Right now, if a dozen people around a block watch the same viral video, all of them will fetch it through the backbone, which is obviously wasteful. If they want to store the video for later, they will have a dozen copies wasting space on their drives, which could again be reduced by more efficient algorithms (eg. erasure codes).
There are lots of unused resources at endpoints (local network links, unused storage space and computing power on most consumer devices, etc) which could be utilized more efficiently. If we can pay for this wasteful centralized approach, why would paying for a decentralized network be more difficult?
And make no mistake, we are all paying for this "free" centralized network now. Whenever you buy a product from a company that advertises on facebook, google, whatever, you are effectively paying for it.
> A decentralized web would be vastly more efficient in terms of resources. A lot of the data could be deduplicated and a lot of backbone traffic could be avoided.
Decentralization usually brings more duplication of data. Think of email: copies of the same email are redundantly stored in many servers. And this is what we want. I have received the email, I decide if and when to delete it. Or Bitcoin. It is "decentralized", but if you want to run a normal client an be part of the federation you will need to spare quite a bit of disk space to store the exact same 90GB that comprise the blockchain.
> Right now, if a dozen people around a block watch the same viral video, all of them will fetch it through the backbone, which is obviously wasteful. If they want to store the video for later, they will have a dozen copies wasting space on their drives, which could again be reduced by more efficient algorithms (eg. erasure codes).
At the same time we do not want people to know that we are watching this video. We can get that video from our neighbors only if the ID of the video leaks somehow (if you reply "TOR", I will answer "exit nodes know"). But that is just a side problem.
The main problem is dynamic content. Almost all my open browser windows contain pages that are generated "just for me" (github, gitlab, travis, HN, a work-related 2TB ownclound installation, amazon, google search, ...). These are the pages I care about, and, sadly, these are the pages that are most difficult to decentralize in acceptable ways. Actually some of them are meant to be centralized (amazon or google search, for example).
I'd love to see more decentralized or semi-decentralized services (planets and their OPML files are great, for instance), but one has to admit that some problems just ask for centralization.
PS: weird, I had to write "google search" instead of "google". Using just the word "google" felt somehow wrong. Brand dilution, I suppose.
While this is true this could be vastly reduced if the internet supported source-specific IP multicast (SSM). That way even a small box in someone's basement hosting a single copy of some niche content could serve thousands worldwide.
Alas, multicast is hard and incentives are not in favor of making it available to the end-user even in limited scopes.
Github/Lab and HN actually seem pretty easy to decentralize - git already is, and the rest are essentially feeds, much like email/newsgroups/RSS/FMS. It's mostly a matter of running the aggregation code in the client, rather than on the server. Google can also be decentralized and federated - see YaCy. Amazon, there wouldn't be much point, since the service would still be contracted with a single party.
It's not a trivial job (in the "I could do that in a weekend"), but it doesn't seem architecturally hard considering the existing distributed platforms.
The current physical topology of the Internet is already heavily biased towards backbone-to-edge connectivity, and is directionally asymmetric.
By comparison, edge-to-edge connectivity that doesn't go through any backbone is much more limited, and edge nodes often don't have the upload bandwidth to effectively deliver content.
Plus, CDNs and other distribution networks already effectively put easily cachable content closer to the edges to relieve backbone congestion.
For example, I live in an apartment and the ISP has a router installed in the building. Coax cables are running to each apartment. I can see 15+ wifi networks from my home. Yet to fetch the exact same front page of a news site my neighbor is reading and has a cached copy of, I am routed through the backbone probably half way around the world, or to a CDN at the other end of my country at best.
I don't think photons really care which way they travel but the shortest the path the faster they will get there and the less photons we send the cheaper it ends up being.
But I'm not so sure about the thesis that it's just a software issue. It seems wasteful to sling electrons through a backbone and back when those bits are (physically) close by, but the cost here is not the distance those bits traveled, but the marginal cost of the physical hardware to deliver those bits to your doorstep.
It's probably much cheaper, on a cost-per-bandwidth basis, to beef up the existing backbone, because significant economies of scale exist.
OTOH, there's a significant fixed cost every time you try to widen edge links (it's very expensive to dig up and replace residential cables).
Couldn't agree more. The internet as we have it today is set up for decentralized consumers and centralized producers.
There is one untapped benefit of the current system which is that there is low latency on the edges. I'm not sure what the killer app to make use of that would be though.
Not only is no one economically motivated to share traffic towards the edge, no one will be building physical network capacity for it either.
I wonder what projected bandwidth numbers look like (if, for instance, youtube or netflix was completely distributed) - is the edge network close to being able to handling this?
In my opinion, it's a bit more nuanced than that. When you purchase a product with cash, no personal data is exchanged. The transaction is final and complete. When you 'pay' for services by exchanging personal data and metadata about yourself, that transaction is persistent and the vendor continues collecting payment indefinitely.
With store loyalty cards and even in-store Wi-Fi tracking and other tactics, even cash payments can be viewed as less anonymous, because the vendors prefer to continue to reap the rewards of profiling and tracking customers. My point is that this isn't a black and white issue, but there are some important distinctions to make about payments in cash and payments in data, or a mixture of both.
Or in social media's case -- we are the product that they make money off of.
My friends and I have been running our own network for a bit and playing with a very grassroots version of the web. Those of us that work in software host and create services to distribute music, movies, etc. that we collect. We also process our own IoT sensors and devices. It's maintained by those of us in software, but we have friends not in software accessing the network and using its resources. It's a lot of fun!
Diaspora has this functionality:
So as enforcement got better and legal alternatives emerged, their 'popularity' declined. There are also people who share Netflix logins, which is a (grey-market) option for those with no money, but is vastly less likely to result in DMCA notices mailed out to you.
If anything, the various decentralized networks of the 1990-2000s were the products of organic demand. There is simply less demand now than there was, or that demand is being suppressed by other mechanisms (such as laws, in the case of copyright-infringing sharing).
Perhaps there's just not a lot of demand for hosting your own identity in some decentralized, federated social network.
As I've pointed out before, all the enthusiasm for micropayments comes from people who want to collect them. There's no vast grass-roots demand for the ability to send pennies.
Actually, we do need a more decentralized web for these use cases:
- I want to send you a big file without going through a server.
- I want to send you a text message without going through a server.
- I want to talk to my home control system from my phone without going through a server.
Technically, we need to fix it so that you can reliably get an Internet connection between any two nodes that want to talk. For starters, all mobile devices and networks need to speak IPv6, so everybody has an address.
David Brin tried pitching a new micropayments concept with repudiation as magick sauce. No.
Of course the loudest mouths will be those looking to profit off micropayments, which then defeats the point of the system, which is why the only successful ones are the ones that start without the selfish motive to begin with.
ipv6 is not going to be a catch all answer. Despite having more addresses than number of computers we could feasibly build ever, rollouts of ipv6 are still allocating tiny blocks to home networks that don't provide enough addresses for guaranteed public IP addressing between all nodes. If Comcast only gives you 8 ipv6 addresses on your home router, your 10th device is being natted again and even on ipv6 you cannot assume a working p2p endpoint.
So far, no established solution exists for doing this. The most promising projects working on this problem that I know of are IPFS and Ethereum Swarm but both are still struggling with the low level protocol that makes sure person A gets paid for hosting files from person B.
IIRC the problem was more that there was an awful lot of malware + the RIAA and MAFIAA.
Which suggests a solution -- keep several of copies of everything (or copy-equivalents with erasure coding) and then download the one(s) with the cheapest/free bandwidth.
People care if you run down their battery or use up their cellular data. Not so much when the device is plugged into AC power and connected to unlimited broadband/wifi. At any given time there should be enough such devices that you don't have to make requests of any of the expensive ones.
Add to that the ability of people who don't want to worry about it (and don't want anyone they know to worry about it) to lay out a few hundred bucks once for a 10TB NAS and leave it connected to their broadband connection, and it feels like an actual solution.
http://gun.js.org/ and we have a storage layer that plugs into S3 that can handle 100M+ records a day for $10 (about 100GB+ of data) https://www.youtube.com/watch?v=x_WqBuEA7s8 .
The goal is that P2P systems are "economies of scales", so the price point is orders of magnitude less than their centralized alternatives. This disruption on price is how we pay for the decentralized web.
The real problem is that developers who actually build the decentralized applications need to be paid somehow. How is that going to be organized? How does the money flow?
It's possible to have decentralised app store. Apps could be hosted by developers. Metadata that could also be hosted by a developer would then determine which app stores the app should be available on. Users searching for an app using the app store could then pick up on this metadata via crawling the decentralised web, as well as sharing lists of available apps with other users. As for payment, plenty of options there (including but not limited to cryptocurrencies like Bitcoin).
But I can imagine a kind of software or device that may become popular that uses/requires this, that will turn this into something that people will get too used to that may become something essential in the long run.
Disclosure: I'm the CEO
Also, just trying to be constructive here, but you guys really need to work on branding. The AMA went about as poorly as it possibly could have because you didn't have good presentation. Here's how you have to pitch the whole "bidding for URLs" thing; "There are permanent, long-form URLs like on YouTube. But people can bid on short-form URLs if they want to as well. You have to bid on them so squatters can't take them all over. If someone outbids you, you get your money back. We make no money from this." Boom, every inane reddit comment about "corporations will buy all the video links, man!" shut down.
To address this, we've written up this FAQ answer: https://lbry.io/faq/naming
We've also communicated the truth better inside of our community, but it is tough to get that message back out to everyone that saw the AMA.
Pay for storage by sharing the disks attached to the computer at home that is connected to the Internet. Run out of space? Buy another disk or reduce the space for sharing. Same of what happens with file sharing.
Pay for bandwidth through the bill of the ISP. This could be tricky because ISPs don't usually make it easy to have servers at home.
Both storage and bandwidth can be outsourced to server farms: a machine on AWS or anywhere else would be as good as one at home. Maybe more convenient, maybe less.
But if you're a big content provider you might end up with large disks and large pipes because you want that your content is always accessible and no random guy can affect its availability.
Much like the internet itself, this new architecture will be ushered in by hackers looking to disrupt the status quo of walled gardens and data warehouses, not corporations. So immediate monetization is not a top concern. We can't even imagine what the economics will look like, just like Tim Berners-Lee couldn't have imagined the current web 30 years ago. If we stop to worry about fitting the decentralized web into our current economic models, we will trap ourselves in a box.
If we just make a better algorithm, people won't find our UI design intolerable.
If we just offer them some privacy, they will pay real money for it.
Long before there was a web, people were giving up privacy and freedom for the convenience of credit cards.
If the decentralized web is going to fly, it needs to START with the social sciences.
We can all nerd out over algorithms that make it happen, or use math to prove that it can't be done, only once the economists have created a model in which it might, theoretically be done. And the social scientists have created a model in which a critical mass of people will make the necessary choices, when presented with them.
As it stands, most people are not willing to pay:
I would be willing to pay that _and more_ in a heartbeat if given the option. But, in social science terms - I understand the problems with centralization much better than most, I'm willing to put up with much worse UI than most, and I'm wealthier than average.
Heck I'd probably pay just because I think the underlying technology is neat! But that makes me a pretty extreme outlier on the broad spectrum of humanity that are choosing, en-masse to use Facebook, which means that I either need to use Facebook, or nothing.
If social scientists could actually predict what people would do, they would have been the ones to start every multi-billion-dollar "why is this so valuable?" startup like twitter, snapchat, etc.
Computer scientists are and will continue to be the driving force behind techno-social change. We have no idea what people actually want, but neither does anyone else, so we should just keep trying stuff and see what sticks.
But as you point out - computer scientists don't know what people want _either_, so.. predictive power itself is not necessary for a thing to work.
But _working_ isn't sufficient either. Facebook _works_. As I understand it is WILDLY profitable. But the way it works is a matter of economics and social sciences. Building a social networking website is a SOLVED PROBLEM in computer science. (yes, I realize they're doing some creative research on scaling it up).
Someone could clone Facebook tomorrow, and offer it for free, right? But.. could they charge $10/mo for it, and say "Facebook but without the spying and advertising?" - Probably not.
I'm not saying computer science is useless here. New algorithms WILL be needed. Bitcoin or any form of distributed ledger seems extremely unlikely to scale as needed, for instance. Perhaps provably unable to scale as needed - I'd welcome a paper proving this with solid math / computer science (and there probably is one, I don't keep up).
But what IS necessary is a working economic model that is not the current economic model. You're saying that nobody can predict the future (true), so just try a bunch of stuff - so far, so good. But then when you find something that works.. how do you pay for it? Nobody in computer science is trying to figure it out... that is a consideration "for later" - the point the article is making! What happens "later" when money needs to be made.. well.. advertising and surveillance. That is literally the only answer anyone has come up with.
That is why, FIRST, predictive or not, someone (probably in the social sciences) needs to come up with the idea(s) about how it _might_ be paid for, that don't include advertising.
The article was a response to the big decentralized web conference "web 3.0" that happened recently. It was all CS people, technologists who got together to enjoy talking about technology. But economics is what will shift this, and _innovation_ perhaps will be needed in economics. And other social sciences - nudging, gamification, etc.
(Wrt scaling, there is no theoretical reason the Bitcoin network couldn't scale to handle any number of transactions in a bounded spacial volume much larger than Earth. There are only surmountable practical limitations like processing power.)
And those are the most "first-world" currencies. People seem to use even more unstable currencies without issue.
I think when it comes to this kinds of money involved, it no longer the domain of social sciences but politics. Some ppl call politics social science but there's no science in it, it's more the inevitable dystopic future than anything else. As someone said on one of the conferences with audience of open source technologists: you can ignore politics as much as you like but the politics will not ignore you.
I mostly agree with this line of reasoning. Users don't care about decentralization, and for developers it's much easier and more lucrative to build a centrally-hosted version of X than a fully-decentralized version of X (for most values of X). Things aren't going to change until the decentralized Web not only needs a killer feature that cannot be had on the Web today, but also needs to preserve backwards-compatibility with the Web while being at least as good in what it can offer.
That said, instead of trying to re-educate users (which would be extremely costly and probably take generations), what if we simply made it so that decentralized applications that otherwise looked like normal Web applications were significantly less expensive to write and operate for developers? If we gave each user a client-side private key paired with globally-unique user-chosen name, then we could refactor most Web applications such that they stored their users' data (signed and encrypted) on the users' personal cloud storage providers (or whatever storage media the user desired). This is great for developers, since they no longer have to host the data themselves or defend it from hackers. This also simplifies a lot of application designs somewhat, since there no longer needs to be a backend in many cases. The user wouldn't even notice the transition, since (1) cloud storage is already pretty much free, and (2) cloud storage is also reasonably performant and highly available (and nothing stops the user from mirroring data across multiple providers). There would need to be some work on the Web browser for automatically managing keys and personas, but this is certainly doable.
I'm not a game theorist, but I think each participants' incentives are appropriately aligned to make this work:
* Cloud storage providers have to compete on space and availability and are incentivized to give users the best cost/performance trade-offs.
* Users are incentivized to keep their data available for as long as they want other users can get at it (and they can take it offline the minute they don't).
* Developers are incentivized to target this development model because it lets them externalize a non-trivial portion of their operating costs without affecting performance or features.
* This design is incrementally deployable--an app can still have some centralized components (like a search index) and we get to re-use as much existing infrastructure as we'd like.
Would love to hear your (or others') thoughts.
Not all incentives have an economic nature, some are political, and I think the latter applies in a greater meassure in the current case.
Just the realization that you give too much power to those running centralized services, may represent the required incentive for people to switch.
The overall material cost to run a centralized vs decentralized service may be comparable, but the political cost is not..
Really, someone tell me what is centralized about the web right now and we can figure out together that this thing actually IS decentralized, it's just that money forces people, companies and govs to work together and keep more of the power in fewer hands.
They do. That's pretty much what bittorrent incentivizes.
Credit systems to prioritize peers who have recently shared with us and promotion of long-running nodes to ultrapeer could really help.
Wow was I surprised to see this on Hacker News! It's my view that decentralized systems will almost always be less efficient than centralized systems. This is because centralized systems can organize information in a more efficient manner and use specialized hardware. Decentralized systems give you privacy, redundancy, anonymity and other benefits at a significant additional cost.
Bitcoin has the potential to provide a mechanism for decentralized systems to set prices and allocate resources. We'll see how this plays out over time.
1. Decentralization won in 1970 - 2005 protocol design because reliable computers were very expensive. It was economically advantageous to have a decentralized network of unreliable computers than centralized unreliable protocols. Reliable compute has now become commoditized by the cloud services. This creates a pressure for centralization.
2. Decentralized protocols have not aged well(TCP/IP, SMTP), have deep security flaws and have proven difficult to upgrade. This has created new centralization pressures to cope with the security flaws(Gmail, Cloudflare)
3. There continue to be pressures for decentralization related to both politics and security pressures. Tor and blockchain cryptocurrencies are excellent examples of these. But decentralization is unlikely to extend all the way to edge devices and instead operate in a federated fashion.
IP allows a maximum of 255 hops for any packet. This inherently restricts the topology of the Internet: As it stands, it can never be a world-wide decentralized mesh. Instead, you end up with large hubs and choke-points.
The IP addressing scheme also makes it very difficult to have a mesh: IP addresses are assigned hierarchically.
The name "Inter-Net" describes the problem directly: The Internet isn't a global network that just anyone can contribute or connect to; instead, the Internet is just a protocol for inter-connecting the world's centrally owned and operated networks.
With the IP Internet forced topology, Economies-of-Scale make massive centralized services cheaper than distributed services (even if similarly massive).
The obvious result: Comcast is your only ISP at home.
Disclosure: I've been working in my spare time for years on a solution, and I think an isochronous source-routed stream-based protocol is the only solution. I've got a proposed spec at IsoGrid.org
TCP/IP and Ethernet were decentralized alternatives both from a design and IP licensing point of view. But the unresolved technical debt at the transport layer is major centralizing force on the Internet right now.
I wish them well, it's good to have lots of people working on this from all angles.
Wrong, the reason centralization happened was profit. You can’t make profit with a decentralized protocol. And that’s exactly why we need more of them.
Companies trying to turn the internet into a for-profit thing are the worst possible outcome.
1) fix the security: https://en.wikipedia.org/wiki/Named_data_networking
2) pay by sidechain: https://bitcoinmagazine.com/articles/greg-maxwell-lightning-...
3) pay players to copy: https://www.wired.com/1994/09/superdis/
The biggest problem, though, is having an incentive to promote the technology and push it to users. Both bittorrent and bitcoin have that incentive, that's why they are successful.
If you're not paying for something, you're the thing being sold.
It may be we end up paying central IT orgs small amounts of money to host apps (I can see people paying AWS for social network applications of their own running on Lambda, for example), but the network pays for itself by users paying for it.
If users are not prepared to do that, they are always going to be sold as commodities, as they are today in most free SaaS consumer-facing applications.
Anti-virus software shows how you motivate non-technical consumers to buy - in cynical terms you basically just scare people enough until ready to pay.
But decentralized webs need not resemble the www. Decentralized webs are radically different.
For me a big part of decentralized webs is people using computers in person, in real life spaces, to share and interact with information. Much of the information will be transmitted in very shallow ways, via home routers and built in wireless hardware, over local networks. People sharing personal files, photos, videos, work documents and using decentralized web applications, together in person or over small, local networks.
The decentralized web could be a huge boon to local economies. For example, imagine a band that plays monthly at a local theater. When you go to a show, they release a new song over the local network, via p2p sharing protocols. Assume that people like the band, and that this is a world where people are familiar with using these kinds of applications. The experience of getting new content will provide additional incentive for people to come to the show. If it is marketed well, the band and the theater will make money from increased turnout. I think this could be huge.
The real question being asked here is: who pays for the world wide web if people aren't relying on it as much. Decentralized webs need not be world wide, though pockets of decentralized webs may interact with worldwide webs, they will not be inherently dependent on it, and will put draw far fewer of its resources.
And in this sense, decentralized webs are paid for by their users when users buy hardware. Much more than users pay for centralized webs.
Maybe there is an argument to be made that software application industries, like the web, somehow incentivize hardware manufacturing. This is probably the case to a degree. For decentralized hardware applications to carry the give people the same incentive to buy hardware as centralized applications do, they are going to need to be dependable and usable by all different kinds of people. If decentralized web applications are an uphill battle for people, then users won't buy hardware with the intention of using it in a decentralized way.
Decentralized networks are a cheaper way to do networking, of course it's not as fast and simple to use, but the big advantage of P2P is that the user pays for the hardware. If your software can work on user hardware (which is a lot of computers), it's a huge economic and infrastructure advantage.
The point is not only to give user the control over their data, it's also to reduce the barrier of having to pay to make a service, since it can also require knowledge and maintenance to build a service.
P2P removes the duality of the internet. The internet is not just a "series of tubes", it's also a large quantity of expensive server rooms and high grade hardware that must sustain large throughput. To me, p2p is the real "web". If today you remove gmail, youtube, google, facebook and twitter, the internet dies because users don't know what to do anything anymore.
Granted, it's not as simple and the software is complex, but if companies can do horizontal scaling in server rooms, it should also be doable with user computers.
A lot of these ideas limit the ability of companies like facebook, google, etc to earn money from a service and/or ignore ecommerce, etc. Take federated systems, for example.. how does a federated facebook earn money?
And while you and I might not care if facebook earns anything, it's going to be tough for existing stakeholders to back it or anything like it if the idea would result in the end of their business.
Similarly, the same limitations apply to startsups.. meaning todays startups, who eventually become tomorrows large companies, aren't going to be based around something that can't be used to generate revenue.
This isn't the 1980s anymore. There are too many users, placing too many demands on systems for it to ever be free.
So to answer the question: I think the distributed web is paid for just like today's web.. because any distributed system that succeeds, will need to be able to support the current methods of generating revenue from the internet.
it's going to be tough for existing stakeholders to back it or anything like it if the idea would result in the end of their business.
The desire for a decentralized and encrypted internet grows with every new data-cap by ISP's and every illegal surveillance scandal carried out by governments. The people find no recourse in their established authorities and will seek to their own devices for security.
Generating revenue is an after-thought of freed men.
It's like evolution: things in society (inventions, ideas, protocols, whatever) exist because they popped up and then survived. One way to survive is to provide a revenue stream for profit seekers who will look after you in return.
Many thing work that way, but many things, such as human languages, and the TCP/IP protocol survive in other ways. Of course all these things do profit people, and are kept around as a result, but through mechanisms quite different to the ones you are talking about.
Here is an example. We'll assume an imaginary decentralized communication platform. Alice wants to chat with Baozhai. Alice speaks English, Baozhai speaks Mandarin. They connect and chat (in their own language). Each capability here, identity, presence, and translation is provided by service providers. Alice and Baozhai have paid for these services.
We might need to invent an easy and secure way to pay first though :)
Edit: and a comment here gave me another idea: offline syscs to friends when your PC is online (or syncs from friends)
Some energy should be spend on making people want a decentralized web. Better privacy might be nice but even Snowden's revelations couldn't change people's mind.
So besides privacy, what additional value would a decentralized web bring to the average person?
In essence users aren't part of the network the way servers are they are simply clients connected through a ISP. That's because routing is a difficult thing to do well without a planned structure.
It's hard enough to route between normal home wireless routers. Could you imagine the complexity involved in routing between wireless devices like Smartphones that physically move around, turn off all the time?
What are you referring to? In a hardware sense, the Internet is decentralized to a large degree. I think you mean that it's Federated, as there is a large degree of coordination within ISPs, but a fairly small degree of coordination (outside of a number of peering stations) across ISPs.
Routing is difficult, but it's not as if there is some central authority managing all routing on the Internet. There are multiple redundant pathways for any given connection, and ISPs manage these things themselves using protocols like BGP.
I wouldn't say the physical architecture of the Internet is decentralized, but it's certainly not centralized.
Better yet, (and I realize there are a couple logical leaps here) these devices will progress The Internet into a mesh amongst neighbors and neighborhoods. And that's when the Decentralized Web will be realized: when a decentralized, resilient, adaptive Internet is available to host it on and that Internet isn't controlled by profiteering gluttons.
server less computing would in applications where this makes sense reduce the latency down to what is dictated by the last mile. on the other hand there is a huge push towards going mobile in many industries (thanks to LTE, LTE-A), but these wireless protocols do horribly poor routing P2P traffic. (this is a problem that operators are just about getting to grips with in their fixed networks).
When it comes to use-cases I think we have some realistic opportunities if an industry is big or desperate enough to also buy data from individuals (rather than a broker). I ran recruitment firm the last few years and a recruiter would always do anything to get their hands on a larger pool of data. If you give them a tiny new cloud service to search for "niche skill candidates" they will pay for it provided they can add additional people to their database. And a distressed candidates would sure rather sell their data directly to the company (earning a few cents along the way) rather than paying linkedin to host their profile and suck up all data.
the bigger challenge (than creating a p2p network) is how to structure services on top, the article mentions btc miners as the profit model for bitcoin. it is for me perfectly logical to have a company with a centralized service/buisness model that builds data applications on top of the decentralized p2p paradigm (most trackers are not really decentralized and the total decentralized bootstrapping of the p2p network itself is also a problem which isn't totally solved afaik).
What doesn't work well right now is supporting development for replacements of services like Google Now, Facebook, etc.
Very popular nodes on distributed systems are essentially just another form of centralization. But portability of data creates a free market, low switching costs, and a more consumer-centric focus on retention.
Groove by the author of Lotus Notes was the same, until M$ bought it.
Don't you see a pattern here? In the real world, you can start a micronation, but it will need the protection of a centralized state pretty soon. Thats why libertopias don't exist anywhere on the Earth, after all this time.
However, online, a lot has already been decentralized. Email is decentralized. The web is decentralized. In practice, we just need to make better, more user friendly servers. Why host your video on YouTube where it can be taken down?
The decentralized web has simple economics: EACH COMMUNITY PAYS FOR ITS OWN HOSTING. They can also have arrangements to have (encrypted, redundant) backup mirror hosting for each other.
What people get wrong is that the natural unit for the decentralized web is an ORGANIZATION and not an individual. Domains correspond to organizations.
One possible solution is the creation of a new type of corporate infrastructure entity that is disconnected from our current financial models. It should prevent outside investment or intent to create profit to keep those models from infecting it.
Also, I would point out that infrastructure can be centralized, decentralized and distributed. Whatever we end up with will likely be a combination of all three.