Continuing the theme of pulling together related stuff from my grand RSS catch-up, here's a few nuggets of fools' gold.
Despite claims to the contrary, the wired world is not flat. It's kind of hilly, with lumps, drumlins, escarpments and dips. Here's some examples.
Aswath et al ponder the need for Skype to maintain supernodes in order that nodes behind a NAT router can participate in peer-to-peer networks. The issue is that this creates a cost on the sevice provider, and who will bear it. The obvious answer seems to be that if you refuse to become a true peer on the Internet, then it shouldn't come as a surprise if you start to have to pay someone to bridge the gap. The seams between private and public address spaces are ditches and dykes on our Internet landscape.
What's unresolved is whether this will be solved within each application (Skype does it's anti-NAT think, MS does it inside MSN messenger, etc.); or whether "stupid" general-purpose overlay networks will emerge that restore global public addressability.
This is at heart an economics problem, not a technical one. How to form a market between those in need of the overlay, and those providing it? If the only "payment" mechamism (bearing in mind not all payments are cash) is attached to an application like Skype, then the overlay will be owned by the application. On the other hand, we might see systems where I and a group of buddies agree to make our PCs into a community overlay/relay network. If I'm out and about, and stuck on a private address space on some hotel Wi-Fi network, I just VPN back into my private relay network -- possibly hosted on a friend's PC in his basement. Skype's model of seeding the market with some centralised supernodes isn't the only way of balancing supply and demand. Why can't my buddies volunteer to be my personal supernodes of default?
Moving up the stack to the application layer, we again see uneveness and wrinkly surfaces. I used to believe that the correct price for a phone call is zero. Now I believe that is only true within social or trust circles. When a communication crosses a social escarpment -- like from a business to a consumer -- or delves over a trust hollow -- like when strangers call you -- then money is likely to change hands.
A fascinating potential for this comes with the end-user adoption of revenue-share numbers in the UK. (See my earlier article for background on this.) In James's example, PSTN/VoIP users in the UK acquire 0844 numbers, and inbound calls generate cash. This cash is then used to fund outbound PSTN calls.
Now think about it. What if you do all your banking online, and hate receiving special credit card offers and mortgage refinancing pitches at 7pm? Just get yourself a premium-rate number, and make that into your virtual number for your bank to call you on. I'll gladly stand there and listen to a telemarketer for $2 a minute!
In other words, there's a brief temporal monopoly -- someone realy wants to contact someone else, and there's only one address to do it on -- which creates an economic opportunity.
Now you can start to see the sorts of business models that virtual operators like Skype might deploy. Their PC software client is fine for user-to-user calling, but not so great as an enterprise tool. What if you were encouraged (financially, if need be) to give a tokenised version of your Skype ID to companies (to protect your privacy), and they were then asked to use the billable Skype enterprise version to contact you?
Skype (or it's rivals -- it's just a shorthand) then becomes a distribution network. It could be for user attention, content, transactions, all kinds of stuff.
So to the extent that there exists market power, however limited in scope and time, between caller and callee, money will pass. That means I disagree with Richard when he says:
If finally there is only "free" IP-IP communication and no real "service", only an applications and products, there is not more business case for specific VoIP providers then for e-mail providers.
This is only true within the trust circles. Is my bank my buddy? Not in the same sense Richard is, and Richard isn't my buddy in the same sense my wife is. I spent 20 minutes yesterday fielding questions from a journalist. Is he my buddy? Answers on the back of a banknote, please.
In conclusion, these lumps and bumps exist, and often for good reason. People partition networks into private address spaces because they see more value in it than the loss of public addressability. People will pay to remain uncontactable by some parties; or will be forced to pay to contact other people. There's money to be made in putting bridges over the ditches and tunneling through the ridges. And it's a feature, not a bug.
UPDATE: Actually, when I was conceiving this article, I had another example in mind. With peer-to-peer networks, there could be an incentive to source content from an "on-net" local node.
If upstream capacity is limited, and -- a cruicial gating factor in future -- costs the user more than local on-net capacity, then you want to copy that file from someone else on your street. This is a problem, since today you don't have easy access to the subnet architecture topology, so it's only by trial-and-error you can guess who is local. Even worse, imagine you're on an ISP plan where upstream access is metered/capped, whereas on-net is not. Your P2P application that sees the world as flat won't work.
Your access provider is then in a uniquely privileged position to offer their own closed P2P distribution system for video and other large files, because only they have access to the unseen pricing and volume meters. Hence the access provider gets to reassert some control over their network, because the world isn't flat after all; IP just provides an illusion that it is.
This on-net/off-net issue is normal on voice plans today. How long before the same applies to data transport?
Posted by Martin Geddes at 3:28 PMTrackBack URL for this entry:
http://www.telepocalypse.net/cgi-sys/cgiwrap/mgeddes/MT/mt-tb.cgi/470