This post was originally posted on Tech Republic.

 

There are more choices for data center upgrades than all but the biggest of companies can afford, and surely the choices for which ones to perform will be hotly debated in CIO offices as 2019 budgets are settled. Servers, storage, networking, virtualization, security, service providers, and numerous other options are all on the table, and don’t call me Shirley.

I polled many vendors and data center providers to see what’s on their minds for next year. What single upgrade, I asked them, should be the most important one for a medium-to-large organization? Some companies replied with real insight; most unfortunately sent thinly-veiled sales pitches. The following are the more compelling responses.

“I think the greatest lesson that medium and large businesses should take on-board is to establish a formal governance routine over their data center, one where the company does not simply look at troubleshooting or firefighting on a day-to-day basis,” said Bart McDonough, CEO of IT security consultancy Agio, based in New York.

“Responsible parties must take a more deliberate review of the environment instead, whether it’s through the lens of security, IT infrastructure, or something else. Establishing a protocol also ensures that the right internal stakeholders are engaged. The identity of these stakeholders will depend on the size of the organization, but it would typically include members of senior leadership, business development executives, and those responsible for deploying new offices, product lines, hires, and company acquisitions. For IT professionals, learning about company developments from these individuals after the fact can create significant problems downstream,” McDonough continued.

On a similar note, “The most important thing you can do to improve your data center for 2019 is to hire a data protection officer or chief information security officer who has a firm grasp on all data governance laws and requirements and who can plan and execute a strategy for the enterprise that doesn’t waste precious resources and time,” added Jackie Rednour-Bruckman, marketing officer at a small Silicon Valley file-sharing specialist called Wanpath, doing business as MyWorkDrive.

We’ll drink some egg nog to that. Agio’s McDonough added other suggestions:

  • Don’t procrastinate any longer about upgrading user desktops from Windows 7;
  • check to ensure that data backups and disaster recovery plans actually work; and
  • ask more (and tougher) questions of your technology suppliers.

Several companies answered that it’s time to wish conventional hard drives good riddance. The question of when to do so has been on the minds of officials at Hewlett-Packard and Toshiba, while capacity issues about solid-state drives are gradually fading away. However, cost is still an annoyance.

“While there are lots of things that can be done to improve the efficiency of a data center, if capex weren’t an issue, a clear winner would be to replace SAS drives with SSDs to reduce power consumption. Typical 2.5-inch, 15K SAS drives will use about 9 watts, while an SSD will effectively use 0 watts while idle and about 2 watts at load, [creating] a savings of around 70-75 percent on disk power utilisation,” observed David Barker, technical director of UK-based 4D Data Centres.

Perhaps smartest of all, “The one single consistent area that can be improved is to avoid lock-in when I talk to IT organizations of all types. This is due to what will change in the future. While I am seeing this trend now, I’m convinced this will be a trend into 2019 and beyond. And the lock-in is across the IT organization,” said Veeam’s Rick Vanover, director of product strategy. “Any decision today that will avoid lock-in tomorrow is a wise one.” (Vanover is a former TechRepublic contributing writer.)

Pause while this reporter drinks more egg nog in strong agreement against vendor lock-in.

A representative for a company that sells unikernels said, “Unikernels would be the main improvement to make in your data center in 2019.” (Shocking!) The person did provide some good reasons: Unikernels are fast, small, and can help you consolidate virtual machines on hyperconverged infrastructure. Still, the very next reply was from a company suggesting disaggregating that same infrastructure, so that your company can upgrade only what’s needed and get better return on the rest of your investments. In other words, there are two sides to every story.

Officials at several data center providers all said the same thing: Pick one that gives you multiple paths to the internet, so your data isn’t inaccessible when bad things happen. Other useful data center tips:

  • Prepare for 5G access;
  • invest in automation tools; and
  • understand the different roles of server owners vs. racking technicians.

Any of the above suggestions would be good investments for your 2019 data center planning. Reach out to me or comment below with your own ideas.