The internet, as we know it today, was shaped by a relatively small number of engineers who saw past the limits of a small network and imagined something far grater. Dr Robert E Kahn was one of those people. Along with Vinton Cerf, Kahn was the co-invetor of TCP/IP – the network protocol that underpins the internet. Add the physical layer, ethernet, that Bob Metcalfe created and Tim Berners-Lee’s world wide web, and much of what we rely on today came from the work of just a few pioneers. I had the opportunity to hear Dr Kahn speak on what he sees as the next big thing when it comes to the digital world.
Before Kahn began discussing Digital Object Architecture, he reminisced about the early days of his research. When the forerunners to the internet were being developed, there was no clear vision about what the public internet would be about.
“It wasn’t obvious that the internet was ever going to happen or even be a good idea,” he said.
Faced with little support from governments that didn’t see the value or potential of the work that was going on, the message Kahn and his colleagues received from their leadership was “If you want to mess about – feel free!”.
That ‘messing around’ kept expanding until “somehow, the monkeys got out of the cage” he added.
What’s interesting is that Kahn never saw the internet as being about a specific technology. TCP/IP, he said was about allowing different technology to connect. And while, at first, that was about people and computers, it has massively expanded with the IoT now a dominant theme in technology discussions. He likened TCP/IP to being like electricity – it’s an enabling technology.
Many of the early design decisions made about TCP/IP have born significant benefits. Kahn said that not other technology in human history has been able to scale like the internet. It has expanded from a handful of connections to billions.
But those connections have spawned a new challenge – how do you make it easy to get to data using a trustworthy system. That’s lead to the development of the Digital Object Architecture (DOA).
When ARPANet was developed, the networking addressing system was 16-bit and was focussed on connecting wires. Then came the internet with 32-bit addressing that allowed a vast number of machines to be connected and addressed on a network. DNS was developed as a way simplifying connections to machines although Kahn said that if they were starting the Internet today, they would make DNS an app rather than an architectural or foundational component.
The development of DOA is about managing information he said.
One of the challenges of the internet today, said Kahn, is that information is very dynamic. A URL that is valid today might be dead tomorrow. And while the data might still be available, finding it can become difficult. DOA was developed with a view to being able to identify and retrieve information that needs to be stored for long periods of time – in the order of decades or longer.
The three fundamentals of DOA, said Kahn, are that it is an open architecture with defined protocols and interacts, it is independent from underlying technology, and it is not complex for users. The idea is that DOA enables interoperability of information whether it is stored on the interest or not and that is non-proprietary.
The building blocks of the DOA are digital objects. Each object cossets of international that is represented in digital form and has an associated persistent identifier.
“Digital Objects are digital views of information with a persistent identifier – like packets on a networks,” said Kahn.
The DOA has three core components.
- a resolution component that resolves identifiers to “state information” about the detailed information. A resolution request yields a handle record
- repositories that store digital objects and enable access by the identifiers
- registries that store metadata that can be used for searching
Each object on DOA is allocated a unique identifier, called a handle. Handles are allocated by The DONA Foundation which has authorised several other organisations around the world to allocate handles as well.
DOA is not a new development – Kahn and Cerf began working on it in the 1980s with further development taking place over the next 35 years. But as the number of devices and volume of data on the internet has exploded, it is seeing renewed interest with various bodies and governments around the world looking at its usefulness.
However, while many elements of DOA have been adopted by government’s wanting to deal with issues pertaining to IoT, a recent meeting of the World Telecommunication Standardization Assembly (WTSA) opted for a framework based on DOA but removed all specific references to DOA and DONA from their resolutions.
It seems that the largely depoliticised world that allowed Kahn and Cerf to create TCP/IP is not allowing DOA to be adopted in the same way.