3 ms·

I've been looking for something like this quote from the original World Bank report: “The economics of the internet favor natural monopolies, the absence of a

I've been looking for something like this quote from the original World Bank report: “The economics of the internet favor natural monopolies, the absence of a competitive business environment can result in more concentrated markets, benefiting incumbent firms. Not surprisingly, the better educated, well connected, and more capable have received most of the benefits – circumscribing the gains from the digital revolution.” My assertion is that this is actually due to the nature of the TCP/IP protocol. IP addresses are assigned hierarchically and packets have a hop limit. Together, this makes a mesh network impossible. The TCP/IP Internet is an engine that centralizes power. I'm working on a side-project to create a (free and open) scalable (wired) mesh network protocol. I'm really glad I saw this, it'll make a perfect citation for my requirements doc. Anyone disagree with the above? I'm looking to try to strengthen my arguments.

Flammy 11y ago
I feel like its going too far to point at TCP/IP as a or the root issue behind growing inequality (especially inequality between 3rd world countries and 1st world countries). To look at the larger historical perspective, we're still headed in the right direction. How much did it cost to copy a book by hand? How much did it cost to print a copy with Gutenberg style printer? How much does it cost to distribute a book's worth if internet in the day and age of the Pony Express? Even within the digital age, we have a lot of alternative methods when issues arise... just look at the growth of community owned ISPs or network end points. And things like P2P transfers, Tor network, or the ad-hoc, informal cuban data sharing network. My point is other factors seem to be significantly higher impact than cost to distribute, or technical limitations behind the access to information. Devices and access are getting cheaper (despite what most of those monopolies would wish) and with that comes expansion. Will pure access to the internet solve all of the problems? Of course not, but I think we'd all agree it will help. And if TCP/IP isn't cost effective (or has another issue) in serving the 3rd world, something else will pop up to fill the gap.
coldnebo 11y ago
I'll bite. Regarding your assertion: I'm not sure that follows. TCP/IP's primary design point was avoiding centralized network failure in a nuclear war, so you can't strongly claim that it is responsible for heirarchies (i.e. Star) we see in modern topology -- that's likely due to other forces like cost optimization (eg branching structure is a common feature of such optimizations). Now this isn't to say that TCP/IP is efficient for all decentralized networks. And I think there are quite a few points that make mesh networks challenging... But you could also argue that for the interplanetary network protocols as well, so I think it's not so much the concept of decentralization, it's more the specific constraints of mesh or interplanetary that are unsuitable for TCP/IP. Regarding the quote: This argument seems to be partly based on economies of scale and how they are currently applied. But internet economies can be platforms as well as products. Platforms have a strong tendency towards monopolies (i.e., Apple App Store, Uber, VISA). But even in platforms, they may centralize wealth by treating customers as resources, or they may distribute wealth by treating them as partners. As we race towards the so-called "gig" economy, I think a key differentiator in platform providers will be the extent to which they grant power to their users as partners in creating value. This changes the classical idea of monopoly quite a bit as the old concept merely controlled a market for maximum exploitation of its consumers. If consumers are given back some power in terms of profit sharing, content creation, value creation, the Internet could be an extremely vitalizing force. You see this in spates: Etsy, Steam, Heroku. Maybe greater collaboration will be necessary to drive increasing efficiencies. The key point might be "central management" vs "empowerment" rather than platform, product, monopoly etc.

Flammy 11y ago

I feel like its going too far to point at TCP/IP as a or the root issue behind growing inequality (especially inequality between 3rd world countries and 1st world countries). To look at the larger historical perspective, we're still headed in the right direction. How much did it cost to copy a book by hand? How much did it cost to print a copy with Gutenberg style printer? How much does it cost to distribute a book's worth if internet in the day and age of the Pony Express? Even within the digital age, we have a lot of alternative methods when issues arise... just look at the growth of community owned ISPs or network end points. And things like P2P transfers, Tor network, or the ad-hoc, informal cuban data sharing network. My point is other factors seem to be significantly higher impact than cost to distribute, or technical limitations behind the access to information. Devices and access are getting cheaper (despite what most of those monopolies would wish) and with that comes expansion. Will pure access to the internet solve all of the problems? Of course not, but I think we'd all agree it will help. And if TCP/IP isn't cost effective (or has another issue) in serving the 3rd world, something else will pop up to fill the gap.

coldnebo 11y ago

I'll bite. Regarding your assertion: I'm not sure that follows. TCP/IP's primary design point was avoiding centralized network failure in a nuclear war, so you can't strongly claim that it is responsible for heirarchies (i.e. Star) we see in modern topology -- that's likely due to other forces like cost optimization (eg branching structure is a common feature of such optimizations). Now this isn't to say that TCP/IP is efficient for all decentralized networks. And I think there are quite a few points that make mesh networks challenging... But you could also argue that for the interplanetary network protocols as well, so I think it's not so much the concept of decentralization, it's more the specific constraints of mesh or interplanetary that are unsuitable for TCP/IP. Regarding the quote: This argument seems to be partly based on economies of scale and how they are currently applied. But internet economies can be platforms as well as products. Platforms have a strong tendency towards monopolies (i.e., Apple App Store, Uber, VISA). But even in platforms, they may centralize wealth by treating customers as resources, or they may distribute wealth by treating them as partners. As we race towards the so-called "gig" economy, I think a key differentiator in platform providers will be the extent to which they grant power to their users as partners in creating value. This changes the classical idea of monopoly quite a bit as the old concept merely controlled a market for maximum exploitation of its consumers. If consumers are given back some power in terms of profit sharing, content creation, value creation, the Internet could be an extremely vitalizing force. You see this in spates: Etsy, Steam, Heroku. Maybe greater collaboration will be necessary to drive increasing efficiencies. The key point might be "central management" vs "empowerment" rather than platform, product, monopoly etc.