NewsCSO Online

    Commvault rolls out an automated recovery mechanism, promising faster restoration of business

    Share
    Commvault on Tuesday rolled out its answer to enterprise backup restoration nightmares. Called Recovery Studio, the vendor promised that the product would “orchestrate the recovery of applications, identities, configurations, infrastructure dependencies, and datasets together in an automated, runbook-driven workflow.” Analysts and consultants applauded the effort, agreeing that enterprises today struggle with fast, accurate and secure restorations. But they also stressed that the new system, which is targeted for early access in December with general availability “in the coming months,” places the onus on IT to keep a dependency map current. Recovery Studio isn’t using any form of AI to figure out proper restoration details. It relies on a map of assets, dependencies, and other relevant information that the customer creates and maintains, and for many enterprises, the challenge of keeping such a document current is the real problem. Even discounting the times when a line of business doesn’t tell IT about an application change, IT teams struggle to keep up with the speed of change in the organization. Commvault’s news release said that its automation allows businesses to “say goodbye to manual and painstaking attempts to restore complicated business services in a fragmented fashion that can result in downtime and data loss.” It pointed out that a large enterprise may need to restore hundreds of applications and thousands of identities, including the configurations and underlying data, all in the right order at the right time. “It’s no wonder that with conventional approaches, recovery to even a minimal level of operation takes an average of 28 days,” the release said. Commvault said Recovery Studio can manage a wide range of components that go beyond containers, VMs, files, and objects, extending to identity systems, application configurations, infrastructure configurations and security and network settings. It also promised that the Recovery Studio can standardize “across all recovery scenarios” including “operational, disaster and cyber recoveries. Unified runbooks automate infrastructure rebuilds, workload recovery, and the orchestration of security controls, such as firewall rules and network security groups and settings – all actions needed to rebuild the business, not just recover data.” The news release said that the new application can deliver other cybersecurity assistance via data security checks, including “automated threat scanning for supported resources in the runbook to identify compromised areas and inform recovery decisions before returning services to production.” Commvault did not provide a spokesperson to offer further details before publication. Depends on IT awareness of the environment A number of factors will influence the effectiveness of the solution, analysts said. Frank Dickson , principal analyst at Dickson Research, pointed out, “nothing in Commvault’s announcement describes a generative model making recovery decisions. This is runbook automation. A runbook is a recipe: do this, then this, then this. It does not hallucinate. It does exactly what it is told.” “That is the real risk. A flawed runbook gets executed flawlessly, at machine speed,” he said. “Automation does not make a bad plan good. It makes a bad plan fast.” That means, said Aman Mahapatra , chief strategy officer for New York City-based technology consulting firm Tribeca Softech, whether Recovery Studio ultimately helps or hurts enterprise restoration efforts depends on how aware IT is of everything going on in its global footprint. “How much it solves depends on something no vendor controls, which is the accuracy of the dependency map underneath it. A runbook is a prediction about your environment and every change since it was last tested makes it a little more wrong: a new application, a rotated certificate, a firewall rule added on a Friday,” Mahapatra said. “Most enterprise dependency maps are aspirational,” he noted, “and CMDBs in particular tend to describe the environment someone intended to build. Commvault’s figure of 28 days to reach minimal operations is plausible, but the gains will go to organizations that refresh and rehearse their runbooks continuously. For everyone else, automation will execute an outdated plan faster.” Abbas Jaffery , principal advisory director at Infotech, agreed. “My concern is that automation can execute the wrong recovery sequence just as efficiently as the right one. Enterprise environments change constantly,” Jaffery said. “Applications are upgraded, infrastructure moves, identities change and integrations are added or removed. If the dependency information behind a recovery workflow is outdated, automation could restore systems successfully but leave the business service broken.” Defined recovery Mike Leone , a VP and principal analyst at Moor Insights & Strategy, added that Commvault is still advancing the space. Recovery Studio “follows defined recovery runbooks rather than a model inventing steps on the fly. That’s much more deterministic and structured,” Leone said. “Sure, automation can repeat a mistake faster than a person can, and a runbook that’s gone stale is a fair concern. But Commvault is validating each piece as it rebuilds and scanning for threats before anything goes back into production. Those are the kind of safeguards and checks everyone should want. The alternative is people rebuilding systems by hand in the middle of an outage, and that isn’t exactly mistake-free either.” But Dickson stressed that automation is not necessarily the perfect fix. “Commvault says the runbooks validate components and can scan supported resources for threats before services return to production. Good,” he said. “But the honest comparison is not automation versus perfection. It is automation versus exhausted people working from a spreadsheet at 3 a.m. Humans get the order wrong, too.” Johnny Yu , an IDC research manager, is not as impressed by the components of the solution as he is by the way Recovery Studio organizes them. “I do see Commvault Recovery Studio as significant, but not from a capability standpoint,” Yu said, adding that all of the product’s components, including recovery runbooks, threat scans, and recovery testing, are already offered by Commvault, as well as by various competitors. “I think the significance comes from how Recovery Studio organizes different recovery workflows for different applications onto a single platform or hub, kind of like a library of validated recovery runbooks to be used for all kinds of scenarios.” He explained that he liked how Studio standardizes recovery runbooks to follow a similar workflow to that of manually rebuilding the environment, recovering the components, such as data, applications and agents, and “validating the recovery before bringing it back into production, but allows recovery nuance for each runbook, such as if you want to do in-place vs secondary-site recovery or whether threat scanning is necessary, such as DR scenario vs cyber recovery scenario.” Jaffery noted, however, that while the promised Commvault cybersecurity controls are attractive, they may also be insufficient. “Following a cyberattack, organizations need to be confident they aren’t restoring compromised identities, configurations or malicious persistence mechanisms,” Jaffery said. “I would not advocate fully autonomous recovery for every scenario. There needs to be appropriate human authorization for high-impact actions, particularly those involving identity, security controls and reconnecting systems to production. And organizations need to test the entire recovery process, not just whether the backup can be restored.” An AI agent recovery plan Commvault on Tuesday also introduced AgentRecover capabilities to help track and control autonomous agents. It would specifically “help organizations audit agent activity, restore agents, and recover the data and systems they affected across environments.” Mahapatra noted that AgentRecover addresses a potentially more difficult problem than system restorations. “Restoring an agent’s configuration is straightforward. Unwinding what it did is not, because an agent’s actions land as side effects in other systems, and some of them, such as an email sent or a payment initiated, cannot be restored at all,” Mahapatra said. “Traditional recovery rolls back to a point in time,” he explained. “Rolling an environment back to 9 a.m. erases everyone’s legitimate work since 9 a.m., while rolling back only what the agent changed requires knowing causality: which agent, acting under which identity, changed which record. Commvault’s mapping of agent activity and the identities agents act through points in the right direction.” Jaffery added that rollbacks of an autonomous agent can be especially daunting. “Restoring the agent itself doesn’t necessarily undo what it has done. An agent might modify customer records, change permissions, invoke APIs or trigger downstream transactions,” he said. “You need to understand what actions it performed and what the consequences were before deciding what to reverse or restore.” Leone also noted that agents can complicate system restoration, partially because of their speed. “Agents don’t wait for anyone to confirm things are safe. The moment a system comes back, agents start hitting it, far faster than any person would,” he said. “If an agent comes back before the data and identities it relies on are verified, it starts acting on stale or incomplete information at machine speed. Get the order wrong and one outage quickly turns into a second one.” Jaffery also observed that Commvault is far from alone in the automated restoration space, and may even be a tad late. “I don’t think Commvault has an uncontested position here. Rubrik, Cohesity and Veeam have all been investing in cyber recovery, application recovery and orchestration,” he said. “Rubrik, in particular, has been pushing toward automated business application recovery and dependency mapping. Cohesity has capabilities around cyber recovery orchestration and cloud application reconstruction, while Veeam has an established disaster recovery orchestration offering.” This article originally appeared on NetworkWorld .

    More on restoration

    Other advisories

    Other news entries

    Source: https://www.csoonline.com/article/4247124/commvault-rolls-out-an-automated-recovery-mechanism-promising-faster-restoration-of-business-2.html

    ID