Ten Need-to-Know Tips for Selecting a DCIM Vendor

Data Center Infrastructure Management's (DCIM) promise has been discussed ad nauseum. So, many potential end-users understand the concept and like what they've heard about theoretical benefits but just aren't quite sure how to begin their journey.

In a Data Center World presentation, Stuart Hallin, a senior technical consultant for Cormant, offered some vendor-neutral tips on how to pick a DCIM solution so users can move toward implementation. He discussed the general criterion required to pick an initial solution and how to select the right vendor among what has become a virtual "sea" of DCIM vendors - all making big promises. Here are 10 things you need to know:

1. Focus on needs not wants.

A big mistake in making a DCIM selection is picking based on features, not needs. A lot of DCIM vendors promise the world, but many are, in truth, specialists in particular areas. One size does not fit all, and your situation is unique. Establish a few of your biggest immediate needs, and select a vendor strong in those areas. Look at DCIM from a cost vs. pain perspective instead of focusing on shiny new features.

2. Have a deeper goal in mind.

You likely will not implement your end game out of the gate. Keep your ultimate goal (i.e. automation, integration) in mind and understand that DCIM will evolve in that direction.

3. Don’t make setup and configuration the center of your evaluation.

Again, DCIM needs to be viewed as a long-term strategy. Easy setup does not necessarily mean that DCIM will evolve with you.

4. Focus on enforcement of process, updates and keeping DCIM accurate.

This one falls under the “DCIM is not a magic bullet” category. The best tools in the world are useless in the wrong hands, and DCIM is an ongoing process. Look to get better measurement tools to improve the quality of what you measure such as RFID and barcodes. In addition, this will make system updates easier on operators and help DCIM stay up to date.

5. Establish a Framework.

Start at a high level and dive deeper from there. Hallin suggests starting with equipment lifecycles. Vendors are more than happy to assist with implementation, so definitely take advantage of it.

6. Share goals and working processes with vendors.

Vendors know what works and what doesn’t, and they want your DCIM implementation to succeed; so be open with them. Other suggestions from Hallin include:

7. Make them prove it.

Vendors make a lot of promises so make them prove they can meet your top needs with a Proof-Of-Concept or trial. Again, this is more worthwhile if you’ve already established your criteria. Hallin suggests paying for the trial, some consulting, or a vendor's on-site presence for the best evaluation. DCIM vendors are increasingly competitive for these deals, so make them work for it.

8. Start with a single site or area for initial deployment and grow out.

Set main stakeholders. Are they in operations, facilities, networking, sysadmin etc.? Then, evaluate accordingly. Look for thorough operations analytics capabilities for the given stakeholder. Keep in mind that each stakeholder looks for different things. A manager wants to easily see the past, present and future; and how much DCIM is reducing costs or increasing efficiency. Taking easy snapshots and enforcing processes are two more ways to help make initial deployment successful.

9. Understand cost of investment.

The investment goes a lot further than just DCIM. In addition to acquiring a license, there’s maintenance and support, professional services, additional hardware like RFID scanners and mobile devices, and time. How many hours a week do you want to invest on training internal staff?

10. Determine the sources for budget and people.

Hallin says to consider tapping budget from multiple departments and determining how much work will be outsourced. The top outsourced item is record conversion. Make one or two people "own" DCIM. While you might want multiple staffers to be able to pull useful knowledge from DCIM, you need to make it clear who is ultimately responsible for its success.