On the Road with DCIM – Affirmation & Embellishment of Our Underlying Thesis
I was part of a Forrester Team that recently completed a multi-country rollout tour with Emerson Network Power as they formally released their Trellis DCIM product, a comprehensive DCIM environment many years in the building. One of the key takeaways was both an affirmation of our fundamental assertions about DCIM, plus hints about its popularity and attraction for potential customers that in some ways expand on the original value proposition we envisioned. Our audiences were in total approximately 500 selected data center users, most current Emerson customers of some sort, plus various partners.
The audiences uniformly supported the fundamental thesis around DCIM – there exists a strong underlying demand for integrated DCIM products, with a strong proximal emphasis on optimizing power and cooling to save opex and avoid the major disruption and capex of new data center capacity. Additionally, the composition of the audiences supported our contention that these tools would have multiple stakeholders in the enterprise. As expected, the groups were heavy with core Infrastructure & Operations types – the people who have to plan, provision and operate the data center infrastructure to deliver the services needed for their company’s operations. What was heartening was the strong minority presence of facilities people, ranging from 10% to 30% of the attendees, along with a sprinkling of corporate finance and real-estate executives. Informal conversations with a number of these people gave us consistent input that they understood the need, and in some cases were formerly tasked by their executives, to work more closely with the I&O group. All expressed the desire for an integrated tool to help with this.
One area that surprised us was the prominence of asset management as a discussion topic for the panel discussions and the consistency of the stories and issues. Although Forrester has cited asset management as a benefit and core function for DCIM in previous reports, it was not reported as a primary driver and received little additional focus. I would still hesitate to classify it as a primary driver like the cost savings associated with power optimization, but it appears to have strong impacts to multiple constituencies in the enterprise. In our panel sessions, the discussions around asset management almost always came back to the same common themes:
- Asset management as it is done today is a problem for many companies. Usually only those whose business it was to manage and deploy assets for clients, such as hosting providers, could answer affirmatively to questions such as “Can you report the real state of your assets today if needed?” and “Can you locate a given asset with no trouble from your asset CMDB?” Most admitted that they could not pass this screen. Anecdotes about “phantom equipment” were common – one large telco provider manager admitted they once had a switch that took two weeks to physically locate.
- Finance databases and I&O and facilities databases were seldom if ever in agreement, and many executives admitted to variances in the 10% to 20% range. The most inaccurate pieces of information were cited as location, upgrade/update status and “is/is not” – do we still own this asset and is it in use doing what we think it is doing? Interestingly, even the finance people seemed willing to accept the DCIM CMDB as the database of record after a physical inventory done as part of a DCIM deployment.
- DCIM vendors still have work to do on integration with existing asset management systems and workflows. Several discussions centered on the complexities of enterprise procurement and deployment policies and how difficult it was to get the DCIM tools to really interface with existing practice in such a way that the assets flowed correctly from procurement to the DCIM database. Issues were cited in terms of interfaces with legacy applications, both bespoke and ISV, complex workflows, bulk deployments and split deployments, where purchasing releases were not necessarily well synchronized with actual physical asset deployment. We would note that both Trellis as well as all other DCIM solutions offer integration capabilities of varying degrees of completeness, but to date the adoption of them has been vestigial at best. My expectation is that we will see successful integration projects rolling out over the next 24 months as users tackle these complex parts of their long-term DCIM rollouts.
Another area that caught our attention was the complexity and mutability of stakeholder communities. We noted several DCIM users citing that the number of licensed users for their DCIM implementations tended to expand over time, with the oldest established user of a precursor product called Aperture (currently moving to Trellis), citing an increase in his user community from 75 to over 800 (the company is a multi-billion dollar integration and service company), and spreading from traditional operations and facilities users to finance, marketing, sales and strategic planning. This was an interesting validation for the premise that these tools are useful across a wide range of constituencies once they are installed and opened up to other groups.
All in all we came away heartened by the enthusiasm of the groups for the basic premise of DCIM – regardless of the fact that this audience was selected because they were expected to have an interest, what we heard was very pragmatic (“How do I get this?” “How do I use it?” and “What can I expect in the way of benefits?”), with essentially 100% recognition of the underlying problem.
All things considered, a very receptive environment in which to be selling a new product. Potential users should expect rapid evolution of product features and very heavy competition.