Ok it seems that more than a few people have had similar misconceptions, so I'm going to try to clear this up. The idea that the internet is a p2p network is a pernicious falsehood. The only real way that the Internet is a p2p network is in inter-AS routing. You are not part of the p2p Internet, your neighbor is not part of the p2p Internet, and your company is probably not part of the p2p Internet--you're just leasing use of the p2p Internet.
By this I mean that the Internet is fundamentally not as distributed as it seems. This is the flaw that most people who argue for wireless mesh networks have. The don't actually understand the technical complexity in how the Internet is run. But anyway, let's admit for the moment that the edge network is not decentralized (it is completely dependent on its ISP). Well, then the Internet is kind of decentralized, right? No. In order to cover the vast areas of land and ocean that connect the Internet we require massive fiber infrastructure development. But if that were somehow provided it would be distributed, right? No. As two Northwestern phd students and I measured in a paper in submission, more than 60% of p2p traffic passes through one of a handful of Tier 1 ISPs.
No matter how you swing it, the Internet is not really decentralized, it is balanced between a few powerful companies and governments (which thus have to operate under the law of their hosting countries). A distributed or unregulated Internet is a pipe dream.
This is the flaw that most people who argue for wireless mesh networks have.
I'm curious what you mean here. I am reasonably sure that mesh networks are actually distributed, at least internally. Are you referring about the Internet gateways that are generally a part of a mesh proposal, or possibly that mesh networks don't scale like centralized networks do?
A distributed or unregulated Internet is a pipe dream.
I think that we could create a truly distributed network. It would not be nearly as efficient as the Internet we are accustomed to, and it would not be based on the current Internet Protocol.
I'm curious what you mean here. I am reasonably sure that mesh networks are actually distributed, at least internally. Are you referring about the Internet gateways that are generally a part of a mesh proposal, or possibly that mesh networks don't scale like centralized networks do?
I'm saying that mesh networks are 1) not connectable on the scale of the Internet and 2) not scalable.
I think that we could create a truly distributed network. It would not be nearly as efficient as the Internet we are accustomed to, and it would not be based on the current Internet Protocol.
No offense but many DS papers have been published in this area and most aren't even positive in the theoretical sense, much less the practical. Crudely, I think it's time to put code where your mouth is ;)
Crudely, I think it's time to put code where your mouth is
Ah, you have me there. Actually, I am interested in this space and would appreciate a good academic overview. Maybe some links to these papers you mention?
