Specify and Verify: Defining Data Center Criticality Levels

This audio was created using Microsoft Azure Speech Services

Determining a data center’s criticality level – how “important” the data center’s operation is to the business with regard to its ability to withstand downtime – is one of the most important decisions made in the system planning process.

Standardized criticality, or tier levels, help to articulate the availability and reliability performance of data center designs.

But more detailed specifications should be used that include unambiguous and defensible language against which a data center can be validated.

Choosing criticality

Three common methods for establishing criticality levels include The Uptime Institute’s Tier Performance Standards, TIA 942, and Syska Hennessy Group’s Criticality Levels™. These methods all support the concept of four common levels of criticality – numbered 1, 2, 3, and 4, with 4 being the highest level of criticality.

A data center’s criticality level depends on the needs of the business planning a new system or a retrofit. A business at criticality level 1 is typically a small business with a limited online presence and low dependence on IT. Downtime is viewed as a tolerable inconvenience.

A level 4 business, on the other hand, would be a multimillion-dollar business in which a majority of revenues are derived from electronic transactions. The business model is entirely dependent on its data center, and the cost of downtime is extremely high.

Choosing a criticality level also depends on whether the project is a new data center being built or a retrofit for an existing center. A new data center presents few constraints, and the criticality decision relies mainly on the type of business the data center is supporting. In a project for an existing data center, criticality is based on the constraints of the existing structure.

Verifying criticality

Once you determine a target criticality level, you need to specify the level, build the data center, and validate it against the specification. The process of validating the center against the specification allows legal recourse against substandard workmanship. It’s not enough, however, to simply choose a criticality level from an organization such as Syska, as this does not represent a verifiable and defensible specification.

A data center specification relays the essential requirements of performance, interoperability and best practice that will enable all physical infrastructure components to work together as an integrated whole.

An effective specification defines a data center with unambiguous, point-by-point specification of the center’s physical systems combined with a standardized process for executing the steps of its construction. “Baseline” specifications define a level 1 criticality data center, with higher level criticality specifications clearly labeled.

Business managers undertaking a data center project can get a data center specification through their organization’s corporate real estate department or via architectural and engineering firms that specialize in critical facilities. Businesses planning a small- to medium-sized data center project that are able to adopt standardized specifications can obtain a complete specification for little to no cost. One example is APC by Schneider Electric’s “Small/Medium Data Center System Specification and Project Manual.”

For additional information on choosing criticality levels as well as defensible approaches for specifying data center performance, see APC by Schneider Electric’s white paper, Guidelines for Specifying Data Center Criticality/Tier Levels.

Tags: , ,