It seems that each year brings momentous change to the data center, and every December there are countless predictions as to what the coming year will bring.
But rather than simply add to the pile of all the obvious changes that are coming our way, I thought I would break the mold this year and focus on some of the not-so-obvious developments that nevertheless portend to radically alter the ways in which data infrastructure is provisioned, managed and consumed.
Perhaps one of the more subtle changes to infrastructure will take place in silicon. Tirias Research’s Paul Teich pointed out recently that core processor performance has ceased to be the key differentiator in the data center now that vendors and system integrators are moving toward accelerators. This will have the twin effect of boosting performance within smaller physical footprints and lowering power consumption because the workhorse chips of the past are no longer needed to support the entire workload. Going forward, much of the connectivity, analytics processing and other high-speed functions will be offloaded to GPUs, DSPs and FPGAs – a change that will remain largely hidden to most knowledge workers.
Part of this is due to the fact that maximum performance and minimal power consumption are key requirements of the hyperscale data center, which is now the focus of most infrastructure development these days. In the olden days, vendors built for top tier enterprise data centers and the rest of the industry had to make do with the leftovers. But as Semi Engineering’s Kevin Fogarty notes, economies of scale now favor top cloud providers like Google and Amazon, forcing the enterprise to follow suit both as a means to control costs and to integrate in-house infrastructure into hybrid cloud architectures.
And this leads us to the next major change in enterprise infrastructure: distribution. The fact is, the data center itself is no longer the sole province of the enterprise but inhabits a vast ecosystem of connected technologies in the cloud, on carrier networks, on the IoT edge and even our personal devices. Datacenter Knowledge recently highlighted a conference sponsored by Vertiv, formerly Emerson Network Power, describing the “Gen 4” data center as one that exists beyond walls and seamlessly integrates core facilities and intelligent edge infrastructure under a single architecture. Somewhat ironically, both the hyperscale facilities in the center and the micro centers on the edge will be built around the same modular, hyperconverged infrastructure, with resource provisioning and management taking place on virtual, software-defined layers.
One key question going forward, however, is how big a role artificial intelligence (AI) will play in this new ecosystem. According to a recent study by Teradata, AI might not be as imminent as some people believe. While the company found that 80 percent of IT and business decision-makers have already implemented AI in one form or another and that interest in expanding its use is high, more than 90 percent see significant barriers to continued investment and implementation. These include a lack of proper infrastructure to support AI, cited by 40 percent of the survey field, and a shortage of expertise on the part of both IT and the user base.
While it is tempting to say that 2018 will be the year that all of these amazing things happen, history shows that major changes to data infrastructure take years, even decades, to play out – and rarely is the transition total, as evidenced by the continued prevalence, even resurgence, of mainframes in the enterprise.
But if the IT executive takes anything into the new year, it should be the knowledge that data is no longer the means to conduct business, it is the business. And that means all the stops should come out when devising data infrastructure that is responsive, secure, inexpensive to operate and, most of all, reliable.
Arthur Cole writes about infrastructure for IT Business Edge. Cole has been covering the high-tech media and computing industries for more than 20 years, having served as editor of TV Technology, Video Technology News, Internet News and Multimedia Weekly. His contributions have appeared in Communications Today and Enterprise Networking Planet and as web content for numerous high-tech clients like TwinStrata and Carpathia. Follow Art on Twitter @acole602.