There’s an old saying that goes, “If it didn’t happen in New York, it didn’t happen.” This was popular in the early 20th Century when New York was the world’s predominant city. If you wanted to make a big splash with the public, you went there instead of Paducah.
A similar thing can be said of the networking industry. Developments can happen throughout the vendor universe, but it doesn’t become a big thing until Cisco either builds it or buys it. This was evidenced again this month with the company’s introduction of Intent-Based Networking (IBN). IBN truly is a big thing, but not just because Cisco is doing it. In fact, you could argue that Cisco’s IBN includes nothing that it doesn’t already offer, given that it is based on existing platforms like the Digital Network Architecture (DNA) system and the Catalyst switching platform.
What’s really going on with IBN, says Virtualization Review’s David Ramel, is the ongoing transition from hardware-centric to software-driven networking, which provides the underlying construct to define network architectures according to how you want to use them rather than what the technology will allow. Gartner, in fact, has been talking up IBN for quite a while, and the firm has even gone so far as to lay out some of the basic requirements, such as translation and validation of higher-level business processes, automated implementation, awareness of the network state and dynamic optimization and remediation. Each of these is likely to be implemented in different ways by different vendors; some may even add their own features to the list. However, the gist of IBN is that users basically spell out what they want, and the network operating system creates it using the available pool of physical and virtual resources.
It comes as little surprise that shortly after Cisco’s announcement, Juniper unveiled its own plans to implement IBN. Its solution will be built around the open source Contrail SDN controller. The new Contrail E2 will provide intent-based network overlays that are expected to be vendor neutral, says ENP’s Sean Michael Kerner. This means it could conceivably oversee Cisco infrastructure, as well as hardware from other vendors. The E2 is currently under development within the GitHub community, where only recently the talk has turned to intent-driven functions. This also causes one to wonder whether, like with Cisco, we are seeing a more refined definition of the use case rather than an entirely new technology initiative.
It was about this time last year, if you’ll recall, that Apstra unveiled its AOS operating system, which from the get-go was described as a vendor-agnostic, intent-based networking solution. To make the system even more self-functioning and agile, the company recently added full transparency to the interrelation of features like topology, policies and telemetry – essentially turning the system from a black box into a highly visible virtual networking platform. A key addition is the “stage blueprint,” which allows operators to model entire networks beforehand rather than applying changes piecemeal to the active blueprint.
All of this represents quite an improvement over existing modes of network management, which should lead to more agile, user-friendly data environments. But it seems that intent-based networking is the logical end result of the abstract networking, automation and programmability that was already happening anyway, not a radical new direction that leading vendors have only recently developed.
IBN is a great idea, which is why the networking community has been striving for it all along.
Arthur Cole is a freelance journalist with more than 25 years’ experience covering enterprise IT, telecommunications and other high-tech industries.