Archived Content

The following content is from an older version of this website, and may not display correctly.

Around five years ago the data center industry became a hotbed of activity. It began to experience explosive growth even during a recession under which the rest of the economy was shrinking.

The growth and excitement around this boom created a knock on effect through the entire data center supply chain.

Data center businesses, vendors, consultants, advisor, investors, all were buzzing with excitement. The explosive growth in mobility, converged on-line services and cloud were driving a seemly endless requirement for data centers world-wide.

In the midst of all the excitement and growth in the data center market, the software guys (and gals) were starting to turn their eye to management, addressing what was effectively a very old fashioned and traditional industry steeped in myth, superstition and FUD (Fear Uncertainty and Doubt).

While there have been software companies that had been around for many years servicing parts of the data center sector, they were very much behind the scenes:  BMS (Building Management System) is one good example.

Eliminating Excel
The default software solution to almost everything else in the data center world was Microsoft Excel. Rack elevations, equipment inventory, space planning, financial planning, capacity planning, all were done in complex, disconnected, manual and error prone Excel sheets. They served the needs of the time.

However, as needs started to evolve and the data center industry started becoming sexy and filled up with investor cash, then many more software businesses started to pop into existence with solutions to supposedly bring all those aspects of data center management into the 21st century.

Some were large and some were small, some had tens if not hundreds of millions of dollars invested into them by hungry venture capitalists all hoping to get a share of the growing market sector that eventually became known as DCIM - Data Center Infrastructure Management - a term coined by the analysts to give themselves something to track and make valuation predictions around.

Once the big tech analysts released their (frankly quite ridiculous) market sizing reports, citing growth, followed by more growth and culminating in a nice dose of explosive growth, everyone jumped on the bandwagon. Suddenly anyone that had any sort of software solution for almost any aspect of the data center were calling themselves a ‘DCIM’ business and collecting the cash from poorly informed investors.

This became a hugely over-inflated bubble in my view, but I don’t place all the blame on the analysts. They were mostly just watching and listening to what the market thought they needed. However there was a lot of transformation in the sector, and talk about energy efficiency and optimization. Clearly software was going to figure in the future of running a well managed and cost efficient data center…and that software was not Excel.

At one point one of the mainstream technology analysts reported there were over 80 new companies claiming to be DCIM companies.

DCIM or not DCIM?
Perhaps ironically now, I spent most of that period arguing with the analysts and some potential customers that my own company was NOT a DCIM business. Much to the frustration of some I flatly refused to be branded or have our own marketing material claim we too were part of the rapidly growing DCIM movement.

Many DCIM vendors believed they were filling an obvious gap between traditional BMS (Building Management Systems) and IT monitoring systems, both of which have been around for decades and are mature technologies. Many believed they could help make data centers more efficient from an energy and cost perspective and followed the mantra “you can’t manage what don’t measure”.

Some tried to pull together all sorts of data points into a ‘single pane of glass’ (Boy do I hate that term!). Some developed mystic algorithms which tried to correlate patterns and find trends in data that were just not there or, worse still, of little actionable value to a data center operator.

There was a time where every conference I attended, every showcase event had DCIM vendors aplenty, some spending huge bucks on industry event sponsorship. The panel and speaking sessions were full of DCIM vendors and proponents talking about how the brave new software enabled world of the data center was going to look.

Many of these DCIM vendors have gone bust, lost their way, run out of gas or been sold for peanuts - or parts - to one of the giant vendors. Most of them would admit that things didn’t quite work out as everyone had hoped. The smarter among them understand why. I say smarter because the reason is not simple, it’s multi-dimensional which always gives us humans a bit of a challenge!

Here are the primary reasons:

1. Boiling the ocean - Too many vendors tried to do everything and ended up failing to do even one aspect of their capability exceptionally well.

2. Blind leading the blind - Most traditional product strategies involve asking the market and target customers what they need. Only just today is the market finally starting to understand what it actually needs from a DCIM solution.

3. Misunderstanding or misrepresenting where the ROI is in DCIM - It is true that asset management, tracking, metering, 3D visualizations and instrumentation all make data centers easier to manage more efficiently. However, many DCIM businesses made the mistake of trying to tie data center opex (mostly energy opex) savings directly to the implementation of their software solutions. Buyers quickly figured out that the savings are in fact indirect, and thus hard to sell to the CFO.

Many data center operators that actually use DCIM today took the parts they wanted, left out the parts that had no value, wrote their own code to complete their solution or glue the parts together so effectively creating a customized in-house solution that best suited their needs.

Conclusion: pick up the pieces

The biggest problem that DCIM could solve in the data center is a capacity-planning gap between facilities and IT. The often-misunderstood element missing in almost every solution though is cost. Capacity isn’t ‘free’ and different types of capacity don[’t cost the same amount.

Yet we’ve been living in what we at Romonet call a ‘service mono-culture’ for decades. This is like having a shop that sells one product at one price. No matter what your requirement is, whether we charge you for it or not, the answer is the same product and it’s always the same price.

This fundamentally ignores the fact that not all data center capacity with an enterprise or service provider is created and operated equally. Everyone knows that each data center and even different areas within the same data center have distinctive build and operating costs (ultimately a different TCO), yet are unable to truly quantify these costs and then reflect them through to the users/buyers of that capacity.

Without differentiated products and differential pricing what you have is a planned economy, not the market economy that we all should have! Planned economies typically aren’t efficient ones as no real competition can occur.

So before the DCIM vendors out there shoot me, I agree that there is value in all the other ‘things’ that DCIM can do. Some folks will need great visualization tools, some will need great asset management and some (fewer) will need all of that stuff you’re trying to sell them.

However, he who solves the issues of service differentiation and differential cost based pricing first will win because this market will continue to commoditize thanks to the public cloud.

Zahl Limbuwala is CEO of Romonet. This article first appeared on LinkedIn.