Managing power, cooling, and space remains a primary challenge in data centers. Here Soeren Schroeder, software director at Schneider Electric, tells us why that is still the case.
The full presentation can be viewed as a live blog-cast here.
The usual thing to hear is that no matter how exotic the feature sets that various providers come up with, the core of the challenge in managing data centers really remains around the three areas: space, power and cooling, which was probably true ten years ago and still carries through today.
The challenges Today
In my opinion we see a dilemma today, a ‘see-saw’ challenge in managing data centers. This is really balancing your availability with relatively ‘new’ business efficiency requirements, which might be lowering the energy bill by a factor of two, saving on the operational expenditures like headcount, being a bit smarter with less carbon emission, etc.
Historically also in the business different management entities have different KPIs. So in the facility management part of the organization we have uptime and the availability of capacities like power and cooling. At the other end we have the IT organization which is typically rewarded and measured on speed of deployment, how to meet changing business requirements, etc. So really the dilemma to some extent between the two groups is to get them to communicate using the same vocabulary
What we have seen in the past, being only 5-10 years ago, really pre virtualization, is we saw a power usage in the data center that fluctuated. It waved between some minimal level and some maximum level which really meant that as long as we kept our total data center capacity somewhere above our worst case power draw then we were safe. The reality that we are seeing today is that this fluctuation has increased in magnitude, so we chose to implement more capacity to the data center. Unfortunately with this increased magnitude of the fluctuation we have seen a situation that becomes a bit less efficient in terms of energy spent and we keep a very power hungry infrastructure system going ‘just in case’.
Cooling is a little bit different. We have a lot of input and a lot of information but really our key challenge is how do we turn all of this information into some tangible knowledge? How can we really make sure that we can manage the cooling element of it rather than just react to input? How can we really understand the cause effect of all of these potential issues that we have in our data centers?
What you need is something that is much more aligned with your day to day operation rather you typically bring in a consultant to facilitate for you.
Space in a sense encompasses all of these challenges. This is where we typically talk about DCIM. DCIM’s nature is really putting the world from various vendors including Schneider Electric to actually facilitate the whole management of power, cooling and space together and a lot more. DCIM is much more than a software package, or a software suite, it’s really a state of mind, it’s a deliberate choice of managing your data centers better.
StruxureWare for Data Centers is Schneider Electric’s contribution to the DCIM community, a world leading suite that facilitates a lot of the challenges that we see today. So taking your existing facility systems, your existing IT systems, coupling them further with a fine level of detail of information from meters and censors etc turning that into one great operational efficiencies.
Using Data to Manage the Data Center
The key message is that it’s still power, cooling and space that are the challenging pieces to manage. DCIM is really there to eliminate risk, it’s really there to facilitate energy management in a more proactive way than what we have typically seen. It’s about using it to improve your existing business processes. StruxureWare for Data Centers really is a key set up tool for you in order for you to be able to take knowledge based decisions that is really supporting your day to day operation.