With the increase in data center creation and popularity, it is more important than ever to build and maintain a reliable data center. At last week’s Data Center World convention, Steven Shapiro spoke about data center decisions based on availability. Here are some of the highlights on availability data center design.
1. A redundant system isn’t always reliable. In fact, redundancy can create a difficult to manage and complex system.
2. Tier 2 and Tier 4 systems vary highly in price. The costs can be high, and it is better to match the needs of the applications your data center supports rather than to keep building.
3. 2(N+1) UPS config is not much more reliable than 2N UPS. These two systems are the same, except for the UPS configuration of 1(N+1). They have the same rate of failure, further proving that redundancy doesn’t always prove reliability.
4. The 2N generator configuration is only slightly more reliable than N+1. While there is a small change between these two, it is a marginal difference.
5. Eight hours after running, 15% of generators fail. Even with a positive start, many generators can stop working during a power failure.
6. Reliability can’t be determined by tier requirements alone. While the tier system can be a good base to begin assessment, there are many factors beyond this including site location, construction of the building, quality of the equipment, and more.
Read more about Shapiro’s presentation here.