Last week we conducted an online webinar devoted to discussing the common mis-understandings and myths associated with DCIM. Over 400 people registered for the webinar and we had a ton of questions and comments afterwards. Its very clear that DCIM is a brand new category of solution for many of the attendees and there are many assumptions and incorrect data points that are preventing many end-users from realizing the benefits of DCIM.
I have selected a handful of the more popular myths we explored during the webinar, and present them here along with a more detailed narrative about the reason for the “myth” and the informed facts that should be considered instead. My goal is to provide the necessary DCIM facts for your consideration and to seed your thought processes as you begin your DCIM journey.
So here are the myths, along with my commentary (in no particular order):
Myth #1: “DCIM is really about Power Management”
This is one of my favorite topics in the category of DCIM. As it turns out DCIM and the energy crisis arrived on the scene about the same time, circa 2007 or so. So it is not surprising that many Data Center practitioners associate the two. And if those practitioners are cut from the Facilities cloth, they would be mostly correct from their perspective. DCIM *can* include power management. But DCIM is actually a whole lot more. For the IT side of the house, there is an entire asset lifecycle just waiting to be managed. Every server and every switch, along with all of its connectivity needs to be managed from the moment it arrives on the shipping dock, to the point in time (maybe 4-5 years later) where it is decommissioned and scrapped. All of the remediation and configuration of each device must be actively managed. Whew, a big job when you consider thousands of devices in a typical data center. Most importantly, while the value of DCIM to the Facilities crew and the value of DCIM for the IT folks are both very tangible, there are many companies that have realized dramatically larger savings from actively managing the IT asset lifecycles than they saw by controlling cooling energy costs and balancing power loads. (Keep in mind that in some places in the USA, we still have 2.9 cent kilowatts and many of us are still paying just a tiny bit more per kilowatt than we were back in 2007)
Myth #2: “Existing data center operational and planning processes work fine as they are”
Your processes were likely born 15+ years ago. You likely had those processes handed down to you from your predecessors. While we all grew up on the premise that keeping everything running in the data center was our main role in IT life, the reality in today’s economy is to keep business services available as needed and at the right cost. Right cost? As needed? What does that mean? Well it means that data centers should no longer operate in a one-size-fits-all mentality. Each business service that is being delivered has a different value and priority to the organization. The acceptable amount of downtime can and SHOULD vary accordingly. Take a company that has internal mail users and the same company that has online ecommerce transactions. It is easy to see that keeping the ecommerce-related business services running will be significantly more valuable than keeping internal mail running. Put into tactics, think Tier-3 or Tier-4 for ecommerce services and Tier-1 or Tier-2 for back-office (email style) services. If this scares you, it should. We didn’t grow up with this level of granular detail and business alignment, but that is EXACTLY where we need to get to, and FAST! Hence, the processes to manage these new alignments must be different than those you are likely practicing today.
Myth #3: “Any DCIM vendor can integrate systems”
The integration of multiple systems from different vendors is a task not taken lightly. Although enterprise software integrations have been performed for years, the specific means used to satisfy those integration requirements can be considered one of the most strategic choices that must be made for long term success. IT business executives that have lived through any enterprise-class integration process will have learned many of the complexities that can arise in making the wrong decision. As a rule, custom code is the root of all evil. While nearly all DCIM vendors will say that they CAN integrate their offering with any external package, there are a ton of details (read: red flags) to consider. These start with the ongoing upkeep and the need to accommodate changing business processes, rules & logic. How can these changes be accounted for when a tactical integration is hardwired in custom code? What about taking advantage of new features introduced by either system? Again, hard to do with spaghetti code. Then there is support and the task of maintaining new interfaces or security mechanisms. And like any programmer knows, its tough to debug integrated system unless the proper tools and auditing mechanisms have been built-in, something notoriously omitted due to cost when custom code approaches are used. And finally, long-term cost! Yes, while many DCIM vendors quote integration cost as a single “professional services” line item, the reality is these integrations live for a dozen years or more and require ongoing maintenance and enhancements… With custom code approaches, the only music you’ll hear is “Cha-Ching”! There must be a longer-term view of cost (i.e. TCO) that must weigh heavily in your considerations. You’ll want to seriously understand the off-the-shelf integration approaches more carefully. Off the shelf integrations keep the long-term engineering costs and overall performance responsibility on the vendor’s shoulders. (A good thing)
Myth #4: “Three-dimensional views are better than 2D in a DCIM offering”
This is an easy one. While 3D views of your data center are pretty, REAL WORK gets done in 2D. Period. Sure there are instances where 3D is desireable, such as looking at front to back collisions on a rack when two devices are planned to occupy the same R-Unit. And then there is CFD studies where 3D may show value to visual thermals at any point in the 3D space. But make no mistake that when changes to the data center are required, 2D is the rule of the road. Users of 3D drop back down to 2D to manage the data center. Work orders are in 2D. Metrics and reports are in 2D. In short, actual work gets done in 2D.
Myth #5: “All DCIM offerings are basically the same.”
Really? Are all cars the same? Is the Hope diamond just another rock? The term “DCIM” can describe a broad range of capabilities for data center infrastructure management and each offering can be competitive or complementary to any other offering. There are IT-centric features and there are Facilities-centric features. Some address power, others address connectivity, and some focus on process workflows. Some DCIM offerings provide environmental sensors or intelligent monitoring. The long-term goal of DCIM is to provide the convergence of everything IT and Facilities that affects the infrastructure and can make it more efficient. Towards that end, each DCIM vendor has chosen which portions of the DCIM pie that they are able to deliver today. While some vendors paint a long-term picture of what their DCIM solution might look like in 10 to 15 years and ask you to trust them, I urge you to consider just what you can get your hands on today. In addition, the capabilities available within each product you review will be directly related to the cost to acquire it. So if you take a look at one solution which costs $80K to purchase, and another vendor has proposed to you a solution that costs $240K to purchase, rest assured you are comparing apples-to-oranges. There is something very different about them, which may take some (well-spent) time to investigate and decipher. Understanding is key. In fact, even if two solutions are identical in price, they may be providing completely different capabilities! Do your homework, ask for help from your peers and trusted analysts!
Myth #6: “Building a DCIM system ourselves would be easy and will suite our needs just fine”
IT professionals have grown up where necessity drives creativity. With generic tools like Autocad, Microsoft Excel and any flavor of SQL database, the people that manage data centers have cobbled together their homegrown management solutions for years. It’s just the way it has worked since there hasn’t been any purpose-built tools like DCIM in the past. The problem with building your own is that these homegrown tools tend to be rigid and minimal at best. And although the biggest value of DCIM is always seen when it is connected to external ITSM and Facilities systems, when you build your own DCIM, you won’t ever get to those integrations. You just don’t have the time or deep programming expertise. One person just can’t replicate the richness and completeness of what dedicated DCIM companies have invested tons of engineers on for years. That means you won’t be building strategically important workflow management, hooking to your existing ticketing systems or connecting to your CMDBs. And if you do begin to build your own DCIM, it will only capture the way one person (read: YOU) thinks about solving the infrastructure management problem. Homegrown tools tend to be very tactical and do not leverage resources well. Worse yet, the person charged to maintain these homegrown tools (again read: YOU) only do so in their ‘spare time’ since they actually have REAL JOBS to do which they are measured upon… not building their homegrown tools. Ultimately, this homegrown solution is not viewed as a strategic asset to the organization at a time where IT business services ARE being viewed as strategic. At a time like this, where economics and efficiency is key, strategic tools to support the delivery of IT business services are essential. Building a DCIM is a fool’s errand and a much more expensive one at that (when you consider ALL of the actual costs, the missed opportunity and capital deferment costs).
Myth #7: “We are going to the Cloud, so we don’t need DCIM”
The short answer is this- while it is very topical for senior executives to stand-up in front of their shareholders and declare their “IT strategy” to be based upon the Cloud and bask in the assumed lower costs of computing in the Cloud, its simply just a small piece of the BIGGER IT story. Today the estimate is that the most aggressive companies that have already stated their Cloud direction are at most 10-15% Cloud based for production IT services today. Move the calendar ahead 5 years and a good estimate of their use of Cloud will be 30-40%, and in 10 years they will be lucky to be using the Cloud for 60% or so. It turns out that a strategy to use the public Cloud is directionally sound, but the process to get there is a much slower one than anyone cares to admit. Fortune 1000 companies tell us that conversions to the Cloud are costly and the current public Cloud pricing models are far from advantageous versus running those same applications in a well-run in-house data center if such a data center existed. The public Cloud is interesting frankly because most in-house data centers have been run sub-optimally for so many years and switching to the Cloud puts all of the burden onto someone else’s shoulders. So while more business services will be migrated to the public Cloud over time, the significant need to manage enormous amounts of gear in-house will not be obscured. In most cases public Cloud and In-House will live hand in hand for a very long time, or perhaps forever. DCIM is the means to make your in-house data centers comparable to public Cloud offerings, and in some cases cheaper over the long run!
The DCIM market is chock full of partial pieces of information which are in many cases presented out of context. There is also a slew of vendor posturing and hand-waving. Vendors just want to say YES to every question in this highly competitive market and look at the whole market as an OJT (On The Job Training) exercise. What should you do? Do your homework and get past asking simple YES/NO questions. Any solution you are looking at should be able to SHOW you the use cases that you define. Rule of thumb: If you can not see and touch any vendor’s stated capabilities, they are not real. Ask the hard questions, demand proof. Do your homework and you’ll see that DCIM can be implemented today, will extend your existing service management capabilities, and will significantly lower your cost structures in doing so.
DCIM hero’s are being made every day. Be one of them!