Introduction and goodbye servers
The internet is not fit for purpose. If it was designed today from nothing, would it use the client-server model? Absolutely not. There have been far too many personal and sensitive online records stolen (1.1 billion in 2014 alone) for anyone to give the internet an architecture as vulnerable as the one it's saddled with.
So why not just change it? One company thinks it's cracked it with an open-source algorithm that crowdsources the internet, making data impossible to hack.
What's wrong with the internet?
It's broken. "Data is extremely hard to secure," says Nick Lambert, Chief Operating Officer at Troon, Scotland-based MaidSafe, who adds that the dominant business model of advertising and surveillance is, by its very nature, a centralising force. That'll be Google and Facebook et al, who give us free services in return for selling our information to advertisers. Privacy, once a non-issue, is increasingly important to us all.
"We're all putting a lot of value and content into the internet, but the benefit of all of this information is centralised around these enormous companies who are merely aggregating it, not creating it," says Lambert of the 'winner takes all' economic model.
Even the so-called 'sharing economy' innovators like Uber are getting more out of the system than their employees.
So what's the solution?
Decentralise it completely by crowdsourcing the cloud. That's the idea behind MaidSafe's SAFE (Secure Access For Everyone) Network. "What we're creating is an infrastructure – it's up to third parties to develop apps," says Lambert, before explaining how a Dropbox-style app would work using the SAFE Network.
"When a file is saved into a virtual file system, the network splits the file into chunks, randomises those chunks, and encrypts each one with bits of the other chunks, then disperses them to all the nodes on the network," explains Lambert.
We're not talking about new hardware here – the SAFE Network works on all existing cables, routers and switches. It all operates on the IP layer of the internet, from web services up through the application layer. Effectively it replaces web servers and data centres.
"The SAFE Network software passes that job down to the spare computing resources of all the users of the network," says Lambert. Users download the software, it connects all the users together, and since all the encrypted fragments of all files are distributed across all users' computers, data is secure precisely because it's unintelligible, and it doesn't reside in one place. In fact, no-one knows where it is – that's the beauty of it.
Goodbye servers
If the concept of a network that relies on encrypted and randomised fragments of ones and zeros distributed across millions of computers is hard to get your head around, it's a model that's already gaining traction elsewhere.
"We're doing a similar thing for data that Bitcoin is doing for trade," says Lambert. "There is no-one in the middle – nobody can be stopped in creating an account … MaidSafe has no idea who its users are."
This system of self-authentication means that when someone uploads a document, only they can reconstitute it with a PIN and password; there are no servers at all on this network. In the current model of the internet and the cloud, it's companies like Dropbox that hold the encryption keys for its users, not document creators.
What if a computer is switched off?
Security may be absolute, but end-points are still an issue – if someone switches off a computer, surely the SAFE Network concept falls apart? Actually, no. As well as everything being client-side, the data locations are constantly changing – there are a minimum number of four copies of all data on the network at any one time, and if one node goes off, within 20 milliseconds the network will relocate all of that encrypted data to a node it knows to be online.
That's complex, but the result is very strong security – instead of a fortress approach, data is dispersed, and constantly on the move.
Business model
A critical mass
Despite it using existing infrastructure and hardware, setting up a network like this is chicken and egg; you have to build the network, popularise it among developers, and have an app ecosystem all at the same time.
There's also the issue of speed. The network needs to reach a critical mass – at least a thousand users – before it becomes quick enough for online gaming. An alpha test of the SAFE Network is due in the next few months, which will bring real-world test results.
"We don't think latency will be an issue," says Lambert, pointing out that online gaming servers often need to be upgraded as the number of gamers grow beyond a certain point. "On this network the more users you have the faster it should become," he says.
However, in the early days, the SAFE Network will be vulnerable because it will be at its smallest – like BitCoin in its early days. "We need to ensure there's as much computing power behind the network as possible," says Lambert. "We may have to add computing power ourselves at first."
What's the business model?
Since it's all open-source (it's all on GitHub now), MaidSafe's business model is built around cryptocurrency – the MaidSafe coin.
"We had a crowd-sale a few years ago where we pre-sold 10% of all the coins that will be created on this network," says Lambert. "The company raised about $7 million (around £5 million, AU$10 million), which allowed the community to have tokens that allowed them to access the network."
The purpose of cryptocurrency, of course, is to incentivise the users to leave their computer on – a key element to the success of the entire concept. If someone provides 50GB on a laptop to the network – via some software – that laptop will then slowly begin to fill with encrypted fragments, and SafeCoin will be delivered into the laptop owner's virtual wallet.
Similarly, app developers can code their SafeCoin email address into the code of an app, and so get paid by the network, not by MaidSafe. "It's a way of transferring value around the network," says Lambert. "But we're just an infrastructure provider."
MaidSafe makes money by taking 5% of all SafeCoins that are created by the network, and by creating its own apps, the first of which are likely to be messaging or VoIP apps.
Is this the end for data centres?
"Within time," says Lambert, "there will be less of a need for data centres," though he admits that 40% of the internet's traffic is from content delivery networks like Netflix, which won't be affected.
"Another way to look at it is data centres could be contributors to the SAFE Network," says Lambert. It's also possible that custom-made, walled-garden versions of the SAFE Network could be used by banks and other security-conscious industries.
There's an environmental angle to this, too – data centres account for 3% of the world's electricity consumption, and since a lot of us are online and running our computers anyway, a decentralised internet ought to be more efficient.
The client-server model currently used by the internet is deeply flawed and inefficient, but the current trend to using giant public clouds – propped-up by data centres – is going to be difficult to reverse.
But should we as internet users aim for the internet, cloud and computing power to all be decentralised? Probably, yes, and the dream of an open-source, secure, crowdsourced internet now has a shape – all it needs to follow is a critical mass.
No comments:
Post a Comment