The challenge of managing virtual and cloud environments as well as physical ones is adding complexity to disaster recovery (DR) efforts -- sometimes resulting in inadequate protection. That's the conclusion of the sixth annual Disaster Recovery Study from Symantec, a Mountain View, Calif.-based provider of security, storage and systems management solutions.
The global report found that 44 percent of data on virtual systems is not routinely backed up, and only 20 percent of respondents use replication and failover technologies in virtual environments. Sixty percent of virtualized servers are not in the DR plans of those surveyed, an increase from 45 percent in last year's report.
The Challenge of Multiple Tools
The report found that "using multiple tools that manage and protect applications and data in virtual environments causes major difficulties for data-center managers." It found the effort to be a "large challenge" for 58 percent.
Dan Lamorena, director of Symantec's Storage and Availability Management Group, said his company expects to see organizations "adopt tools that provide a holistic solution with a consistent set of policies across all environments." The keys for data-center managers, he said, are simplification, standardization and treating "all environments the same."
Toward those aims, Lamorena recommended using integrated tool sets, simplifying data-protection processes, prioritizing planning and tools to "automate and perform processes which minimize downtime during system upgrades," identifying issues as quickly as possible, and avoiding shortcuts that could have "disastrous consequences."
Since the study found that about half of mission-critical applications are being run in the cloud, inadequate DR planning poses a major problem. The respondents recognize that fact, with 66 percent reporting that security is their chief concern about using the cloud.
Backup Frequency, Outages
About 80 percent of backups take place weekly or less frequently, instead of daily. Resource constraints, including people, budget and space, were the biggest issue, with 59 percent seeing it as the key challenge.
Downtime after an outage was twice what survey respondents thought it would be. Asked how long it would take to recover from an event that destroyed their main data center, respondents said they expected downtime to be two hours before they were up and running. In the last such survey, taken in 2009, those responding estimated four hours.
But Symantec found that the median downtime for each outage in the last 12 months was five hours, with an average of four downtime incidents for each organization in that time period. Most frequently, the outages were caused by system upgrades (72 percent), followed closely by power outages and failures (70 percent). For 63 percent, outages were caused by cyberattacks.
'Interesting But Not Greatly Surprising'
Al Hilwa, a program director at IDC, called the survey results "interesting but not greatly surprising." The findings, he said, reflect that one of the core value propositions of virtual and cloud-based computing is to consolidate "many smaller workloads for better hardware utilization." Because of that factor, he said, "it makes sense that there is a pattern of lower DR attention and inferior practices around such workloads, compared to larger workloads."
But Hilwa contested one possible interpretation of the study. "Correlation is not necessarily causation," he noted, in that cloud and virtual environments by themselves do not "necessarily" complicate DR efforts. "There are many effective tools for managing DR in a virtual environment," he said.
Hilwa pointed out that the migration of many separate apps and content to a single virtual or cloud environment is not more complex because of the environment, but because there were separate DR plans -- or none at all -- for those individual items in their former homes.