Unifying ideas and initiatives: Data Center Stack Framework & OpenDCME

The current indexes for data center performance, such as DCiE, EUE and PUE are not sufficient to drive data center efficiency. These indexes focus only on the power or energy consumption of the facilities. Each metric in itself says nothing about how efficient a data center really is. In order to drive and improve efficiency, a common framework that will describe any data center, anywhere, doing anything is required. The next step is to apply industry established metrics for each block that is running in the data center. The combination of a framework and the metrics can form the basis of real data center performance monitoring.

And here come two things together.

Data Center Pulse (DCP), a non-profit, data center industry community founded on the principles of sharing best practices among its membership is working on Standardized Data Center Stack Framework Proposal The goals of the Stack are: Treat the data center as a common system which can be easily described and measured Provide a common framework to describe, communicate, and innovate data center thinking between owner/operators peers and the industry.  So the aim is simple – provide one common framework that will describe any data center, anywhere, doing anything. The next step is to apply industry established metrics for each block that is running in the data center.

Datacenter Pulse Stack Framework

Datacenter Pulse Stack Framework

Another initiative is the open source Open Data Center Measure of Efficiency (OpenDCME). In this model 16 KPIs that span the data center are used to measure data center efficiency. As stated “This first version of the OpenDCME model is based on, amongst others, the EU Code of Conduct for Data Centres best practices in combination with the feedback of applying the model to a large number of data centers.” Mansystems , a European IT specialist in service management, consultancy & support, created and released OpenDCME. The proposed measures belongs to the community and is open for contribution by using the Creative Commons license agreement. The model consists of four domains:

  1. the IT assets that are located in the data center,
  2. the IT assets efficiency
  3. the Availability, Performance and Capacity of the IT assets,
  4. the efficiency of data center IT processes.

The radar plot shown below is the presentation of the 4 domains and the 16 KPIs (4 per topic). The OpenDCME model, in its current version, does not tell you HOW to measure the 16 KPIs.

OpenDCME model

OpenDCME model

Comparing the Stack Framework and the OpenDCME model initiatives you can see that both are complimentary to each other. Bringing these to initiatives together can accelerate the development of performance monitoring and management of data centers.

Lets see what happens …

Bookmark and Share

Doing Your Energy Math

Energy Metrics

To improve the energy efficiency of existing data centers, as well as making decisions on new data center there are some metrics being used: Power Usage Effectiveness (PUE), Data Center Infrastructure Efficiency (DCIE) and Data Center Productivity (DCP). Ideally, these metrics and processes will help determine if the existing data center can be optimized before a new data center is needed.

The PUE is defined as follows:

PUE = Total Facility Power/IT Equipment Power 
and its reciprocal, the DCiE is defined as:

DCiE = 1/PUE = 
IT Equipment Power/Total Facility Power x 100%

And the DCP definition is:

DCP = Useful Work / Total Facility

Data center productivity is much more difficult to determine than PUE or DCiE. That is because of the issue how to define ‘useful work’. The Green Grid organization feels that DCP is a key strategic focus for the industry. DCP defines the datacenter as a black box – power goes into the box, heat comes out, data goes into and out of the black box, and a net amount of useful work is done by the black box.

PUE developments

On the site of Green Grid is stated “The Green Grid will also consider the development of metrics that provide more granularity for the PUE and DCiE metrics by breaking it down into the following components:

PUE= 1/DCiE= Cooling Load Factor (CLF) + Power Load Factor (PLF) + 1.0

Where all factors are ratios that are divided by the IT Load and:

• 1.0 represents the normalized IT Load. Effectively this is the IT Load Factor (ILF) but is always 1.0.

• Cooling Load Factor (CLF) = total power consumed by chillers, cooling towers, computer room air conditioners (CRACs), pumps, etc. divided by the IT Load.

• Power Load Factor (PLF) = total power dissipated by switch gear, uninterruptible power supplies (UPSs), power distribution units (PDUs), etc. divided by the IT Load.”

Calculators

On the Web there are several calculators available to do some about the data center energy usage. Simple PUE / DCiE calculators such as offered by 42U but also more advanced calculators such as APC Data Center Efficiency Calculator. With the APC calculator you can also do some interesting What If / Trade Off analysis on different Data Center configurations. APC offers also other Carbon footprint and Energy Usage calculators. Another type of calculator that can be found is an efficiency savings calculator. Examples are 42U’s energy efficiency calculator and the European Free Cooling Tool an online tool has been developed to help data center and facilities managers easily determine how much free cooling and free evaporative cooling is available for individual data centers. An often forgotten thing about PUE and DCiE is that these figures are NOT static, they are dynamic. Workload but also the environment (outdoor) change on a daily base. A fine example of these dynamics can be find at Google Data Center Efficiency Measurements. Although these calculators can wetting your appetite in reducing energy usage and the carbon footprint and/or saving money real results can only be made when a proper energy monitoring system is in place Usage of an architectural framework is a must to get an energy monitoring environment working.

Bookmark and Share