After more than a decade of stagnation, the data recovery market is going through a pervasive transformation, says one industry expert.
The most visible aspect of this is a clear shift in focus from backup to recovery, said Ray Paquet, vice-president and research director at analyst firm Gartner Inc. in Stamford, Conn. Paquet was addressing a press contingent Monday at an international media conference hosted by Cupertino, Calif.-based Symantec Corp. in New York City. In the end, backup doesn’t matter, recovery does. Ray Paquet>Text
In Paquet’s view, the change is both welcome and long overdue. “In the end, backup doesn’t matter, recovery does,” said Paquet, noting that it has taken an inordinately long time for businesses – big and small – to realize this fact.
He said while the new license spend in this market has been impressive – US$2.5 billion in 2005 alone – the market itself has been remarkably stagnant, and the bulk of that expenditure has been on backup – mainly backing data on to tape. “The truth is you can backup all the data you want but if you can’t recover it, it doesn’t do any good, because it is recovery that end users actually seek.”
But all that is going to change quite dramatically, and quite soon.
By 2008, the analyst said, enterprises will be able to recover data to any point in time, as opposed to today, when recovery time for most logical errors is 24 hours.
Self-service recovery
Going by current recovery market trends, Paquet said, self-service recovery will be a reality by 2008-09. “Instead of IT having to deliver recovery, end users will be able to do their own recovery using self-service mechanisms.” He said eventually businesses would opt for Single Instance Store (SIS) as a means to store and recover data more efficiently.
SIS and self-service recovery, in turn, will be the fallout of the move towards continuous data protection (CDP) – a trend we will witness in 2007-08. With CDP, the focus is not just on disk, but also on ensuring more granular recovery, he said.
According to Paquet, there’s a clear move right now towards using disk to disk (rather than tape) as the primary recovery mechanism, so as to be able to recover even more significant points of time. This trend, he said, would continue through 2007.
According to Paquet, the growing popularity of disks has to do with price as well as performance. Serial ATA disks are driving disk prices down, he said, making the disk a more affordable recovery mechanism. In addition, it’s a known fact that backups often fail and “disks decrease the frequency of backup failure,” he said. “That’s because with disks you can backup data without actually moving it. With no operators involved, failure rates decrease.”
He said while in 2005-06 the spotlight is on recovery time, and the goal to be able to recover quickly, in 2007-08, the focus will be on “recovery points” – losing less data, and in 2008, on end user self-service for recovery.
The recovery of static data will remain the number one problem in storage up until 2015, the analyst said. “Our IT infrastructure is as old as it’s ever been. We are finding more uses for older equipment – rather than retiring it. This means that older equipment must also have a level of recoverability.”
Recovery in Gartner’s Top 10
Paquet related the results of the Gartner CIO Survey for 2005 that’s based on responses from more than 1,300 CIOs.
Their biggest priority, he said, is managing and reducing IT risks [and] business continuity and security are the key investment areas to reduce IT risks. The second focus area is developing a flexible infrastructure, and recovery will ensure that infrastructure is in fact flexible.
High on the list as well, he said is the desire to transition the IS organization from a functional to a services one. Paquet said CIOs are beginning to realize that recovery is in fact a service an IT organization can deliver to end users. “That’s because end users will accidentally delete a file and create a recovery incident for them.”
Bottomline: recovery is a critical aspect of all the top 10 priorities for CIO.
According to Paquet, the new regulatory environment is one of the big drivers of this renewed focus on recovery. “Compliance is a major driver of technology spend in general as well as specific to ensuring recovery. Associated with this is managing risk. IT organizations have to manage risk for reasons of compliance and general good practice.”
Replication vs. backup
According to Paquet, continuous data protection (CDP) is not a new market, but a constant form of point in time copy, which could be used in the backup or the replication markets.
The difference, he said, is replication doesn’t involve any sense of time and only addresses physical errors. “We can take an existing replication engine, put an indexing mechanism in with it and create continuous data protection.”
By contrast, he said, backup – whether by disk or tape – is time-based and can addresses logical or physical errors.
“But backup is unique in that it has a catalogue of everything that’s been moved – a catalogue of all the files.” Such a catalogue, he said, makes it easier for end users to deliver on self-service. “We do much better in providing self-service because end users aren’t going to think: ‘I need to recover this block; they’re going to think: I need to recover this file, or this application.’”
Recover less!
According to Paquet, a crucial but often ignored way of improving recovery is to recover less. “The reason is we store too much and never use it. In a typical environment, 70 per cent of the unstructured files are untouched after 90 days. Yet a full backup copies them every time. This means almost 70 per cent of our full backup costs are on stale data. If we can archive that data off, we can reduce our backup media costs for full backup by 70 per cent. In addition we’ve just improved recovery by 70 per cent as well, because we only need to recover the 30 per cent that we’re using right away.”
