David Hughes
Don’t go to many conferences. Isenberg dragged me here. Don’t like to do anything for people. Just go there, do it while they watch, let them learn, I have no intention to stay behind.
Hedy Lamarr, Inventor of Spread Spectrum. 1941. Never got recognition. I nominated her for award that got her world fame. Never occurred to these people who used her image in marketing she might still be alive (84 years old). She sued and for $5m.
Picture of her actual patent, released for public access in 1985.
Classified for 40 years by FCC. Too secure for intercept.
1995 and Israelis were playing with the technology. They then loosened up. Three NSF years at the edge. Rain forests of Puerto Rico, permafrost in Alaska, lakes in WI, outer islands of Chesapeake.
Nobody is doing modeling of wireless properly, just need to apply known biological science model techniques.
Towers, trees, rain, rats problems in Puerto Rico. Need to cut through canopt. 902-928 Mhz radios.
Lake Wisconsin, problems with ice, wireless in middle of lakes. Tower height, $7000 to get 125ft up, above trees. LOS problems? Use helium balloons. Small radios in slings. Hooked into wells [against wind and snow].
Chesapeake, offshore. Univ of Virginia studying out there, 14 miles of water. You can get range out of wireless, just need to know how. Maybe a little bit of civil disobedience in public interest. Only thing that stayed up during hurricane was this one link.
Alaska. Fairbanks. Central AK. 16 miles down the river to collect data from sensor, swap memory modules.
Real-time demo of webcam up in AK, -7C , date and time is now.
Caribou, vertical solar panels, not enough power still.
Power is biggest issue of mobile.
Everest. 18000 ft up. Can’t put solar panel and satellite up on iceflow next to base camp, as glacier moves 4ft a day. Called up friend in Cisco. Asked to give radios in return for PR. Got it all hooked up. VSAT on hard ground, local link to basecamp. Picture of Sherpa with directional antenna. Can then charge $1/min to climbers who can communicate with rest of the world.
Then off to Namche, but 76 years old. Had to train in CO. 2000 feet in 2 hours without collapsing. Namche surrounded by cliffs. Start with Lukla, don’t bring US radios, just being cheap ones from Singapore with low power needs.
Cybercafe. Success. Any climber could get to top of Everest and send email home to mum.
In Thame [nearby village], very poor, trying to raise money for books (much poorer than Namche). 5 relay radios. Picture of Moats [HELP! I know nothing about these micro smart-dust radios, even how to spell their name] Future of radios, tiny.
Economic development next topic. Examples of Bhuddist art they are selling.
Demo of SJPhone, ready to call – contact Mingma, sherpa in Pittsburgh, learned English in kathmandu. Couldn’t afford to go to CMU. Distance learning VoIP to classroom. 1500ms latency.
[Wifi problems in auditorium.]
Moats demo. TinyOS Linux derivative. Looks small. But two big dry cells. Haven’t paid any attention to alternate power. Sensors – sound, accelerometer. Smart dust 2mm x 2mm. Stansby power 1 uW. Transmit 1.5mW. Receive 1mW. Mesh range @915Mhz, not spread spectrum.
Duck Island in ME off Bar harbor, had Stormy Petrel birds. Why survive there and not next island. Put 150 old style sensors in, no room for bids in nest!
Fuel cell technology. Fuel cell with printer techniques, H2 hydride fuel cell. Three 1V cells. Smallest ones will fit on a bee’s back.
Want 20 years unattended of a fuel cell harvesting power. Tiny power, tiny storage. Want to know how to tap evergy in changes in atmos pressure. Why not? Trickle charge. Can last for years that way. I don’t want to be the guy who goes out there to change the batteries [Audience: what’s that PhD students are for!]
DI: Tell them about the award you won! We just think he’s great, given him the only standing ovation of the show.
Flattered, humbled, astonished. Association of graduates of west point. Schwarzkoph. President of Phillipines. Am Distinguished West Point Graduate for 2004. Going in May. Only 7th colonel to get it. Aldridge and Borman who went to moon before me.
GC: Crossed political correctness boundary.
Women at West Point – any nation that permits and encourages its women to join an institution aimed at killing people and destroying people is morally bankrupt.
Q: First time I’ve heard spread spectrum being locked up by military. Wanted to confirm this had been around for a while. Defence department has affected direction of [economy].
[Story of Hety Lamarr - you go Google for it yourself]
GC: I visited David in 1989 in CO. Will have Global bulletin board. Working stiffs will see this. Thatcher will talk to Brezhnev. Only 7 years later was putting Mongolians onh the Interet.
Third world can’t go wired except in main cities. Must be wireless. PDA sized thing.
Two interesting speakers on the real opportunity in telecom, emerging markets, a.k.a. places you want to visit on vacation but not live in.
Intro: Lane has been wiring Africa. Rahul is public policy researcher at CMU.
Lane Smith
Mickey Leland, congressman from Houston when it was very racist. Fought against hunder. Plane crashed in Sudan and died. Asked for $25m from Whitehouse, so USAid stepped in.
Connectivity map of Africa in 1995. Very little. By 2000, all capitals and many major cities connected.
Two great privileges in life. Started with Peace Corps, now with USAid. Funded by tax dollars. A privilege to be a steward of those dollars.
I didn’t do any of the roll-out. All done by Africans. I just sat in Washington. ISP from Morocco installed gateways into 10 countries.
Knew old telecom model of monopoly wouldn’t work. 40 countries at a kick-off conference. All countries wanted internet. Only 10 wanted retail competition at Internet. How many willing to offer cost-based tariffs to make it work. Same 10. Worked with those countries. Last country to sign up is Eritrea.
To give idea of bandwidth scarcity, USAid bought “gigantic” circuit of 128kbps for 1st year. Soon had ISPs wanting more bandwidth. Go to telcos. Either couldn’t work out how to do it, or refused (competition with own ISPs).
Issues: Policy reform, infrastructure access, network of users.
All countries had regulatory countries, not independent, staffed by current or former telcos people. Needed 4-5000 people trained in regulation. Have formed industry regulation associations.
Nigeria: 450k phone lines, 20k cell phones at start, deregulated, 18 months later, 2m cell phones. Still growing that fast. Catalytic infrastructure. Incremental.
Rahul Tongia
Carnegie Mellon University. Engineering and public policy. Faculty of school of CompSci.
Background academic, lots of real-world experience. CMU involved with India number of years ago. Presentation to prime minister officer. Don’t treat voice and data the same. Build all-optical network, leapfrog. 1998. Spent three years building Internet++ for India. Walked out because of politics. Need to be ready to play dirty. Can’t just win on merit. For example, CMU project funded by DARPA, think it’s a back-door to take over India.
Lessons: International is very important. UN meeting recently. 90% of brazil tax returns done online. Internet important. UN important. Therefore UN = Internet. Hmmm… Scares engineers.
Compare voice versus Internet. Developing countries have call completion charges. Raises forex. In Internet would they have to pay to send and receive data.
2.8bn live in poverty (<$2 day). Internet far from the minds of people with lack of food and electricity. Developing countries don’t know how to react to change. US has incumbents. Developing countries the incumbent is the government. Restrict VoIP, unlicensed wireless spectrum.
World telecom development indicators.