For the immediate future at least, federation seems much more feasible than wholly decentralized systems. At least for websites / applications that are expected to gain any mainstream traction.
Building a federated Wikipedia, Reddit, Twitter, Amazon, App Store, etc. that does not have a single point of censorship / failure is perfectly doable.
Hell, any of those websites could start federation initiatives today if they so wished (Waiting any day now for a Big Name to experiment with a federated model).
However from my experience p2p services (IPFS / Freenet etc.) still require a combination of technical skill and political awareness (to go out of one's way to use a service that is much less convenient -- and to develop for a platform that will make you no money) that the average user does not have.
Not to say they have no place: the internet itself was a fringe, nerds-only technology at one point. But it only became a mass phenomenon once uses were found that brought 10x improvement on things the nontechnical user already needed (sending a letter, looking up information).
Decentralized technologies (today) offer no such improvement.
And let's face it, for the average Joe, p2p offers nothing: a p2p service can at best be of equal quality as a centralized service. There's no way around it: there's nothing simpler and more performant than talking directly to a central host.
With federation, the user needn't even know the difference.
While it's clear that p2p is going to face an uphill battle to demonstrate its value to the average Joe, I would also argue that federation doesn't address the underlying issues with user control of data.
Email is federated. Does it give us any more control over our data? No, because one company controls the largest email service and monitors not just that user's emails but also the ones I send to them.
XMPP is federated. One company managed to get a large enough market share, at the first opportunity they lock other federation servers out and change the protocol.
Federation for established websites is long a thing, CDNs are taking care of that for them. Central points of contact combined with P2P distribution like WebTorrent is the worst of both worlds: The centralized site still owns all your data, and in addition you're paying with your bandwidth and CPU resources to reduce their bill.
A fully client-centric model isn't anywhere near likely to make it big in the near future. However, client-first (with personal servers provided by friends & non-profits) is the only feasible approach for reclaiming ownership of our data.
For-profit companies don't have any incentives to let you store large amounts of encrypted data on their servers for free, at best they'll use it for cross-selling. At worst they'll push for a modified protocol version that relies on the central company server for crucial functionality, scheming the next lock-in.
If the past is any indication, the future looks bleak for idealists who (merely?) want a user-controlled alternative to those data-driven monopolist platform ecosystems.
The technology might be getting there, but the average user's priorities make it a super hard sell. In the face of the Google/Apple duopoly, open standards are more and more an optional gimmick - instead of standard data transfer protocols we get CarPlay and Dropbox, instead of wireless standards we get Chromecast and Apple Multipeer, instead of mini-computer smartphones we get constrained consumption devices with tracking built in as a core feature.
The value of client-first P2P is really not about censorship or failure. It's about control. By giving control over protocols and platforms to a handful of large service providers, we got more convenience in exchange for our opportunities to tinker and improve on the status quo. Instead of being able to change things as a society, now we have to rely on shareholder-controlled entities to decide what's good for us. It used to be just code and Windows vs. Linux, now it's data and Google vs. basic assumptions of privacy.
Agree wholeheartedly. I'm a linux/GPL proselytizer among my friends. I've customized my development environment beyond recognition; I love it. That's what technology is to me: the ability to transform the world around me using my imagination. It's a beautiful, empowering feeling that I wish everyone could feel.
Every day I am saddened by enormous drift between what I know is possible with today's already existing technology (we cary supercomputers in our pockets!) and what we are actually doing with it (supercomputers we use as neutered mass-media consumption and ad-delivery devices).
I also know a lot of non technical people. Even young people, smart professional people, politically aware people; just not technical. And to 99% of the public, control is even more of a non-issue than privacy / censorship / security. "This glowing rectangle is a black box, why should I care if company A locks it down with binary blobs or if company B releases it under some GP-I-dunno-what?"
Even technology companies choose slack over improving and extending open standards. Convenience is a bitch.
Sadly, the only selling point I see today for p2p is piracy. The only issue that will get people riled up is when they lose free access to their favorite TV show. Which only further stigmatizes the hacker mindset as a front for would-be criminals. Privacy is somewhat making its way to the general population, but the message is muddled and confusing to most.
I wish I had a solution. How do you sell the inherent beauty of open, hackable technology to non-geeks? Selling it on its social merit alone seems too abstract.
Building a federated Wikipedia, Reddit, Twitter, Amazon, App Store, etc. that does not have a single point of censorship / failure is perfectly doable.
Hell, any of those websites could start federation initiatives today if they so wished (Waiting any day now for a Big Name to experiment with a federated model).
However from my experience p2p services (IPFS / Freenet etc.) still require a combination of technical skill and political awareness (to go out of one's way to use a service that is much less convenient -- and to develop for a platform that will make you no money) that the average user does not have.
Not to say they have no place: the internet itself was a fringe, nerds-only technology at one point. But it only became a mass phenomenon once uses were found that brought 10x improvement on things the nontechnical user already needed (sending a letter, looking up information).
Decentralized technologies (today) offer no such improvement.
And let's face it, for the average Joe, p2p offers nothing: a p2p service can at best be of equal quality as a centralized service. There's no way around it: there's nothing simpler and more performant than talking directly to a central host.
With federation, the user needn't even know the difference.