Venyu Blog

As Mean Time Between Failures (MTBF) and Mean Time To Recovery (MTTR) give way to Recovery Time Objectives (RTOs) and Recovery Point Objectives (RPOs), backup and recovery’s importance to the sustainability of data-dependent organizations is magnified.

Writing in the latest edition of Processor Magazine, William Van Winkle points to a recent industry survey that finds although business continuity/disaster recovery (BC/DR) is a key priority in the enterprise data center, 46% of respondents are not currently meeting their RTOs and RPOs in a consistent manner. Moreover, and perhaps even more alarming, forty-three percent of respondents now require more than 24 hours to complete a backup.

Quoting Rachel Dines, an infrastructure and operations analyst with Forrester Research, the reason that many companies aren’t meeting their RTOs is because they are relying on tape backups for operational recovery. As Dines observes, “offsite vaulted tape can take 24 to 48 hours to recover… these days most enterprises are measuring their RTO in hours, not days, for mission, and even business-critical applications.”

Rapidly supplanting tape media as a way to improve RTO/RPO ― according to Winkler a “game changer” ― is server virtualization as a way to expedite server recovery.

The benefits of server virtualization are well-known and reinforced in the column by an Info-Tech Research Group analyst who suggests that when server virtualization deployed to production systems are coupled with a robust SAN, an administrator can take snapshots of all virtual machines, mirroring them to a redundant facility which does not require redundant infrastructure. As the analyst concludes, [Virtualization] allows IT to build a more responsive, more recoverable infrastructure.

Moreover, as virtualization is applied across multiple connected data centers, making them centrally managed and mutually redundant, it can drop recovery times to minutes or even less.

Winkler also proposes that data deduplication is another key resource available to sustain effective RTO/RPO. Data deduplication keeps one instance of data, replacing other instances of that same data with a pointer back to the “master copy”, reducing disk capacity by as much as 90% or more. Combined with incremental and/or differential backups, which only record files that have changed since the last backup or full backup, respectively, deduplication can yield very small backup times and greatly extend RPO windows.

Winkler sums up his perspective on RTO/RPO by introducing the often forgotten human element in this discussion suggesting objectives can only remain realistic if real-world data growth, changing network topology or other factors are taken into account in an organization’s ability to meet its SLAs. In small and midsized businesses particularly, which often don’t express those SLAs in terms of recovery objectives – meeting them on their systems is one thing and what the business expects quite another. In other words, when it comes to setting RTOs and RPOs, getting on the same page and being prepared to find an alternate route to your destination is essential.

The metaphor Winkler uses – managers measuring backup and restore operations using a stopwatch mentality – seems apt. Because no matter how much time you think you have for backup and recovery it’s better to spend a few hours of planning ahead now in order to gain a few precious minutes later on when you may really need them.