Future internets: building a better system

Is the internet broken?

Hesselman: "Not at all. The internet's an outstanding success. However, it's the product of a different era. When it was developed in the seventies, no one could have envisaged the way it's used today. The original idea was to enable universities to make efficient use of expensive hardware by connecting their computers together in a network. Security, content distribution and mobile access weren't really considered. Such things simply weren't part of the problem they were trying to resolve. Things are different now, because the internet has changed the world enormously and we need to respond to that."

Frissen: "The internet has done a lot for us, in both economic and social terms. We now have access to all sorts of services that simply wouldn't exist without the internet. However, the internet's success has unintended side-effects. As a result, many of the positive things we've looked forward to for years are at risk. There's a danger that people's confidence in the internet will be eroded."

Why is that?

Frissen: "A small number of players have increasing power. The internet is increasingly monopolised by a handful or major international corporations. And, in some parts of the world, the internet is being used as a medium for state control. Certainly in countries such as China, the way things are going gives cause for concern. Ordinary people are more and more dependent on the internet as a vital social resource, yet they have less and less influence over the way things work. Labs and the Fund both want to restore the lost balance as far as possible. And one way of doing that is by contributing to 'future internets'."

Hesselman: "The point is that technology isn't neutral. It makes a big difference where something is developed. Just look at the way data is handled. Broadly speaking, there are three models (see lecture by Professor José van Dijck of Utrecht University). There's the American model, where corporations own users' data, the Chinese model, where the state owns the data, and the European model, where personal data belongs to the people it relates to. The question is: how do we sustain the European model if we're largely dependent on technical systems based on the American or Chinese model? When the internet took off in the nineties, governments tended to leave things to the market. However, now that the internet's become a critical infrastructure, it's important that Dutch and European society has more control. Do we want Silicon Valley and China to define the parameters, or do we want to decide for ourselves what form our societies should take in the future?"valerie

Is there a lot of interest in this subject?

Frissen: "At the European level, it's definitely a topical issue, more so than here in the Netherlands. European privacy legislation can certainly be seen as a significant attempt to restore the lost balance, at least partially. Europe is considering whether competition laws should be applied more strictly, and similar moves are afoot in countries such as France and Germany. Europe is also promoting the development of Responsible AI, artificial intelligence based on human values, with a responsible attitude towards data and algorithms. Considerable research funds are being invested in programmes such as Future Internet and Next Generation Internet – in other words, how best to update the existing internet."

Hesselman: "On the technical side, and especially in academia, people have been asking for a decade or two whether the existing internet infrastructure could be improved. Numerous ideas have been put forward for alternatives to the current Internet Protocol (IP), which forms the basis of the internet as we know it. The IP is very flexible: you can extend it in lots of ways, but the extensions are ultimately all just add-ons. That's the case with DNSSEC and secure routing (BGPsec/RPKI), for example. If you want a fundamentally different network concept, you need to come up with something new."

What stage of development are these ideas at?

Hesselman: "A few are at quite an advanced stage. That's the case with NDN, for example, and with SCION. They're still essentially academic research projects, but both initiatives now have testbeds with dozens of machines. In the US, the National Science Foundation has invested millions in NDN, while ETH Zürich currently has a team of twenty people working on the development of SCION. The SCION work is being done in partnership with companies such as Swisscom and SWITCH (operator of the .ch domain). ETH Zurich started the project back in 2009." 

What are SIDN Fund and SIDN Labs doing in this field?

Hesselman: "In October, we launched a future internets research project, as part of a consortium with the University of Twente, the University of Amsterdam, SURFnet and NLnet Labs. Within the project, we're concentrating on issues that are relevant to people's everyday lives: security, availability and transparency of network communications. Also, SIDN Labs is now connected to the SCION test network, which is also active in this field. By doing a lot of experiments and developing new applications with our consortium partners, the aim is to take these future internet concepts to the next stage. Our ambition is for the Netherlands to play a lead role in this field, as we did with introduction and adoption of the existing internet in the eighties.

Frissen: "For the last five years, SIDN Fund has enjoyed a lot of success with open calls for project proposals, leading to the funding of a large number of very varied projects. However, we're changing tack slightly and adopting a more programmatic approach to funding. Future internets is definitely one of the fields we'll be focusing on. However, whereas SIDN Labs is concerned with infrastructure, we'll be concentrating more on applications and making them accessible to as many people as possible. A good example is The Things Network, a project that offers a public infrastructure to people using smart devices connected to the Internet of Things (IoT), enabling users to take back control over the data collected and used via the IoT Having started in the Netherlands, it's now active in about eighty countries around the world. I think that we have a lot to gain by working with SIDN Labs. Cristian and I will be discussing the options over a working lunch in the near future."

Where are the biggest opportunities?

Hesselman: "Programmable network equipment is now coming onto the market. Such equipment can be an enabler for new internet technologies, because the devices can be set up to operate on the basis of new non-IP protocols. So, together with our research partners, we're creating a programmable network here in the Netherlands for experimentation. Personally, I foresee the emergence of a more diverse network environment that enables us to use not only the existing internet, but also various alternatives. So you might have an internet specifically for health care applications, or for the Internet of Things."

What are the main challenges associated with a new internet?

Hesselman: "I see deployment as the number-one challenge. There are numerous technological options, but to demonstrate their added value they need to be made available to a sufficiently large group of users. That's easier said than done. How do you persuade people to start using them? Again, programmable networks could play a major role there."

Frissen: "Often, the stumbling block isn't the technology, but people's habits. Habits are hard to change. As long as people prefer the service they're getting from a monopolist to a better, more secure alternative, nothing will change. New technologies and applications need to be convenient and to offer users real control and real benefits. At the same time, we have a responsibility to promote awareness. That's been one of our priorities at SIDN Fund all along. Another significant challenge is governance, I believe. The era of laissez faire is definitely over; more regulation is needed, as well as sound agreements and forms of interaction. A new internet can't function without a good governance structure. In view of its experience within bodies such as ICANN, SIDN can definitely play a significant role in that context."