Digital Transformation the Smart Way

Thursday Jun 29th 2017 by Arthur Cole
Share:

In order to enable digital transformation, networks will need automation and — eventually — machine learning capabilities.  

Back in January, I wrote a post entitled “No Digital Transformation without Networking.” And while the crux of the idea— that networking ties together all of the pieces of a digital enterprise — is still valid, it has become clear since then that an added element is needed for all of this to come together: intelligence.

Basic automation has already emerged as a core requirement in modern data center networks. As demand for services increases on both internal and customer-facing infrastructure, the need to alter network configurations rapidly is exceeding the capabilities of even the most advanced manual processes. But as trends like big data, the Internet of Things and mobile data services enter the mainstream, the network must respond not just to the expected connectivity issues of any given function, but to the unexpected ones as well.

According to Juniper Networks’ Steve Wexler, networking is at the heart of digital transformation, and intelligent automation should be the ultimate goal for any enterprise that hopes to compete in the 21st Century.

Fortunately, this does not have to happen all at once. The initial change will encompass human-driven automation, in which admins define the changes while software carries them out. What follows will be event-driven automation that has systems orienting themselves around different outcomes. Finally, we'll have machine-driven automation that can learn and function largely autonomously. Wexler says the ROI for network automation is extremely high, with payback in as little as six months and five-year returns approaching 350 percent.

Already, says Cisco’s David Goff, it is becoming clear that most failures of IoT and digital transformation initiatives are the result of inadequate networking. A recent IDC survey reported that enterprises which have invested in modern network capabilities are seeing two to three times better revenue growth, customer retention and profits than those that rely on legacy infrastructure, while at the same time seeing twice the success rate for transformation projects. With literally billions of connections coming online in the next few years, a network that cannot think for itself to some degree will be the chief barrier to everything from product and market development to customer engagement and strategic planning.

This is at the heart of Cisco’s latest push toward intuitive, intent-based networking, says ENP’s Sean Michael Kerner. The idea is to convert the “dumb pipes” that exist in today’s networking infrastructure into intelligent entities that can enhance business value. To that end, the company is combining elements of its Digital Network Architecture (DNA) with the Application Centric Infrastructure (ACI) and several new initiatives like SD-Access policy enforcement and SecureX context-aware security. And all of this will be deployed on the new Catalyst 9000 switch portfolio consisting of LAN port, access and core models.

One of the most valuable aspects of intelligent networking is the ability to identify and resolve bottlenecks and other issues before performance degradation becomes noticeable to the user. A company called LiveAction has been working toward this goal with its LiveNX Insight platform that uses deep learning and predictive modeling to bolster both security and network management. The system incorporates a mix of basic task automation for functions like fault monitoring and bandwidth management, as well as real-time correlation of network metadata and policy-based performance optimization and customization. In this way, network operations can actually be simplified even as the data environments themselves become more complex.

It certainly isn’t a stretch to think that in an age of smartphones, smart cars and smart cities that the enterprise should have a smart network as well. To be sure, there is still a lot of work to do before intelligent systems can take on IoT and cloud-scale production workloads. But fortunately, these systems have the capacity to learn how to do this on their own, rather than wait for someone to program their instructions.

And going forward, the biggest job for humans will be to figure out what we want our networks to do, and then sit back and watch them do it.

Arthur Cole is a freelance journalist with more than 25 years’ experience covering enterprise IT, telecommunications and other high-tech industries.

Share:
Home
Mobile Site | Full Site
Copyright 2017 © QuinStreet Inc. All Rights Reserved