Ever get the feeling that events are moving too quickly?
Current political drama aside, the enterprise networking industry is evolving at a rapid clip now that hardware limitations are no longer a barrier to what can be achieved. Global Market Insights projects the data center networking market to hit $53 billion by 2024, a 14 percent compound annual growth rate. When the only thing preventing new networking solutions from entering the market is our own imaginations, it’s small wonder that the channel is suddenly filled with novel ideas.
But is this necessarily a good thing?
Could it be true that having too many choices is just as bad as having no choice at all? Multiple options on the table can often produce wrong decisions, or at the very least doubt over having made the right one. At the same time, it takes more time, money and effort to wade through competing claims, particularly when the ultimate goal keeps shifting with the advent of advanced cloud and edge infrastructure.
Network analytics firm Kentik recently polled more than 500 networking professionals and found that the sheer pace of new developments is causing a fair amount of anxiety. One of the reasons is that before SDN came along, network technology had not seen a major advance since the advent of MPLS some 15 years before. Now, suddenly, techs are having to deal not just with software-defined architectures, but the automation, artificial intelligence and broad federation of resources that accompany them. Underlying many of these concerns, of course, is that few enterprises have a clear handle as to how their legacy networks have evolved over the years, making it difficult to determine how to position themselves for the future.
Perhaps the biggest item on the networking wish-list is automation. As Data Center Knowledge’s Mary Branscombe notes, most enterprises are eagerly integrating application development and infrastructure management under emerging DevOps models, and automation is a way to boost the pace of network management and provisioning to support the continuous integration and continuous delivery of digital services. At the same time, automated operations are better equipped to handle self-service requests from both users and autonomous applications, further contributing to high-speed data environments, greater reliability and more efficient resource consumption.
These factors will likely play leading roles in the development of hyperscale data centers, which is growing at more than 26 percent per year and is likely to top $102 billion by 2023. Most of these facilities are being built around advanced fabric technologies, many of them homegrown, to accommodate the greater flow of east-west traffic. At the same time, the steady stream of advancements in silicon and other basic technologies by firms like Broadcom, Cavium and Cisco are starting to hit the data center market at large as it strives for greater convergence rather than physical scale.
Confusion over the range of choices to guide networking into the future is certainly warranted these days, but enterprise executives should not let it rise to a point that it paralyzes the decision-making process. Today’s decisions, after all, should address the finite problems that are hampering data flow now, whereas broader architectural changes can evolve over time in accordance with long-term strategic visions.
At this point, the only thing the enterprise needs to fear more than making the wrong choice is being left behind in a high-speed, highly flexible digital economy.
Arthur Cole is a freelance journalist with more than 25 years’ experience covering enterprise IT, telecommunications and other high-tech industries.