The Future of the Network – and the History of the Computer

NEXT Berlin has kicked off with two very different keynotes – but with linking themes. René Obermann – Deutsche Telecom René kicked off his talk by addressing the issue of post-digital head on. “The fact that everything is digital is no longer sexy,” he stated. And he’s right. People are interested in what they can do with digital technology, not with the technology itself. However, that’s creating a problem for the network operators, as demand for bandwidth scales ever upwards. As prices come down, traffic goes up. That means infrastructure is going to have to be built out fast.

NEXT Berlin has kicked off with two very different keynotes – but with linking themes.

René Obermann – Deutsche Telecom

René Obermann at NEXT Berlin

René kicked off his talk by addressing the issue of post-digital head on. “The fact that everything is digital is no longer sexy,” he stated. And he’s right. People are interested in what they can do with digital technology, not with the technology itself.

However, that’s creating a problem for the network operators, as demand for bandwidth scales ever upwards. As prices come down, traffic goes up. That means infrastructure is going to have to be built out fast. €2bn to €3bn is needed to invest in a fibre network to provide the bandwidth to support emerging services – which is tough to do as margins reduce. The regulatory regime is still taking a lot of money out of the system through licences and regulatory intervention.

At the moment, service providers are making money out of the network, but the only people funding their expansion are customers. That’s a problem, suggests Obermann. But there are solutions, particularly in more intelligent network management.  “The dumb pipes need to become smarter and very, very efficient,” he says.

And the telecoms companies need to change. “We don’t believe in walled garden strategies,” he said. “I’m serious about it. We have to open to two-sided markets.” Telecos will change and are opening up, he insisted.

Deutsche Telecom is launching an incubator in Berlin – hub:raum –  they want to help startup companies gain an advantage by partnering with them. They have advertising systems, and the distribution network they have for their stores, which they can open up to their partners to smooth their path to market. And they can offer infrastructure expertise.

They’ve signed a memorandum of understanding with General Assembly – to help them launch an educational campus in Berlin. They’re opening up network APIs, giving developers easier access to network features, like voice and conference calls, and billing. They’re committed to standardising those APIs.

Their vision? A virtually hyper-connected service mall. The networks are the linking elements between a plethora of service models. The networks provide service models and enabling models, to allow developers to build better products.

George Dyson

George Dyson at NEXT Berlin

George, a science historian, is taking us through a fascinating tour of the history of computing. He’s highlighting some significant figures like:

  • Gottfried Wilhelm Leibniz: In 1679, working in Hanover,  Leibniz outlined how to build a digital computer  – using marbles. The same principles that chips work on today were described with marbles.
  • Alan Turing – “being digital should be of more interesting than being electronic”. And he was right. At 24 he wrote a paper called On Computable Numbers With An Application to the Entscheidungsproblem  
  • John von Neumann – 1948 “Let the whole world consist of one long piece of paper tape”. Two dimensional compute matrix – what every computer today runs on.
The world has flipped from reliable code on unreliable machines, to reliable hardware but bugs in code…
Where will this go?
Four ideas:
  1. Three-dimensional computing. Turing gave us 1D, von Neumann gave us two. The ability to add multiple machine sat the same times has barely been explored yet.
  2. Template-based addressing. von Neumann’s system is still based on fixed numerical address in sequence. We need to move to template-based addressing. Google is starting on this.
  3. Pulse-frequency coding – how the brain works. It’s not what the code is – it’s how frequently the links connect.
  4. Analog computing – all this brings us back to that.