Energy Saving Service Opportunities from Veridity

Veridity announced EnergyCenter Version 2.0 on June 21 2011, and now represents a new opportunity for service providers to sell software and services to help clients achieve significant savings in power costs, and improve utilization of IT resources in mid-size and larger data centers.

The Achilles heel of such services is the setup costs, with a prerequisite to crocodile-up the data center (i.e. place crocodile clips on all the devices that consume power so energy consumption can be painstakingly measured). This approach takes time (manpower and elapsed), there are no standards, and is difficult, very difficult to justify.

Each week, Wikibon receives many presentations and the Viridity team impressed with a deceptively simple approach, and the impressive execution since our last briefing eighteen months ago. The secret sauce is a badly named “Genome” database, which contains measured and interpolated power measurements from a wide range of data center components, with an emphasis on servers and power/cooling equipment (these two categories account for 80%+ of data center power consumption). No hardware or agents are required, and the assets are discovered automatically by software. The software allows each frame in the data center to be populated, with average usage and server utilization.

Potential Services based on Veridity

The traditional approach to measuring the efficiency of the data center is the GreenGrid metric of Power Usage Effectiveness or PUE. This ratio is the Total Power usage in a data center divided by the power used for IT systems. The problem with this metric is that the denominator (power used for systems) contains two components that overstate the efficiency:

  1. The machine usage include the power used by fans in IT equipment, which should be included only in the numerator
    1. Veridity can tackle this by including fan ratings and estimating fan usage according to other factors.
  2. More importantly, the denominator include equipment that not doing useful work; servers and storage that is powered on, but not running any application.
    1. Veridity can tackle this by identifying the servers (and in the future storage) which is doing no useful work.

Providing a service to identify stranded assets that are not doing useful work will save most IT data centers a significant amount of power and maintenance costs, and could be a simple way to introduce the software and service. Additional services can help plan the introduction of new equipment and ensure that the power and cooling costs are optimized. Advanced services based on simulation of different data center and facility strategies to optimize the total cost of computing would be a natural build up, and offer very high ROI for data centers that had grown to trust the service providers.

The list price for the Veridity software is $500 per frame per year, with significant discounts for perpetual licenses. The software focuses on services, and can discover the utilization of processors.

The Bottom Line:

The cost of sale, time to benefit, ability to offer simple services growing to advanced services and low risk of the approach makes power management software very attractive for service providers and VARs/SIs looking for additional revenue streams and differentiation. The key investment is developing people that can communicate in Megahertz and Megawatts, and can express the value of the service and data base for power measurement without hardware or agents.

 

About David Floyer

David Floyer is Wikibon’s resident CTO. Floyer spent more than twenty years at IBM holding positions in research, sales, marketing, systems analysis and running IT operations for IBM France. He directly worked with IBM’s largest European customers including BMW, Credit Suisse, Deutsche Bank and Lloyd’s Bank. Floyer was a Research Vice President at International Data Corporation (IDC) and is a recognized expert in IT strategy, economic value justification, systems architecture, performance, clustering, and systems software.