There was much discussion around skills – IT, facilities and the skill sets that are lacking in between – at the recent DatacenterDynamics Converged event in California in June. Some discussions focused around what was missing while one roundtable took the discussion back to its very infancy, asking panellists to identify if there is a skills gap at all.

DatacenterDynamics CTO and board director Stephen Worn proposed three key questions – around skills ‘myths’ –  to a panel which included HP Critical Facilities Services distinguished technologist Richard L Sawyer, Stanford University research computing strategist Phil Reese and Digital Realty director of technical operations for the west region Danny Johnson. Here is what they said.

Stephen Worn: Myth 1 – There is no skills shortage. All I need to do is pick up a phone and I can get any capability I need, or it’s all in the Cloud so I don’t have to worry about it at all …

Richard L Sawyer, HP Critical Facilities: It is always interesting to ask people [in data center engineering] where they came from because nobody ever starts out their career by saying “I want to be in the data center business”. This makes it really interesting to figure out how people got into this business. It also means there is no large pool of people out there that we can draw upon consistently.

There are no standards today for data center operations that are universally accepted across the whole industry.

And there are no certification programs that are generally accepted industry wide.

DatacenterDynamics (with DCProfessional Development – the company’s training organization) has made real progress in getting its certification programs out there, and there are competing ones, but there are still no broad standards for educating people.

Danny Johnson, Digital Realty: I agree. There is no standard training or certification process for that would produce a data center-ready pool of people through some sort of training mechanism. That is a challenge for us as a data center operator/owner but it is more of a challenge for the industry, which needs to be thinking about how to actually shore up that skills gap.

Digital Realty has internal training and processes we follow – methods of procedure for doing some works – but there is definitely a need to have an accredited training program that people can easily go through.

Phil Reese, Stanford University: My perspective is slightly different because I am with a university.

I will say while Stanford doesn’t offer a degree  in data center management we do find professors much more interested in looking at the data center statistics and data center data center systems data. They are building that into their research portfolio for studies, and I think that is a great thing.

Stephen Worn: Myth 2 –  Data center infrastructure management (DCIM) – automation – solves this problem. You need fewer people working in the data center, there are fewer issues with downtime because of human error. After all, there is an app for that …

Danny Johnson: There is an app for that but there is not an app for human interaction. DCIM does provide an extraordinary amount of data and quite quickly but sometimes data becomes quite skewed when you are talking about a workload.

When looking at the amount of data DCIM produces (and if it is implemented in the correct way that obviously can be a study in itself) how do you find the information you need when you need it and then how do you utilize that information?

I say you will still have to have a person to take that information and use it. So DCIM does not replace the need for the data center operator. It is a tool – something that allows us to take a large amount of data and use that to more effectively or more efficiently operate the data center.

A human being still needs to come in and switch breakers, do emergency operations, ensure the generator is taking on the facility or call somebody when something doesn’t switch right. Human interaction will never be replaced by DCIM.

Richard L Sawyer: You look at the DCIM offerings out there and the best you can hope for is a very good systems integrator holding the UPS information, BMS information, workload management information, energy capping systems all on the facilities side. Then you replicate that on the IT side and get that usual dashboard with the information overload. I mean you almost have to have a heads-up display going on for people.

How DCIM deals with depth in each one of those channels is where the human element comes in.

Let’s face it, data centers have three basic components – there is the IT  architecture, the facility infrastructure then there is a people element. You have to spend the same amount of concern and resources putting processes into the people element as you do any other piece of the data center,  otherwise you are going to lose it.

Phil Reese: DCIM is just getting started. There are a lot of problems with DCIM still – a lot of gaps. Is it the ERP (enterprise resource planning) of the future for data centers? That is a scary thought …

Stephen Worn: Myth 3 – IT and facilities, you can cross train them, not a problem. The facilities guys, they can learn IT  and IT we can train them to do anything, they are pretty smart …

Phil Reese: I don’t think that is quite the reality. I think IT has enough trouble trying to keep on track of its own stuff.

I do, however, think the facilities side and the IT people have to work closely together. For years this has been an issue – are you IT or are you facilities? And I think this is still an issue. That myth hasn’t gone away.

I think there are strengths in each role learning more about the other but there is not a full overlap at all. I think there is still a differentiation between IT and facilities.

Richard L Sawyer: As an industry we actually have the answers. If we borrow from the facilities side and get answers from the IT side things will work out much better.

It is about making sure you can deliver the process effectively at an acceptable cost.If we take these principals and put it into an ITSM  (IT service management) framework and tie that to facilities, then we are looking at three things – the capacity of the compute, the availability of that computation and doing this at a reasonable cost. If everyone’s objectives – including the IT and facilities people – are geared  toward meeting these three requirements for service delivery in terms of an SLA, everyone will be on the same page.

When we get down to it data centers are big electron factories. We take large amounts of electrons, high voltage, high current, and send it through a data factory so we can put out 1s and 0s. And guess what? Our electrons are all at one definable state – and that is high frequency.

This is what we package, this is our finished product. The sooner we get to realize that we are a data factory and the IT processing is the machinery and facility infrastructure is what supports it and people are what make it happen, the better off we will be.