Magic shows can be extremely entertaining and the best have genuine “wow” moments. Coins are produced from nowhere, buses and buildings disappear from before your eyes, padlocked straightjackets are escaped from – all done with sleight of hand, mirrors and distraction; oh, and a sprinkling of magic dust.
Data centres are not illusions but real buildings and rooms, fitted out with a mass of physical infrastructure made from metal and plastic and silicon. They are fed by the invisible resource of electricity and produce invisible bits of data and that is magical in many respects. It requires a great deal of sophistication in the processes and people running those critical facilities and you would imagine that all aspects are monitored automatically and that all components are known and inventoried. They are not. Cue Data Centre Infrastructure Management or DCIM.
Where do data centre managers and operators find the magic in DCIM? Where do they find independent validation of it? Perhaps they do not have to look very far: Gartner published its “Magic Quadrant for Data Center Infrastructure Management Tools” recently. But it did not seem to contain much magic. Market penetration for DCIM is assessed as being in the single-digit percentage. There are a handful of vendors that are very large and there are a lot of small ones. Comparing products is hard because the span of capabilities is wide and variable with some being not much more than slideware. Pity the poor customer trying to make sense of this and find a product that hits the spot.
The world of DCIM is a very noisy one, most of it generated by vendors aided and abetted by market research companies. Yet the voice of the customer seems to be a very quiet one. It is all well and good to keep shouting “Look at me!” but how much better to hear “I am looking for a solution to a problem in my data centre, can you help?” coming from the other side. There is no shortage of self-promoting vendor ‘white papers’ and hype reports forecasting that the DCIM market will be worth >$billion tomorrow. But what of the reality?
Vendors do not share real revenue numbers for their DCIM software licences with market analysts so analysis is arguably guesswork. Press coverage tends to use the amounts of venture investment raised as a proxy for success and the focus is almost exclusively on US vendors. Yet we are seeing DCIM companies which have raised tens of $millions being acquired for buttons or simply closing down. We are inundated with announcements of partnering deals between vendors, sometimes promoting technologies that few if any data centres care about or are prepared to adopt. What we do not hear much of is the data centre manager who says “I need DCIM and I need it now”.
We need to provide clarity where there is confusion. Data centres are business critical facilities and tend to be run by very small teams who have too much to do with too few tools to help them. They are swamped with alarms that they have to spend valuable time filtering and assessing, using experience as the principal yardstick. They are asked for reports on all aspects of the facility’s performance. They are expected to find ways to reduce energy consumption because the utility bills are skyrocketing. Above all, they are responsible for keeping the facility powered and operational.
If you are going to sell to these time-starved people you need to put across a simple, compelling message that says “you need this, you need it now and you can afford it; in fact, you cannot afford not get it.” Not as easy as it sounds. The environment of the data centre itself is a conservative one and decisions to change things are not taken easily or quickly unless there has been a crisis. Introducing DCIM may be perceived as introducing risk but, far from it, introducing DCIM is about reducing risk and increasing service availability and reliability. Vendors who can reduce the friction of deployment by making it possible to start small, quick and safe and then scaling up can be attractive. Better still is a solution that requires little or no professional services beyond supporting the initial installation. Professional services are good to offer as a helping hand but not to promote as a dependency.
Perhaps that is the real magic: software that can be deployed quickly and remotely, with minimal customer effort, to provide infrastructure monitoring such that reliability and availability are improved for an investment that stacks up against energy efficiency gains. Abracadabra!