The Server Needs To Die To Save The Internet

D o we have the Internet we deserve? There’s an argument to say that yes, we absolutely do. Given web users’ general reluctance to pay for content. We are of course, paying. Just not with cold hard cash, but with our privacy — as digital business models rely on gathering and selling intel on their users to make the money to pay (the investors who paid) for the free service.

Users are also increasingly paying with time and attention, as more ad content — and more adverts masquerading as, infiltrating and degrading content — thrusts its way in front of our eyeballs in ever more insidious ways. Whether it’s repurposing our friends’ photos and endorsements to socially engineer selling us stuff, or resorting to other background tracking and targeting tricks to divert our attention from whatever it was we were actually trying to do online.

The commercialization of the web is the ugly reality of the hidden cost of all the datacenters and servers required to power the Internet. And that commercialization is compounded by the power of the big digital platforms that dominate the web we have today: Google, Facebook, Amazon. Increasingly we’re forced to play by their rules if we want to participate in the digital space where most of our friends are.

But perhaps there is another, far better way — that benefits individual web users and startup developers alike.

A UK startup called MaidSafe, based in the small town of Troon in Scotland, reckons the myriad problems with today’s Internet can be linked to a design quirk of its underlying architecture. And that the answer to solving the web’s most perennial issues, such as finding sustainable digital business models for content, safeguarding user data and privacy, and thwarting hacking, malware and overreaching surveillance, is to begin again — with a whole new re-architected Internet.

No one said this was a small problem with an easy fix. MaidSafe has actually been working on its new network since 2006, finally coming out of stealth earlier this year to begin detailing its grand plan. It’s now rolling out the first of three test networks, to test the underlying network without any apps on it as yet, ahead of a full beta launch in Q4. The initial test network comprises 180 nodes, located in Singapore, San Francisco, Amsterdam and New York.

So what exactly is this startup building? MaidSafe’s Nick Lambert summarizes the product as “a fully cross-platform, fully decentralized autonomous data and communications network”. What that means in practice is a network that does away with an intermediary layer of servers and datacenters — replacing that with peer-to-peer infrastructure. (No surprise then that MaidSafe counts Michael Jackson, the former COO of P2P pioneer Skype, as an advisor.)

Basically, the users of the network are also acting as the network infrastructure by donating a portion of their spare hard drive capacity — with built in incentives for them to do so in the form of a network specific cryptocurrency (called SafeCoin).

There are no other networks that combine being autonomous and serverless

So, in a similar way to Bitcoin mining being incentivized by the creation and distribution of new Bitcoins, users of the MaidSafe network will be compensated for the computing resource they contribute by earning SafeCoin. (Currently one SafeCoin is worth around 2 US cents but there was also a time when Bitcoin was worth as little — and MaidSafe obviously expects the value of SafeCoin to scale up as usage of the network scales.) It calls this resource donation process farming.

“What we’re building is software that connects together all the computers on the network to form — think of it as one giant computer, or effectively one giant cyber brain. So it really connects together all the nodes on the network and allows them to effectively become a very large datacenter, without of course the datacenter,” explains Lambert. “It’s a network infrastructure that will replace datacenters — and hopefully large technology companies.”

That’s right. This startup wants to reconfigure the current Internet hierarchy too — pulling the power and centre of gravity away from the owners of the biggest datacenters and platforms, and putting it back in the hands of individual users.

And individual developers too. Development costs for building an app on the MaidSafe network would be lower than via today’s hosted model — being as a startup wouldn’t need to pay for any hosting costs. AWS, Rackspace — not required! There’s also no upfront fee required for the privilege of developing on this network (it’s no Apple App Store). And the MaidSafe API is free.

Users of MaidSafe’s network contribute unused hard drive space, becoming the network’s nodes. It’s that pooling — or, hey, crowdsourcing — of many users’ spare computing resource that yields a connected storage layer that doesn’t need to centralize around dedicated datacenters. And so doesn’t need any middlemen serving data. The user directly accesses the network, and the network directly accesses the users’ computers. And that’s it.

“This is, we believe, to be the world’s first operating autonomous and serverless network that enables self authentication,” says Lambert. “It’s self managing and self healing… If data is lost through nodes going offline it recreates them. It’s able to be resistant to viruses as well… It’s utterly serverless. There are no other networks that combine being autonomous and serverless.

“There are other peer to peer networks that you could argue don’t require servers but if you want to join some of these programs at the end of the day you have to create an account to a lot of these services. And in order to create these accounts you have to go to a server and someone has to authenticate you onto the network… With this network self authentication is enabling the user to authenticate themselves onto the network.”

So, instead of paying for digital services with privacy, users on MaidSafe’s network pay with hard drive capacity they’re not even using. Which — frankly — sounds like a far fairer, more egalitarian ‘client/server’ relationship than the one we have now.

The only way they can actually stop the Safe network in a country is to shut off the whole Internet. It just can’t be shut down

Meanwhile, data being transmitted over MaidSafe’s network is encrypted locally then broken into fragments by its software and distributed randomly across the nodes so it’s stored in a massively decentralized way, thus — so the claim goes — thwarting hacking and snooping.

The problem with centralising data storage and transmission by using servers and datacenters is that data becomes inherently vulnerable, argues Lambert. Vulnerable to theft by hackers or other prying eyes — whether that’s corporates tracking us or governments snooping on us. Or regimes trying to control what we have access to. Ergo, all the more reason to throw the middleman away.

“MaidSafe doesn’t use DNS. Everything goes through routers in a thing called encrypted RUDP packets, so basically the packets go through the router but nothing can tell what’s inside it. And it doesn’t use DNS so they don’t know that it’s traffic for the MaidSafe network. So countries can’t shut MaidSafe down. So China couldn’t shut off the Safe network the way they could now, or Turkey couldn’t shut it off the way they shut off Twitter. So the only way they can actually stop the Safe network in a country is to shut off the whole Internet. It just can’t be shut down,” he says.

“The other interesting thing is the NSA could train all their resources on one router and they still wouldn’t be able to stop and detect MaidSafe network traffic.”

Lambert adds that the Internet we have today was also never designed to support so many users — making another argument for overhauling the underlying structure, on network resilience grounds.

“As it was originally designed the Internet was never meant to have 2.5 billion people on it. And that’s why it creaks at the seams sometimes. So I think what we’re doing is a kind of evolution — decentralization is a much more efficient way of doing this. And I think whether it’s MaidSafe or somebody else… someone will do it. I think it’s just an evolutionary step.”

“A lot of these large [digital technology] incumbents will not be overly happy with us but I think what we’re doing is natural evolution,” he adds.

How does MaidSafe ensure resilience with such a massively distributed infrastructure that it has no direct control over? “We keep a minimum of four copies of every chunk of data at all times. And the reason we do that is obviously people will turn their computers off and on, and people will have hard drive failures so what the network needs to know fairly quickly is does that piece of data still exist?” notes Lambert, likening the technology MaidSafe has built for this portion of the network to systems used by file-sharing sites like eMule — except, he says, it’s far, far faster. Because it has to be to make a viable network.

“Our network knows within 20 milliseconds if the status of a piece of data or a node has changed. It has to happen that fast because if you turn your computer off the network has to recreate that chunk on another node on the network to maintain four copies at all time.”

“We’re often maybe confused with a decentralized Dropbox,” he adds. “And that is maybe one very good use for our network… [because] it’s completely private, we don’t know who our users are, we don’t hold your cryptographic keys so we can’t access your data… But effectively MaidSafe is a data and communications network so you could put any service that you get on the Internet today onto the network — absolutely anything, YouTube, Facebook, Dropbox, basically everything.”

MaidSafe calls its network the Safe network — aka Secure Access For Everyone. (The MaidSafe name itself is short for ‘Massive Array Of Internet Disks — Safe Access For Everyone’. It’s also, according to Lambert, a play on RaidSafe — and indeed ‘Made Safe’.)

Lambert says the network needs an absolute minimum of 60 nodes to be viable, and likely a couple of thousand to “make it much more established”. “Which I don’t think is really too much when you consider there is 2.5 billion regular Internet users,” he adds.

So how is MaidSafe going to incentivize developers to build apps for the Safe network? That’s also built in to the design, via the SafeCoin cryptocurrency. Developers will be able to hardcode their SafeCoin wallet address into their applications — setting their own usage price (which can also be free if they like) — and then the network will pay them based on their app’s usage.

“There’s a built in revenue stream for them already. They don’t have to go down the advertizing route, or the support route if they don’t want to,” says Lambert.

MaidSafe has already had more than 650 developers registering their interest in the network — and it isn’t even beta launched yet.

With SafeCoins being distributed based on app usage, that allows for a network that rewards developers for building services that people actually use. Now just think about that for a minute. Apps that people love and use could receive payment just for being loved and used on this network — rather than their founders having to resort to stuffing ads in front of eyeballs, or selling user data off to marketing entities. That’s potentially a sea change in the kind of digital services that are possible.

Consider the problem content producers such as publishers have in getting people to pay for accessing their information. Again, this network could offer a way to earn money from readers without having to rely upon — or resort to — advertizing. This could very well be the micropayments dream that’s long been talked about and hoped for, but not yet executed effectively in practice.

From the user perspective, although there can technically be a cost for accessing the digital services on MaidSafe, if you’re acting as a node for the network then the money you’re using to pay for those services is money you’ve earned by being part of the network (i.e. SafeCoin). So there is effectively zero cost involved.

“The concept of free is an interesting one, if you have earned SafeCoins by using and providing space to the network then these SafeCoins are used on other services, no money (digital or otherwise) has left your pocket,” notes Lambert.

The user also has the option of donating storage space and just converting the SafeCoin they earn straight to fiat through an exchange. In other words they could put MaidSafe’s software on their computer purely to generate a revenue stream from being part of its crowdsourced, distributed datacenter-in-the-cloud.

SafeCoin is earned by Safe network users above what they are using on the network so spare capacity is a requirement. Users will also be able to access and use the network for free up to a certain capacity, according to Lambert — which will open it up to mobile users and mobile usage.

(And for cryptocurrency nerds, there is a finite number of SafeCoin that can be farmed, circa 4.3 billion, because of a technical limitation, however Lambert says MaidSafe may end up recycling a small percentage, say 2% to 3% per year, to avoid reaching a hard stop — in the ten or so years it takes for all the SafeCoin to be farmed.)

Despite open sourcing the network technology itself, MaidSafe is also a startup business — and has raised some $6 million over the years from private investors (including a big chunk from a crowdfunding sale in April) to fund development of the project — so it needs its own business model here too.

It’s looking at three areas for that. Firstly it takes a 5% cut of all SafeCoin generated on the network in order to maintain the core code. As the network scales up, that 5% cut scales with it. And if the value of SafeCoin rises it’s also earning a larger sized slice, since 5% of a bigger pie amounts to more pie.

MaidSafe’s team also plans to make apps itself to earn money on the network — it now has a team of around 16, skewed towards developers (C++ is the dev language for building on the Safe network). “We’ll be well placed to capitalize on the network because we know it better than anyone else,” says Lambert. “We will create money by making applications that are well used.”

In addition, MaidSafe has filed multiple patents — and will be looking at opportunities to license its technologies for use outside the Safe network. “We’ve got about ten granted patents, about 22 pending just now and more on the way. And some of these libraries and technologies that we’ve created can be used outside the network. So, for example, an existing Content Delivery Network… like Akamai they could use our routing and RUDP libraries to basically make their distributed servers run much more efficiently and much faster,” he says. “Anyone using our patents within the network is absolutely welcome to do so. The patents, just to point out, are purely defensive.”

Although MaidSafe is a for-profit startup business, there’s also clearly a huge determination to see the vision — the concept of a truly decentralized Internet — succeed and “change the world”, as Lambert frequently puts it. So it has set up a not-for-profit foundation that is the largest shareholder in its company. The crowdfunding raise — selling SafeCoin that will be redeemable when the network launches — was also a way to incentivize success across a wider group of supporters.

Plus, it’s working on decentralizing its own network knowledge (its own ‘corporate IP’ if you will) — and ultimately diluting its own paycheck by sharing that 5% cut with others — in order to increase the resilience of the network and boost the overall chances of success.

“We’re trying to decentralize the Internet, we’re also trying to decentralize the knowledge of the safe network so we’re trying to create global development pods which are not part of the company — these are little separate groups of people that exist in different parts of the world that will effectively start to contribute to the MaidSafe core codebase, and start to take some of that 5%,” he says.

“Our intention is to try and create these competing pods. And what that will mean is there’s no central point of weakness to the technology. And it will also make the network much better and much faster. These pods will hopefully start to become self sustaining. As they start to earn SafeCoins for their contributions to the underlying code and taking a portion of that 5% total. They’ll be able to start having a fund and potentially become a legal entity themselves if they choose to do so.”

Pods have been set up in San Francisco and Montreal so far, with Washington DC next in line.

Talk of pods sounds rather organic and that’s unlikely to be an accident given that the original inspiration for the Safe network came in part from one of the founders, David Irvine, being a fan of physicist Richard Feynman, and drawing on his advice to focus on the natural world when trying to solve a problem. That led Irvine to spend time studying ant colonies — as a model for an alternative, serverless connectivity topology.

“Ants are interesting because if you think of them like nodes on our network… they perform pretty much one function — but when you bring them all together they become this great powerful colony,” says Lambert. “So if you think about the nodes being ants, and each of those little nodes performing one function pretty well, actually when they come together you’ve got this amazing, powerful, safe and secure network that will — when it reaches critical mass — be more powerful than any supercomputer.”

So, while this small town startup has (mostly) been sequestered a world away from the big digital technology hubs where Internet giants are forged in bubbling development cauldrons liberally fed with VC-cash, you’d never guess that from the scale of its ambition.

“Hopefully the network will replace the entire Internet” is a phrase that rolls off Lambert’s tongue more than once during our conversation.

And while it may not have big VC backing, given its radical vision and open ethos — the network is being “given away” for free, as Lambert puts it, and once it’s up and running won’t be something MaidSafe has the ability to shut down — it’s not a startup that’s easy to ignore. This idea has the potential to turn the Internet as we know it on its head — for the better.

“The network is open source. People can fork it. If we start doing stupid things that people don’t like they can just fork the code from Github and just create their own. We’re very respectful that we’re kind of custodians of this network now, which is a fairly risky strategy given how much money and effort has gone into the network thus far, just to give it away — but I think it’s the only way that it will work,” he says.

Lambert jokingly refers to the company as the “oldest startup in history”, given its eight+ years of development, most of which were spent in stealth mode. So why has it taken so long to re-architect the entire Internet? When phrased like that it’s not really much of a question…

“Doing what we’re doing is exceedingly hard. Which is why it’s not really been done before. We’re different because we’ve set out to decentralize the Internet. And we’re also different because nothing like MaidSafe does exist. People will say that’s just like that or that but there is no other network in the world where you can privately log in to your own data, without anyone else knowing. And store data and share data — without the use of an intermediary, I’m just not aware of any network that you can do that on,” says Lambert.

“It probably requires different thought,” he adds. “Developers are trained from a very early stage that the infrastructure is: there is a server, there is a client, the client logs into the server and so on. And I think that has become utterly pervasive throughout the programming community. And I think why it’s not been done before is MaidSafe is such a different mindset.”

When Irvine was designing the Safe network he began by trying to figure out what the problem was with servers — and ended up with the realization that servers are the problem.

“The mindset has been how do we make servers better. Instead of saying servers are the problem let’s try and remove them.”

Now there is a thought.

[Image by Tristan Schmurr via Flickr]