The Internet of Things and 5G could be among the benefactors of an upcoming $20 million U.S. government cash injection designed to come up with new architectures to replace existing public internet.
FABRIC, as the National Science Foundation-funded umbrella project is called, aims to come up with a proving ground to explore new ways to move, keep and compute data in shared infrastructure such as the public internet. The project “will allow scientists to explore what a new internet could look like at scale,” says the lead institution, the University of North Carolina at Chapel Hill, in a media release. And it “will help determine the internet architecture of the future.”
Bottlenecks, security and overall performance are infrastructure areas that the group are looking to improve on. The “Internet is showing its age and limitations,” Ilya Baldin, director of Network Research and Infrastructure at the Renaissance Computing Institute at UNC-Chapel Hill is quoted as saying in the release. “Especially when it comes to processing large amounts of data.” RENCI is involved in developing and deploying research technologies.
“Today’s internet was not designed for the massive datasets, machine-learning tools, advanced sensors and Internet of Things devices,” FABRIC says, echoing others who, too, are envisioning a new internet:
I wrote, in July, for example, about a team of network engineers known as NOIA, who also want to revolutionize global public internet traffic. That group wants to co-join a new software-defined public internet with a bandwidth- and routing-trading system based on blockchain. Others, such as companies FileStorm and YottaChain, are working on distributed blockchain-like storage for Internet adoption.
Another group led by researchers at University of Magdeburg, whom I’ve also written about, want to completely restructure the internet. That university, which has received German government funding, says adapting IoT to existing networks won’t work. Centralized security that causes choke points, is just one trouble-spot that needs fixing, it thinks. “The internet, as we know it, is based on network architectures of the 70s and 80s, when it was designed for completely different applications,” those researchers say.
FABRIC, the UNC project, which has begun to address ideas for the architecture it thinks will work best, says it will function using “storage, computational and network hardware nodes,” joined by 100Gbps and Terabit optical links. “Interconnected deeply programmable core nodes [will be] deployed across the country,” it proposes in its media release. Much like the original internet, in fact, universities, labs and supercomputers will be connected, this time in order for today’s massive datasets to be experimented with.
“All major aspects of the FABRIC infrastructure will be programmable,” it says. It will be “an everywhere programmable nationwide instrument comprised of novel extensible network elements.” Machine learning and distributed network systems control will be included.
The project asserts that it’s the programmability that will let it customize the platform to experiment with specific aspects of public Internet: cybersecurity is one, it says; distributed architectures, could be another.
“If computer scientists were to start over today, knowing what they now know, the Internet might be designed in a different way,” Baldin says.
