The Internet Becomes The Interdata: Interview With Stefan Decker

Stefan Decker is the director of the Digital Enterprise Research Institute (DERI). One of its key funders is Science Foundation Ireland. It is based in NUI Galway, Galway, Ireland, and with a staff of over 130, it is the largest web research institute on the planet.

DERI works with industrial partners such as Ericsson and Avaya to road test and develop new ideas in Semantic Web research. Cisco, for one, is rapidly moving from a company which used to produce hardware and routers to a company that is connecting people much more than machines, and who are basing new technologies on work that has been done at the institute. Developments like SIOC are becoming a global standard in representing information about how humans communicate, and for the foreseeable future this sort of research can only grow in importance and relevance.

In a previous article, “The Fragmentation Of The Semantic Web“, a view was expressed that development in Semantic Web technologies may be side-tracked or derailed by vested interests and proprietary systems. However, Stefan takes a far more positive view of the situation.

“What we see right now is that our ideas are gaining ground and making progress and are becoming part of reality. It’s the reason why companies like Microsoft, Google and Facebook are actually using the same ideas.

“The same thing happened when it became clear we would need global information systems. MSN and CompuServe tried to do this, each with their own information systems, but they basically failed and they gave into open standards. Companies are doing the same thing again, but as Nova Spivack says, unless you really gain your own monopoly, really gain dominance in the market, you are not able to work together with everyone else. For working together, you need open standards: that is exactly where the Semantic Web comes in.

“In other words, what Google is doing, what Microsoft is doing, and what Facebook is doing is actually supporting us. It’s just a first step into a much larger game, which will end up as people and companies working together based on the standards that are being created. The only danger is one company like Facebook reaching global dominance.”

But the issue is not only open standards. Interoperability is the key to future developments.

Stefan goes on to say, “The semantic search engines that we are building are right now one of the largest repositories for structured information that exists on the web for open standards.

“What we see right now is a fundamental shift in the internet architecture. Previously, the Internet has been concerned about sending bits from one piece of the network to another piece of the network. But that is now more or less solved. We know how to send bits. What we now need to do is ensure that we are able to make sense out of those bits.

“In other words, the Internet at this point is in desperate need of a new layer in the internet architecture which takes care of data interoperability. It’s not so much about internetworking anymore but inter-data-working – making sense of each other’s data – which is applicable and available for all the specific applications.

“That requires a new data layer and that is something that we are building right now: the Linked Data Layer.”

The diagram at the top of the page shows where the Linked Data Layer fits into the current web architecture. In the current Open Systems Interconnection (OSI) stack, which consists of seven layers, there is a presentation layer which works a bit with data formats. This is directly followed by the application layer which is designed to make sense of the bits. The Linked Data Layer is a new layer between the application layer and the presentation layer which makes sense of the data in such a way that it establishes interoperability between different applications.

Stefan points out, “A recent Wired article indicated that a lot of the traffic is moving away from the Web into other applications. As the technology changes and allows new uses to come into being on new platforms, mobile phones and such-like, the need increases for the applications and other operational requirements to be able to work together – to be interoperable.

“This is one of the core topics at DERI, so one way or another everyone is working on the Linked Data Layer. It is all about interoperability, about making things work between different applications on the Internet.

“If you look at future internet initiatives that are going on around the planet, you have a lot special interest group discussions about security, about e-health, about smart cities and so on. They all have different information requirements. Then you have the mobile service providers who again have different information requirements. But what they are missing is a joint layer, something they can base their applications on.

“We all know how to exchange bits on the wire, but to make these different applications interoperable – to make smart cities use the security standards that have been defined – they need a common layer to understand each other’s problems as well as each other’s solutions. That is exactly what a Linked Data Layer can provide.”

With research and development well under way at DERI, Stefan’s next step is to get the word out to the larger Semantic Web community and beyond.

“One of my fundamental aims now is to make clear the need for this Linked Data Layer, and there will be an assembly in Ghent in December in which we will have a Linked Data Layer session. We will invite all the different communities in the Future Internet Initiative in Europe to help define what the role of a Linked Data Layer will be and what it can contribute to bring the communities together.”

7 thoughts on “The Internet Becomes The Interdata: Interview With Stefan Decker

  1. good piece, one remark:I would put the linked data layer below the presentation layer, more in a sense of seeing it as a common middleware infrastructure, which is accessed by applications (whereas applications employ their own presentation techniques/means as it best fits them.)The main role of applications is in my view to make the (linked) data from the middleware level actionable for the enduser.


  2. Hi Daniel – I guess it depends on how you interpret the Presentation Layer (…, which does format information for display/UI purposes in applications but is more about the formatting itself of the data / information – I guess if you can get the data in the right format, than you can define the meaning of the formatted data.


  3. Linked Data and ISO/OSI model would seem to be orthoganal at best. The Internet runs on the four layer TCP/IP stack where the top (application) layer consists of application protocols such as HTTP, FTP etc. How linked data can be seen as something which sits between application and transport (ie, between the HTTP and TCP protocols in a typical HTTP based linked data architecture) is bizarre to say the least.Linked data is an extremely powerful concept, and there is no doubt that it will dramatically change the Web, the Internet, and the way we run our businesses and lives. This diagram is confusing and wrong.


  4. Interesting comment. In the diagram I did not propose this as a TCP/IP extension as you may have noticed. You could equally call the presentation layer bizarre in relation to HTTP and FTP, which is part of the ISO/OSI stack. The ISO/OSI stack is conceptual and I used that on purpose.But you are right – data and transport are orthogonal. That is exactly the point. And I wanted to detach it from the Web or other application protocols – and you just confirmed that the diagram does this.But maybe there is a better way to do this. I am looking forward to your suggestions. We need to work towards a new abstraction of the internet architecture. Currently many different communities propose extensions (security, smart cities,… etc.). What is the abstraction you propose to get the different communities working together amd creates an abstraction that all these communities can work with? I believe the proposed didagram can work as a starting point. Anything that would work better?Best regards,Stefan Decker


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s