The road to network automation is a bumpy one. Not only does the enterprise have to implement broad architectural changes to today’s static infrastructure, it has to devise new operational parameters to ensure that the results they expect are the ones they actually get.
But network automation is no longer a choice for the modern enterprise. Data loads are too large, user demands for speed and agility are too great, and competitive pressures to cut costs and increase performance are too intense to delay any longer. Any enterprise that has not already started this transition is already behind the curve — perhaps permanently.
According to KBV Research, the market for network automation in North America alone is set to expand more than 40 percent between now and 2023. The field covers everything from deployment and configuration to ongoing testing and management of both physical and virtual resources, all of which is falling under the domain of an integrated software stack that also provides enhanced security and visibility for this increasingly dynamic data ecosystem. If done right, the end results are a dramatic improvement in operational efficiency, reduction in downtime caused by human error and lower operating expenses.
The key words here, of course, are “if done right.” If done wrong, all you end up doing is automating the same convoluted management processes that are diminishing performance today. Juniper’s Mike Bushong notes that companies often make the mistake of setting out to automate infrastructure when they should be thinking about automating workflows. This is particularly crucial when it comes to the actions at the boundaries of systems, people and organizations, where the hand-off of data requires multiple remote and repetitive steps.
But how, exactly, does the enterprise transition from manual networking to automated? The details will vary, of course, but Mark Vondemkamp, executive VP of business development at network orchestration developer AppViewX, charted out a good path to success on eWeek recently, starting with identifying the workloads most in need of automation, progressing through infrastructure segmentation and standardization, and culminating in iterative, self-serviceable operations. This will require multiple tools and scripts at each stage, but ultimately the enterprise gains a new, agile application delivery architecture.
Early adopters are already seeing these benefits. Lee Congdon, senior VP and CIO at educational software developer Ellucian says that eliminating manual tasks frees up employees’ time to concentrate on harnessing the vast new arrays of tools at the enterprise’s disposal to improve products and business models. In many cases, this is leading to new opportunities to leverage automation in both internal and external cloud environments. In this view, today’s networking professional becomes an expert at managing and optimizing connectivity, not performing basic oversight and operational functions.
It should be clear by now that automation is not a one-time change, nor is there a fine line between an automated network and a manual one. For quite some time, the enterprise will be dealing with a mixed environment in which some processes are automated and some remain as they are.
And even when basic automation of the full end-to-end network is complete, there will still be a lengthy cycle of improvements and upgrades, including new intelligent and cognitive capabilities, to keep IT staff busy for quite a while.
Arthur Cole is a freelance journalist with more than 25 years’ experience covering enterprise IT, telecommunications and other high-tech industries.