New Research Looks at Cost of Downtime for Mission-Critical Applications

Although infrastructure owners and DBAs are embracing strategies to increase resiliency, the challenge of planned and unplanned downtime in organizations remains intractable. The problem is adversely affecting operational efficiency, revenue potential, and the ability to provide a trusted IT environment.

A new survey of 315 data managers and professionals who are members of the IOUG (Independent Oracle Users Group) scrutinized the issues related to mission-critical application downtime. According to the survey report, authored by Joe McKendrick, among respondents with at least two data centers and rapid replication solutions, 46% are less than satisfied with their current strategies. For sites with three or more data centers, this figure drops to 35%.

Satisfaction With Data Continuity Strategies at Sites With Continuous Availability

(Among respondents reporting having asynchronous replication, or continuous data protection.)

                               1 data center     2 data centers   3+data centers

Very satisfied/satisfied  0%                   18%                  24%
Satisfied                       43%                  36%                  40%
Less than satisfied        57%                  46%                  35%

The survey explored database availability and resiliency challenges and solutions, and discovered that service-level agreements (SLAs) are becoming more rigorous—and application owners are struggling to meet them. Close to one-fourth, 24%, of organizations in the survey, have SLAs of “four nines” of availability or greater, meaning they must have less than 52 minutes of downtime per year. However, approximately the same number of respondents indicated that they meet their SLAs sometimes, not frequently enough, or not at all.

Service-Level Agreement Time Frames

Less than 99.9% availability    40%
99.9% availability (<8 hours, 45 minutes downtime/year)    30%
99.99% availability (<52 minutes downtime/year)    13%
99.999% availability (<5 minutes downtime/year)    11%
Other    5%

The demands of recovery time objectives (RTOs) are also placing stress on organizations. Roughly 40% of respondents to the survey said their organizations had an RTO of less than 2 hours in which they were required to have everything back up and running after a disaster. One-fifth of respondents said their RTOs were even more stringent—less than 1 hour.

More than 25% of respondents experienced more than 8 hours of unplanned downtime in the past year. And, while likely to be far less disruptive, planned downtime of more than 8 hours was experienced by 50% of respondents’ organizations over the past year, contributing to low levels of satisfaction from IT directors and database owners.

In addition, thwarted by their lack of the right tools and technologies to meet the most demanding SLAs, only one-third of companies indicated they have the potential to support near-instantaneous replication.

Source: “Bringing Continuous Availability to Oracle Environments—2013 IOUG Mission-Critical Application Availability Survey” conducted by Unisphere Research, a division of Information Today, Inc., and underwritten by EMC Corp.

Image courtesy of Shutterstock.