Survey says: Enterprise IT needs a Zoloft and a life coach

Aging data centers, cloud computing, high server workloads and more are stressing data center professionals out. New research shows, though, there’s no need to panic.

Enterprise IT needs a Zoloft and a life coach
Credit: Thinkstock

For the past seven years, I’ve conducted Uptime Institute’s Annual Data Center Industry Survey (over 1,000 end user respondents from around the globe, conducted by email).

Every year, some trend jumps out as the main theme. Maybe it’s because I’m turning 40 this year, but my takeaway from 2017 is that enterprise data center professionals need to relax—and reevaluate what’s important to their organizations.

Over the course of the survey, I’ve watched our respondents wrestle with uncertainties as the IT profession continues to evolve. But the data from this year’s survey illustrates that many of the industry’s concerns are not coming to pass, meanwhile chronic management problems go untended.

Let’s look at some of the recent data points to allay our collective existential dread, then take stock on what we can improve with a clearer mind.

If you’re not growing, you’re dying!

Most server footprints aren’t growing (60 percent of enterprises report flat or shrinking server hardware). Increased performance from technology advances, continued adoption of server virtualization, and offloading of workloads to the cloud are slowing the pace of server deployment.

But unless you are in the business of selling hardware, that’s OK!

enterprise it consolidation Uptime Institute

Celebrating stasis in an industry so driven by growth and consumption can seem radical or even subversive. But the fact that companies are able to meet the demands of their business with less hardware is a good thing. Effective hardware utilization is the key to overall efficiency and profitability of enterprise IT.

With that shift, a lot of enterprise IT organizations in mature markets aren’t building new data centers.

Unless you are meeting pent up demand in a developing economy or working for a colocation company or hyperscale, you probably aren’t getting $50 million next year for a new site, even if your data center is starting to show its age.

Which brings us to our next overblown concern.

How will our legacy data center handle new workloads?

What we see in the data is half of enterprise IT organizations are upgrading facility infrastructure. About half the respondents report dealing with an aging chiller system or building management system (BMS).

Of course, it’s going to be challenging to do system upgrades to live sites. But if this is your organization’s directive, they’ve accepted those risks. Also, you are prepared. Most of you have spent your entire career in the mission critical space, preparing and planning for outages, for things to go wrong.

If you need to replace your chiller systems, maybe it’s a good time to install that economizer you had been considering. Maybe the BMS upgrade would be a good excuse to get that Data Center Infrastructure Management (DCIM) software project started.

The fact is an operational, utilized and depreciated facility asset is good for the balance sheet. And it’s not as if the data center you built 10 or 15 years ago can’t hold up. Uptime Institute has been studying server rack density for years, and the consistent finding is that high density computing is not materializing for most companies. You may have pockets of high density computing for specific projects, but the vast majority of sites still operate at less than 6kW per rack.

server density Uptime Institute

We’ve got to keep up with what the hyperscales are doing!

Every data center conference for the past 10 years has had someone from Facebook, Amazon, Microsoft or Google on stage talking about all the very cool stuff they’re doing. Guess, what? It doesn’t apply to you!

Your company doesn’t need to install a data center on a barge in San Francisco Bay, integrate a wind farm or deploy something that looks like a modern art installation.

One of the most-cited hyperscale trends of the last few years had been Facebook’s Open Compute Project, a way to totally rethink the integration and design of the entire IT infrastructure stack, from data center to the processor.

It’s innovative, disruptive and incredibly cool. It was going to transform the industry!

But only 2 percent of data center respondents said they’ve deployed it. Over 40 percent had never even heard of it.

The cloud is going to put me out of a job!

According to the survey, 67 percent of respondents reported workloads that would have previously resided in their enterprise data centers have gone to the cloud.

But the data also reveals that while the IT organizations are moving some of their workloads to the cloud, the percentage of workloads residing in enterprise-owned/operated data centers has remained stable at 65 percent since 2014.

hybrid server deployment Uptime Institute

Many critical, legacy business applications aren’t able to run in a cloud architecture. A lot of workloads are safe, for now…

But many senior enterprise execs have been pushing to get out of the data center business. Data centers require specialized staff and real estate for a non-core business function. Enterprise planners find it impossible to match the cloud’s flexibility and agility. And lastly, financial departments faced embarrassingly poor cost accounting or accountability for enterprise IT. 

Public cloud computing is forcing organizations to look at their enterprise IT in a comparative light, and many senior business execs don’t like what they see:

  • Poor capacity planning
  • An inability to scale and respond to the business quickly
  • Poor visibility and transparency into cost

You may be able to deliver IT services more reliably and with less cost than a cloud provider, but if you can’t articulate that effectively to your stakeholders, you’re dead.

To recap: Critical workloads still reside in the on-premise assets, but your execs probably aren’t happy about it.

If you’re retiring in the next few years, feel free to stop reading. Otherwise…

The hard work ahead

The specter of cloud means your role is under more scrutiny than ever before. You’ll be asked to do more with less budget and staff. And nobody is going to sign off on building you a new data center when your CIO is reading news sites claiming Amazon is going to be running everything in 2020.

So, you need to focus on three things:

  • Prepare for an extended stay in an aging site.
  • Get serious about chargeback.
  • Insert yourself in the decisions about compute venues.

Management and operations: If you’re sitting on a site your company spent a lot of time and effort building 10-15 years ago, you aren’t in a bad place. A lot of thought and investment went into those sites, and you’ll be able to do really well in those data centers—if you focus on operations and maintenance. Update your maintenance programs, invest in cross training staff, and evaluate your equipment for replacement. If you are upgrading live sites, sequencing, training, and procedures become even more critical.

Focusing on training and maintenance will extend the life of your site, improve your employee retention, and allow you to be more responsive to deploying your company’s private cloud strategy.

Become the CFO of IT: For many companies, there is a disconnect between IT infrastructure costs and the lines of business. Accurate forecasting and asset utilization are not prioritized if no one is held accountable for those functions.

Chargeback is a method of charging internal consumers for the IT services they use. Instead of bundling all IT costs under the IT department, a chargeback program allocates the costs of delivering IT to the business units that consume them.

In 2015, less than a third of survey respondents said their organizations had deployed a chargeback accounting method. Ongoing conversations with large IT organizations suggest that many companies are still struggling with this effort.

It is critical to be able to discuss the cost and benefits of on-premise enterprise IT to budget-holders on the business side.

Service provider evaluation: According to 2017 data, 70 percent of the respondents reported that their organizations’ processes for evaluating cloud, colocation and on-premise computing options needed improvement; 15 percent described the process as incoherent.

Taking your IT off premise isn’t a panacea. Managing a portfolio of assets across multiple platforms, geographies and providers will increase complexity. Your lines of business need your help to bring governance, diligence and stability to this runaway train.

But in order to have a meaningful role, infrastructure professionals need to stand back from day-to-day activities so they can understand what is important for the whole of the organization.

Join the Network World communities on Facebook and LinkedIn to comment on topics that are top of mind.
Must read: 10 new UI features coming to Windows 10