'Largest Distributed Peer-To-Peer Grid' On Earth Laying Foundation For A Decentralized Internet (forbes.com) 80
Forbes reports on ThreeFold, an ambitious new "long-term project to rewire the internet in the image of its first incarnation: decentralized, unowned, accessible, free."
"We have 18,000 CPU cores and 90 million gigabytes, which is a lot of capacity," founder Kristof de Spiegeleer told me recently on the TechFirst podcast. "It's probably between five and ten times more than all of the capacity of all the blockchain projects together..."
"It's a movement," de Spiegeleer says about ThreeFold. "It's where we invite a lot of people to...basically help us to build a new internet. Now it sounds a little bit weird building a new internet. We're not trying to replace the cables... what we need help with is that we get more compute and storage capacity close to us." That would be a fundamentally different kind of internet: one we all collectively own rather than just one we all just use.
It requires a lot of different technology for backups and storage, for which ThreeFold is building a variety of related technologies: peer-to-peer technology to create the grid in the first place; storage, compute, and network technologies to enable distributed applications; and a self-healing layer bridging people and applications. Oh, and yes. There is a blockchain component: smart contracts for utilizing the grid and keeping a record of activities. "Farmers" (read: all of us) provide capacity and get micropayments for usage.
So instead of a Bitcoin scenario where some of the fastest computers in the world waste country-scale amounts of electricity doing arcane math to create an imaginary currency with dubious value (apologies, are my biases showing?) you have people providing actual tangible services for others in exchange for some degree of cryptocurrency reward. Which, in my (very) humble opinion, offers a lot more social utility...
ThreeFold and partners have invested more than $40 million in make it happen, de Spiegeleer says, and there are more than 30 partners working on the project or onboarding shortly. "So it's happening," he says.
In the interview, de Spiegeleer points out 80% of current internet capacity is owned by less than 20 companies, arguing on the podcast that "It really needs to be something like electricity.
"It needs to be everywhere and everyone needs to have access to it. It needs to be cost effective, it needs to be reliable, it needs to be independent..."
"It's a movement," de Spiegeleer says about ThreeFold. "It's where we invite a lot of people to...basically help us to build a new internet. Now it sounds a little bit weird building a new internet. We're not trying to replace the cables... what we need help with is that we get more compute and storage capacity close to us." That would be a fundamentally different kind of internet: one we all collectively own rather than just one we all just use.
It requires a lot of different technology for backups and storage, for which ThreeFold is building a variety of related technologies: peer-to-peer technology to create the grid in the first place; storage, compute, and network technologies to enable distributed applications; and a self-healing layer bridging people and applications. Oh, and yes. There is a blockchain component: smart contracts for utilizing the grid and keeping a record of activities. "Farmers" (read: all of us) provide capacity and get micropayments for usage.
So instead of a Bitcoin scenario where some of the fastest computers in the world waste country-scale amounts of electricity doing arcane math to create an imaginary currency with dubious value (apologies, are my biases showing?) you have people providing actual tangible services for others in exchange for some degree of cryptocurrency reward. Which, in my (very) humble opinion, offers a lot more social utility...
ThreeFold and partners have invested more than $40 million in make it happen, de Spiegeleer says, and there are more than 30 partners working on the project or onboarding shortly. "So it's happening," he says.
In the interview, de Spiegeleer points out 80% of current internet capacity is owned by less than 20 companies, arguing on the podcast that "It really needs to be something like electricity.
"It needs to be everywhere and everyone needs to have access to it. It needs to be cost effective, it needs to be reliable, it needs to be independent..."
And for legal defense? (Score:5, Funny)
Are all the proper backdoors in place for law enforcement's access?
Re: (Score:2)
"90 MILLION GIGABYTES!"
If only there were an SI unit for that number...
Re: And for legal defense? (Score:1)
Re: (Score:2)
Re: (Score:1)
Re: (Score:3)
Are all the proper backdoors in place for law enforcement's access?
Or, alternatively, some nice unavoidable block-chain audit trail built in that anybody can use, including law enforcement. You know, Internet-like, but finally with logging and monitoring as part of the system.
Re: (Score:1)
Exactly, if books are kept, even if by everybody, it's not "decentralized"
Re: (Score:2)
Exactly, if books are kept, even if by everybody, it's not "decentralized"
In a sense, yes. But calling it "decentralized" is a nice piece of misdirection intended for those that do pattern matching for terms instead of actually trying to understand things (i.e. the vast majority of people).
Re: (Score:1)
But calling it "decentralized" is a nice piece of misdirection...
Madison Avenue is always on the ball. The Decentralized Quantum Cloud® is better than gold
Re: (Score:2)
Indeed. Love that decentralized entanglement!
Re: (Score:1)
George Washington warned against entanglements in his farewell address. I think he stole the idea from Eisenhower's farewell address against the military industrial complex.
Re: (Score:1)
if you ever wanted to kill your ex, it will be there 100 years from now
Missing the point (Score:5, Insightful)
It's not the internet that is the problem - it was designed to be decentralized and resilient. The two problems are the last mile and the apps people use. The last mile infrastructure is a natural monopoly because the cost is high. Some apps, like search, don't scale well in a distributed form. Mostly though, they're just badly written and some have developed into a monopoly through proprietary barriers - e.g. facebook vs the interoperable distributed versions that are the way social media should be implemented - this is how email has always been, and how the web is. The lack of decent apps is perhaps what threefold is trying to fix (it's hard to tell through the buzzwords), but https://xkcd.com/927/ [xkcd.com]
Re: (Score:1)
Re: (Score:2)
A major point was the number of routes. That was a problem going back to the early 90s in that the routers just couldn't cope with the routing tables if most sites were dual homed. Even today getting a v4 /24 routed won't happen in many places assuming there is even free address space.
Re: (Score:2)
OK, but remember what spam did to Usenet. Better filters are needed. I've heard that 90% of spam never makes it to the end users, because it gets filtered out along the way. So this system would need end user filters that filter out that 90%.
Re: (Score:2)
Communist from Google
I do not think that word means what you think that it means . . .
Re: (Score:2)
Instead of Facebook and its insidious censorship, we can run our on little Apache web server.
We could run one, until they decided we couldn't.
Every time a commie uses the "muh private company" excuse to censor ideas that go against their programming, they follow up with "if you don't like the rules, build your own".
Then, inevitably, someone goes and builds their own and once they reach a few people, the commies come in to DDoS the site and if it's a message board of some kind, spam it to hell with illegal pornography and swastikas then call their ISP and get them terminated.
Then, if you want it to
Re: (Score:2)
Instead of Facebook and its insidious censorship, we can run our on little Apache web server.
We could run one, until they decided we couldn't.
Every time a commie uses the "muh private company" excuse to censor ideas that go against their programming, they follow up with "if you don't like the rules, build your own".
How do you maintain the cognitive dissonance necessary to call capitalist CEOs and executives of some of the riches for-profit companies on the face of the earth communist? "muh private company", "if you don't like the rules, build your own" are the most conventional statements of neo-lib philosophy I could possibly imagine.
Re: (Score:2)
capitalist CEOs and executives of some of the riches for-profit companies
If you're not aware that communism and capitalism can coexist quite happily, let me introduce you to an obscure country called China.
The point was that even if you go out and build your own service, if you do not conform to the dictates of the government or the big, rich, "capitalist" corporations, they can shut you out of Internet and banking services. "Good!" you might say, if this insidious form of corporatocracy suits your ideology. Just remember how much you loved it when it comes gunning for you.
Not the worst idea (Score:5, Interesting)
It's not the worst idea in the world, but the problem will be getting people to adopt it.
As for "donate services, get paid" kind of issues in a P2P system, that just invites the same problem that ride sharing (Uber) and airBNB cause. Some enterprising people decide to use their own/others property as a kind of unlicensed node on this P2P network. We see it in the present, and we've seen it in the past going back to the early 90's. DDoS's wouldn't be possible if machines were secured, and the sheer amount of wasted computational bandwidth that gets wasted in a DDoS could instead be redirected to unauthorized storage and transmission nodes and the malicious actors controlling the botnets would be PAID for doing so.
So uh, no thanks, think bigger. Decentalized is good, but putting an incentive for people to actually contribute to this P2P system by buying physical devices that are maintained automatically to just plug into their fiber gateway, on an IP address not routed to the user's LAN, and connect via a VPN system to any nearby transmission node on the ISP before doing anything. ISP's can then opt-in by providing their own transmission gateways in cities to save their own costs instead of piggybacking off customers.
You'd end up with a near ideal situation, eventually where end users connect to ISP-owned/maintained gateway nodes that are completely anonymous (eg all the gateway nodes can see each other, but not what is transmitted between each other.) Storage nodes can be located anywhere in the world, and realtime services (eg VoIP, TVoIP, gaming, etc) would automatically setup and tear down multicast/unicast channels to transmission nodes if anyone requests the stream and others would benefit from replicating the stream. That's where the real bandwidth can be saved. Storage nodes for VOD, images, text just need to store hashes of their video, audio, text (eg subtitles) so that content only needs to be downloaded to the storage node close by, and won't disappear unless metadata in the originating node suggests the content should be disappeared after a set time. (eg for temporary data with no real value, or data that might be confiscated if physically known the contents of.)
Re: Not the worst idea (Score:1)
you can generate electricty so why is there a grid (Score:2)
yes legacy
secondly its VERY hard to do routing reliably and speedily it requires investment at the core...
your not going to get that unless you convince edge users to pay for it
John Jones
Re: (Score:3)
be redirected to unauthorized storage and transmission nodes
That already happens, a local company was appalled to find that the FTP server on their multi-function printer was being used to host a kiddie porn site.
I'm not sure I understand (Score:4, Interesting)
Is this something like Tor where everybody can help extend the network? I see a lot of big company names listed as backers... If it isn't user driven, won't this alternative to the internet that's in the hand of like 20 corporations just be another set of large corporations?
What benefit is that to the end-user? Exchanging the devil for beelzebub?
Re: (Score:1)
Re: (Score:2)
Are they trying to reproduce the plot of 'Bing Bang Theory'?
Re: (Score:2)
No, it's totally different: TOR is a honeypot run by the NSA.
Re: (Score:2)
It's Forbes so it's expected to be gibberish scams.
Re: (Score:2)
What benefit is that to the end-user?
None because it fundamentally doesn't understand the problem with the current internet and doesn't attempt to resolve it.
Problem with the internet: Facebook/Google owns x% of it and you can't do anything without accessing their servers at some point.
Their proposed solution: Provide CPU and storage space for Facebook/Google for free.
Ok stupid example, but the reality is the issue is fundamentally that the systems sitting *on* the internet are under corporate control. You can't change that by messing with the
Re: (Score:1)
Um, yes. If he really had all power that he could just steal all the bitcoins and retire.
Re: (Score:1)
If his claim of being 5-10 times the size of all blockchains (not just Bitcoin) is true, then I can foresee someone renting this monstrosity to make money with it. Well, fraudulently, but "money talks"; I've been leaning more towards the spiritual side than material, lately.
eg. You could steal all the bitcoins in the world in 24 hours.
Use (Score:2)
It needs to be everywhere and everyone needs to have access to it. It needs to be cost effective, it needs to be reliable, it needs to be independent..."
It needs to be of use to people also. If you're going off to build your own Internet, better add the blackjack and hookers as well...
Re: (Score:2)
And porn. The consumer-centered Internet relies on porn for its existence and always has.
Re: (Score:2)
That's more of an issue than you might expect. All these decentralised networks will attract, in development, certain groups: Those people who have a need to shift information around the world, but are also unable to access the more conventional and established means of doing so - either for legal reasons, or cost reasons. So when you set up your new decentralised network, you will be flooded with pornography distributors, criminal fraud, and copyright infringement on an unimaginable scale.
That's how the p2
Well the Internet itself is kinda decentralized (Score:2)
At least in areas where you have open Internet Exchanges.
The Internet itself doesn't need a lot of computing power. Where the power is needed is when you want to push data over copper wires or the air. That requires serious amounts of computation. Websites on the other hand need so little computation that people can easily waste insane amounts of it in inefficiency and it still works. After all when you want to be "cloud native" it doesn't make much sense when you only need a fraction of a Raspberry Pi.
Utopia - ain't gonna work (Score:2)
80% of current internet capacity is owned by less than 20 companies
This mirrors human societies, and it's bound to get worse. And just like any human endeavour, if you try to redress the situation and do things differently, it'll end up being even worse - or the idea won't even get off the ground - because humanity is stupid and fundamentally corrupt.
Nice idea though - utopic though it is.
Re: So What ? (Score:1)
This post was kinda interesting until the end. Reading it was like watching an interesting street performer who was doing a good job and then all of a sudden stopped and just lit himself on fire unexpectedly.
Re: (Score:2)
This post was kinda interesting until the end. Reading it was like watching an interesting street performer who was doing a good job and then all of a sudden stopped and just lit himself on fire unexpectedly.
Haha. Thank you for that simile.
Though I suppose the level of interesting would skyrocket once the immolation starts.
Re: (Score:2)
In the interview, de Spiegeleer points out 80% of current internet capacity is owned by less than 20 companies, arguing on the podcast that "It really needs to be something like electricity.
I don't think that he understands how the electrical grid works at all. Instead of 20 companies there are around 10.
Re: (Score:2)
Electrical grids are marvels of accounting as much as they are of technology. I can buy my electricity from many different companies, which charge different rates and tiers,and yet the electricity all comes from the same wires. At the generating end too, those companies buy their electricity from generating companies who pump it into the common grid - a country-spanning network, beating a steady rhythm fifty times each second. It all works so long as everyone agrees on the fiction that one company's electri
Re: (Score:3)
Nonsense, the Pareto "principle" is one of the few things more incorrectly named than Moore's "law" - it's not some inherent phenomena of nature as inequality enthusiasts would like us to believe, it was just an observation about Italian land ownership at one point in time, at least as fleeting and meaningless as CPU speed increases from the '70s-'90s. We can and have done better and there's zero reason to think we can't.
My model for such an internet (Score:2)
I think a truly decentralised internet needs to be a cross of bittorrent and freenet and git.
I am thinking of truly decentralised content delivery network with each node acting as part of the network storage. End nodes would be incentivised to provide larger storage capacities by making download speeds dependent on the amount of storage the node provides to the network.
The more popular content would then be mirrored more widely to increase download speeds for that content and to reduce resource contention.
T
Re: (Score:1)
ACTION TODAY (Score:1)
Re: ACTION TODAY (Score:1)
Right. I don't see how your post or the parent is any different from the early web, has everyone forgotten already?
If you create content, it's sharable.
Re: (Score:2)
I'd go for something similar if I were designing it, except I'd use IPFS for the static storage - it has a very elegant data structure that would be much better suited to browsing websites than bittorrent is. The actual protocol kind of sucks, but the data structure is great.
Am i stupid? (Score:1)
Even Better (Score:1)
18k cores isn't much, actually (Score:2)
These days, that's about the equivalent of 500 servers, or about what would fit in about a dozen racks in a datacenter. Interesting, no doubt, but an insignificant amount of capacity to do anything of this sort.
Re: (Score:1)
And since there are entire data centers that do nothing but run Bitcoin on multi-cored GPUs his very first claim is bullshit.
Re: (Score:3)
GPUs are for Ethereum. You can't mine Bitcoin in anything less than single-purpose ASICs now.
This sounds exceptionally fishy (Score:2)
Probably just yet somebody else that tries to get rich on the blockchain. Maybe also an attempt to get an Internet with surveillance and monitoring already built-in.
Has anybody looked at the tech details? Would be interesting to see what they are hiding in there.
Cables are the problem (Score:5, Insightful)
"We're not trying to replace the cables... what we need help with is that we get more compute and storage capacity close to us."
The whole idea of the internet is that capacity doesn't have to be close.
The problem with the internet is the cables. Who owns them, where they do or don't put them, how much they charge.
Competition among services keeps them relatively honest. But there is no competition for who provides your connection in most places.
Re: (Score:2)
If you can get absolutely everything encrypted, then the power held by those who control the cables is greatly diminished.
Re: (Score:2)
If you can get absolutely everything encrypted, then the power held by those who control the cables is greatly diminished.
They can still unfairly throttle connections to services they don't control/don't approve of. I'd say it's only slightly diminished.
Re: (Score:2)
They can, but it becomes more difficult. If limited to mere IP blocking, and now that so many servers are shared infrastructure, blocking will often risk also inadvertently blocking other things.
Re: (Score:2)
True. If Verizon's block of rival VOIP operators blocks your Aunt's recipe site, well, that will certainly convince Verizon to remove the block.
Except (Score:2)
This will run into the same problem, eventually. It will become centralized into the hands of a few, regardless of origins. Sensing a pattern yet?
You hear that? Anarchy is calling!! (Score:1)
This is yet another plot for some people to take centre stage. Another Facebook, another Google, another you-name-it. And the lure is a promise of absolute freedom, perfection, superiority, and a happy world of anarchy with everlasting evasion from law and police.
This is only a super-highway to deliver more malware, porn, crime and other shit right into people's homes. I'll pass.
Something feels off. (Score:3)
I don't know... its the page, I think. It's very well done from a publicity perspective, but it lacks in details. It talks about how they replace 'four of the seven layers of the internet' even though their new stack has to run on top of the existing layers - plus the internet only has four layers, the OSI model only exists in textbooks now. Lots of social buzzwords. But where's the technical description? What makes the people behind this initiative think they will succeed where so many other projects have tried, only to die in obscurity?
Their network seems able to report real-time records of available resources. 20722 cores, but only 546 nodes. Average of 38 cores per node - and 154TB. Those are some mighty nodes - I'm not sure if it means the numbers are off, or that some of their corporate sponsors have been lending some serious hardware to bootstrap the project. Plus all the nodes report an uptime of between one and two months.
Some parts are dumbed-down to the point of making no sense. Their cryptocurrency, for example - they claim that new tokens are only ever generated when new capacity is connected to the grid - but when what happens when that capacity is lost? What's to stop someone joining, transfering their tokens to another account, disconnecting, wiping their drives and repeating?
I wish for a true decentralised internet as much as anyone - both for the technical advantages, and some nostalgia for the lawless days when there was more to the internet than facebook and twitter. But there have been so many projects promosing that - if I am to give this one any attention at all, I need to see proof of success. Give me demo apps. Give me an interface I can easily install and play with. Show me someone who is using it for genuine, serious work. Fold some proteins. Set up a web-gateway so that any curious person can access that demo app.
In short, give me proof!
blockchain? (Score:2)
Why does this need to involve blockchain? Does a peer-to-peer network need 'compute' resources or 'network' resources?
This sounds like a bunch of bullshit with a giant waste of power to complete buzzword bingo.
Re: (Score:1)
This dude didn't want the hassle of laying cable so he redefined "Internet capacity" as storage and CPUs. Then he convinced a bunch of suckers to install his SETI at Home clone.
It's a decent bet he's mining bitcoin.
Nice idea. Too bad it won't last (Score:1)
18000 cpu cores... (Score:1)
HPE marketing sh*t (Score:2)
https://wiki.threefold.io/#/be... [threefold.io]
Only the purchase of certified hardware will give you the option to become a certified farmer.
List of certified hardware: https://wiki.threefold.io/#/hp... [threefold.io]
Consisting of 100% HPE servers.
I've experimented heavily with (Score:2)
p2p software like IPFS, FreeNet, ZeroNet, etc.... to find a network that maintains privacy and security. This was about 3 or 4 years ago. In checking today I find fewer English speaking YT videos and Internet sites promoting these p2p technologies. Of the ones I've tested I'll mention two:
IPFS - I had up to 500+ connections to my computer while running IPFS. I allocated 20MB of disk space to each connection and 1/2 of my Internet bandwidth to the total. It slowed my laptop considerably. However, the