Image: Posonskyi Andrey/Shutterstock
The internet needs to be replaced by something else or it will be useless when 5G comes along, says a TSSG researcher who claims to have the answer.
Telecommunications Software and Systems Group (TSSG) researcher and chief architect Miguel Ponce de Leon said the internet as we know it is not fit for purpose and he wants to replace it with something better.
De Leon is part of an EU-wide research project called Pristine to develop a RINA (Recursive InterNetwork Architecture), a computer network architecture that unifies distributed computing and telecoms.
‘When you think about, when the internet was designed originally back in the late 1960s, we never thought about mobility or security, which are huge issues that we see currently’
– MIGUEL PONCE DE LEON
De Leon believes that the current internet architecture that we are used to with broadband, mobile and Wi-Fi is inefficient.
Not fit for purpose for 5G
De Leon warns that the internet will be utterly useless when it comes to deploying 5G in the future.
RINA is a theory of how a communications network should be built, and the TSSG researchers have combined the theories of John Day with the ‘grandfather of the internet’, Louis Pouzin, who was involved in the creation of the original ARPANet.
“Look, it is just not fit for purpose,” De Leon explained. “When you think about when the internet was designed originally back in the late 1960s, we never thought about mobility or security, which are huge issues that we see currently.
“One of the things that we’ve done is that we are looking at replacing the current internet.
“We hear about 5G that it is going to be fabulous with all this quick activity. But the core network is just not ready. It needs to be replaced.”
De Leon likens today’s internet architecture to a two-lane highway built for vehicles with 50cc engines. But when 5G comes along, it will need three- or four-lane highways with turbo engines.
“It’s not right. You need to be able to rearchitect how the internet could and should be used for the future.”
The grandfather of the internet was right all along
De Leon and his team have gone back to the foundation of the internet and interviewed Louis Pouzin, who came up with the idea of the datagram, one of the first ideas around how you could move packets of data around a network.
“We know the father of the internet is Vint Cerf for TCP/IP, but ask Vint about who was there at the time running the groups – it was Louis. He gives great insights into whole lots of alternatives that could be used.
“What we’ve done over the last number of years is we’ve gone and implemented them and tested them against TCP/IP and … there are some good alternatives here that could and will be used in the future.”
Surely this would be expensive? De Leon thinks not: “When you look at how cloud data centres are put together today with virtual machines, there is now much easier ways to seamlessly replace the current internet architecture with alternatives”.
“We were at the SDN World Congress in October and we showed demonstrations of how seamless it is … to move over from the current internet to what we are proposing, which is a much better-architected alternative.”
De Leon indicates that the current limitations of the internet can be seen when homes and office networks experience congestion control when users try to access various apps or games.
“You still don’t get connectivity to the apps that you want,” De Leon said.
“On our alternative, you get a network for your application – a dedicated connectivity from your application all the way to the data centre or the peer application that you are talking with, guaranteed QRS, what we call a guaranteed network slice.”
What De Leon is proposing sounds wonderful but how soon will it become a reality?
“We are just applying TRL 6, which means we are proving the concept. The top of the range is TRL 9. This is in the next two to three years, to show how you can integrate this into products on site and for delivery.”
If De Leon and his colleagues are correct, then network operators and equipment makers need to pay attention, in case they put a massive data cart before the horse, when what we really need is V8 engines on a four-lane motorway.
