Intent-based networking (IBN) is a great idea. With enough automation and even a bit of intelligence, IT infrastructure can start to configure itself according to the outcomes that users demand. No more mucking about with arcane management practices, undecipherable coding or the tedium of manual processes.
But it’s important to keep in mind that few organizations are even close to implementing IBN. So it will be a while before we can actually gauge the difference between how it works on paper versus how it works in live production environments.
According to a recent report from MarketsandMarkets, the overall network automation field is set to expand seven-fold over the next five years, jumping from today’s $2.32 billion to nearly $16.9 billion by 2022. The hottest segment of this market is expected to be IBN, which the firm says will deliver enhanced access control, scalability, security and multi-vendor device management for increasingly complex network ecosystems. Ideally, IBN will allow users to input policy-based requirements into management engines, which will then implement networking configuration and design commands and dynamically manage the environment through real-time routing and other functions.
Cisco was not the first to come out with an IBN platform, but it put the technology on a faster track to the mainstream last year when it launched its solution, Cisco Intent-Based Networking. Earlier this month, the company announced one of its first deployments at the Felix Platter Hospital in Basel, Switzerland, which will use it for key functions like the hospital information system, asset tracking, building automation and voice/video communications. By utilizing IBM in conjunction with Cisco’s DNA Center portfolio of integrated hardware and software solutions, Felix Platter is hoping to see simplified management, improved automation and segmentation, and a heightened security stance.
Still, leading experts are cautioning against high expectations for IBN in the near future. Michael Bushong, vice president of enterprise marketing at Juniper Networks, which has its own robotics-based IBN platform, pointed out recently that some 85 percent of all network interactions are handled through the CLI, which means IBN will not be a dominant solution until it has garnered a good 40 percent of that action. Given that most organizations are highly risk-averse when it comes to their networks, such a dramatic change is unrealistic for the remainder of this decade. Next decade, however. . .
Indeed, says Amy Gregory, of hybrid IT solutions provider OneNeck, IBN is a classic example of trying to walk before you can crawl. The fact is, most enterprises are just starting the transition to SDN and advanced network automation, both of which come with fairly steep learning curves, not to mention the level of integration and performance testing required of the actual technology. So before anyone starts looking ahead to fully automated, user-friendly network management capabilities, it would help to master the basics first. Right now, the focus should be on integrating more machine learning and data analytics into network architectures and figuring out how these can enhance orchestration and IT performance.
As a long-term goal, IBN is a worthy one. Not only will it greatly enhance the productivity of serverless computing architectures, containerized applications and machine-to-machine operations, it also has the potential to democratize the network provisioning process for emerging DevOps models of IT management.
And when that day comes, the enterprise will realize how productive it can be when it no longer has to worry about the technical side of infrastructure.
Arthur Cole is a freelance journalist with more than 25 years’ experience covering enterprise IT, telecommunications and other high-tech industries.