Surviving the datacenter apocalypse: Skills, skills skills

My best advice to everyone who is potentially threatened by Cloud-based automation and datacenter elimination is to get thee some new skills. Quickly.
Written by Jason Perlow, Senior Contributing Writer

In a piece I wrote at the end of June, I discussed how recent layoffs in the tech industry have highlighted how IT pros working on physical datacenter infrastructure could be automated out of the job market.

I am sad to say that for the most part, it is unavoidable, as the primary driver behind cloud computing is the overall desire of large organizations to reduce IT spend in physical assets, software licenses, and human resources.

The fundamental problem is understanding the scale of of what that displacement will be in order to prepare for this sea-change event, which is difficult to quantify at this point, as we are in the early stages of the cloud adoption buying cycle.


Will the cloud eliminate as many jobs as it creates? All of this depends on how much infrastructure moves to the cloud and what kinds of organizations take to it. SMBs are likely to move first, as they have smaller IT departments and are woefully understaffed to begin with, so those are probably going to be a wash.

Great Debate: Will cloud adoption bring about a datacenter and IT job apocalypse?

The real test will be when the largest corporations start thinking about eliminating infrastructure in favor of a hybridized cloud solution, where some of the infrastructure resides on-premises — perhaps as low as 20 percent — and the balance of it lives in the cloud.

The ratio that each organization determines best suits them is also going to determine the amount of the displacement.

So what jobs are most threatened by the cloud? I think that the rank and file system administrator and network infrastructure person are most at risk. Also probably any teams dealing with PC support, especially once technologies such as desktop as a service (DaaS) become more commonplace.

It's also unlikely that employees in these types of roles will be able to substantially contribute or influence the buying cycle in order to preserve their jobs. The technology buying cycle right now is almost entirely driven by CFOs, who are putting heavy constraints on the CIOs.

That's what I've been hearing top down from every single large organization that I've dealt with in the last two years, and I expect that to be even more so in the future.

Anyone who is directly datacenter facing is going to be at extreme risk, and needs to re-educate themselves. Infrastructure architect or enterprise architect types are probably safe, but they need to keep their skills up to date.

The most skilled Cisco/network infrastructure and storage people will still have many opportunities, but they may have to end up changing employers to those that are actually cloud providers themselves as the balance of private datacenter infrastructure is eliminated. Sysadmins are a big unknown here, because tremendous advances are being made in provisioning and automation, and I expect that group to take a big hit.

So what kind of IT professions are the most valuable in a cloud-centric world?

Any job that deals with architecting, implementing, and managing cloud stacks and the virtual infrastructure that sits beneath, whether they are on-prem, hybrid, or totally off-prem. Cloud security and cloud storage experts. SaaS transition/software migration types of folks, who understand how to move a legacy system with complex multi-tier apps to a hybrid or off-prem model.

Additionally, application developers and software lifecycle management types will continue to play an important role, and will most likely be the group that is least affected by the transition because app delivery is the cloud's primary focus.

I also think that based on what we have seen with IBM's most recent resource actions and where the layoffs have occurred from a geographical perspective, we can expect the mass outsourcing of the 2000s to be largely undone by the end of this decade.

And I would add that due to recent security concerns that have come to light in wake of the Snowden debacle, it is far less likely that employers will want foreign nationals having access to remote systems with large amounts of personally identifying information (PII) on it.

Of course, there are IT jobs that cannot be automated. Any job that needs to make the kinds of decisions that relate to adapting to the speed of the business itself, and also, as I said before, software development. You also can't automate people who have to touch physical infrastructure or maintain facilities, but, as I said, there will be fewer of those kinds of people needed.

Can people who perform certain kinds of roles be re-trained or cross-trained to ones that are more valuable in a cloud-based world? For instance, can database admins become data scientists? Could a network admin move to more of an architecture role?

The problem with this sort of thing is that not everyone has the personality or wherewithal to make these sorts of changes, which is coupled with a fundamental skills acquisition and subject matter expertise issue.

I believe there are logical tracks that certain types of IT professions and practitioners can move along, but jumping tracks can be difficult from a skills adaptation perspective, and there is also the issue of experience and how long it takes to acquire the skills. A sysadmin or an integration specialist could become an architect, but, for example, it took me 10 years to do that myself.

And what of the hardware industry? The server and datacenter equipment manufacturers? What about jobs at those companies?

I think the server hardware industry is clearly feeling the crunch already, and if enterprise-class equipment is mostly going to be owned by cloud providers in the future, then clearly we can see where the derivative effects are, and there are too many types of job roles to count that fit into just this one bucket.

In addition to the hardware decline, we also have to look at things like subscription-based software models contributing to the overall displacement of resellers and VARs, although partner relationships will probably mature to more of a broker-based model, as they will want to continue to control the billing relationship.

So ultimately, does the cloud mean that many technology roles can be handled by more generalists, or line-of-business (LOB) folks?

I think savvy generalists as well as LOB people will continue to be extremely valuable, as will SMEs, for specific types of apps and workloads, but what we are ultimately seeing is a mass compression effect, where we don't need as many of types of people in all of the traditional roles.

My best advice to everyone who is potentially threatened by cloud-based automation and datacenter elimination is to get thee some new skills. Quickly. Learn Azure, learn AWS, and start familiarizing yourself with the various vendor private cloud stacks and public cloud offerings.

I don't just talk the talk with this one, though. I've also had to walk the walk and make fundamental changes to my own career. But it hasn't been easy, and I am continuously challenged as I keep learning new skills and adjust to new ways of doing things.

I am now, at this time, employed as a cloud SME and strategist at a software company that has reinvented itself as a devices and services company. I am intimately involved in helping our partners in the hosting and provider industry in assisting their own customers make the transition from a datacenter-oriented to a hybridized infrastructure.

At the moment, I'm on the beachfront of getting these stacks implemented. I think I can say that it has already drastically affected my career path, as I went from datacenter optimization expert to cloud SME in the course of a year.

So I certainly understand what being agile means. I would like to think that the experience will empower me to help more and more organizations make the same kinds of transitions, in the most painless way possible.

Are you preparing yourself for the datacenter apocalypse? Talk back and let me know.

Editorial standards