The Empty Data Center
In case you didn’t notice we are in the midst of a revolution in our economy. Historically, we’ve seen these before as the agricultural societies of yore gave way to the factories of the Industrial Revolution, and the factories of the 20th century gave birth to a whole new class of workers who were actively involved in the consumption of what they produced. The late 90’s and 2000’s saw the birth of the “Information Economy” where the intellectual element of human capital rose in importance. This continual revolution has been marked by the displacement of workers as old skill sets gave way to new ones. While for many the consequences of this economic obsolescence had negative consequences, these same shifts within the marketplace also offered opportunity for both those who could adapt, and generations of new workers who possessed the new skills required. Obviously, this all took place on a purely human level, but what happens when the new skill set required for a job is possessed only by a machine or lines of code? If the vision of folks like Vinod Khosla comes to fruition, we may witness the impact of this “creative destruction” within our own industry.
Speaking at a recent conference, Khosla stated that the most important opportunity in the business of IT is getting rid of all the IT people. While this may strike some as a little Malthusian, Khosla views it as a more of an effective method of cost control since, “People are a big cost in IT. Not equipment itself. Let’s take that out”. Ouch. While Khosla may have a point, I’m sure most IT professionals would probably prefer a more personal reference then to be lumped under the umbrella heading of “that”.
While Khosla’s delivery may be a little…blunt, it also doesn’t mean that everyone needs to start polishing up the old resume. What it does mean is that as our insatiable needs for computing capacity and storage continue to escalate, the corresponding need for large centralized facilities to scale and adapt will drive new hardware and software innovation. Rather than seeing these new requirements as the precursor to relegating “IT professional” to the same level as “blacksmith”, “cobbler” and “cooper” in terms of career potential, Khosla thinks that this shift in future data center requirements presents a world of new opportunity.
While I think that Khosla may want to tone down the whole “let’s eliminate people” meme a little bit, I think that he is both right, and wrong. Like most things, the architecture of data center networks is not a “one size fits all” proposition. The high volumes of data and the increasing speed in which people want to access it dictates more distributed structures with large centralized data centers working in conjunction with multiple edge locations, and, in some instances even micro facilities. As Google, Microsoft, and many of their peers, have demonstrated, the need sheer size and complexity involved in making the modifications associated with scalability do indeed necessitate the need for automated alternatives to human intervention.
This doesn’t mean, however, that data centers will be devoid of humanity tomorrow, or ever for that matter. The requirements of centralized mega-centers are not the same as their edge counterparts where scalability is more “labor intensive” including the performance of tasks such as adding or removing servers, drives, etc. The delivery on Khosla’s vision will be primarily centered on the firms, both new and existing, that will be focused on developing the hardware and software applications that will be needed to support these new scalability requirements of the central structures that form the backbone of these distributed networks.
As would be expected these “hyper requirements” of these web-centric giants will gradually filter down to the enterprise. The longer gestation process for this market will provide the broadest field for new product development with multiple vendors competing with nascent start-ups. For many, the emergence of these new firms will translate into some of today’s IT professionals becoming tomorrow’s start-up employees.
The rapid rate of evolution in the data center industry can be viewed as a microcosm of the historical impact of economic evolution. The constant pressure on data center operators to support higher volumes of information will require both automated alternatives to many of today’s people oriented processes in mega facilities and the development of new data center network architectures. As daunting as this may seem to some, this evolution should not be viewed with trepidation but as an opportunity to translate existing skill sets and knowledge into new areas of opportunity.