Google Urges IT to Abandon Data Capacity Planning
Will the Internet of Things make capacity planning moot? Or will the discipline become even more critical to IT operations?
Nearly everyone in IT has realized that the industry needs to plan for disruption after continual disruption. Yet Google is arguing — given the massive increase in data traffic expected to result from the blossoming Internet of Things (IoT) — that disruptions will soon be so great and so frequent that any attempts to plan for capacity usage or data storage will essentially be futile.
Google has a point: capacity usage is exploding, with complexities popping up in every direction. But is this ample excuse for IT managers to ignore capacity, or should it force companies to look at resource usage with a doubly critical eye? In other words, if the IoT doesn’t make capacity planning truly impossible or impractical (as Google seems to imply), is there really any incentive for organizations to leave themselves at the mercy of the tides of the market?
It’s commonly accepted that enterprises must estimate their future capacity needs to account for, as examples, the launch of a new application or a cloud migration. But Google is warning that, while planning ahead is a good idea, the IoT may render predictions of anything less than five years into the future wildly inaccurate, according to Computer Weekly.
Undoubtedly, the number of internet-connected devices is skyrocketing. Cisco predicts that global IP traffic will grow threefold by 2019, when it will reach an astonishing 2 zettabytes (ZB), annually. For perspective, watching just one month’s worth of video content sent over the internet at that time would take you over five million years.
The numbers involved are indeed getting massive, which will make capacity planning even more difficult and unpredictable — a fact that few deny. Still, there are a number of reasons why this growing complexity will actually make capacity planning efforts all the more urgent.
It’s always prudent to estimate your future resource usage, but planning a half-decade in advance is unrealistic, even today. Organizations still have every reason to develop a clear strategy for next week, next quarter, and next year.
And while the IoT will require increasingly dynamic forecasting with shorter horizons, no company wants to hit high-traffic events like Cyber Monday without first analyzing trends from the past 12 months. In short, looking both ahead and back a year will never stop being a best business practice.
It’s also important to remember that, aside from the Googles and Amazons of the world, few companies will be dealing with capacity on the order of zettabytes or exabytes. In fact, that’s the exception to the rule (terabytes or petabytes will be much more routine). Google’s situation stands as a great bellwether for what companies should come to expect down the line, certainly, but it doesn’t present an immediate business case for the vast majority.
The implied argument is that housing such massively complex environments — hybrid cloud, multi-cloud, private cloud — all under one roof will make capacity planning impossible, full stop. However, IT ops and IT management vendors are actively anticipating these trends, adapting their tools to handle intricate, cloud-scale, and container-driven applications.
And these tools will be increasingly geared towards solving mission-critical complexity problems. While they may not hope to solve to solve every conceivable IoT-related problem, there will still be immense value in picking and choosing.
For instance, TeamQuest’s Vityl Adviser can help organizations prioritize their efforts by turning thousands of data points into a simple indicator of Health or Risk. IT managers don’t need to understand the entirety of IT at every moment, but they do need to address highly specific problems that can cause future issues — like an application inefficiency that will force a bottleneck in six months.
And within even the most complicated IoT environment, IT managers still require perspective. For instance, while Google’s argument focuses on data capacity planning, the realms of CPU and capacity planning will still be very much in fair territory, and may deliver just as much business value.
(Main image credit: Wikimedia)