Network Computing is part of the Informa Tech Division of Informa PLC

This site is operated by a business or businesses owned by Informa PLC and all copyright resides with them. Informa PLC's registered office is 5 Howick Place, London SW1P 1WG. Registered in England and Wales. Number 8860726.

The State Of Enterprise Backup

  • A popular catch phrase among technology pundits these days is "backup is broken." While it's easy to argue that backup is broken, the reality is that no single technology in today's data center is as important to continuing operations. It might be true that traditional file-by-file backup to tape is too slow and inflexible to meet the modern data center's dynamic needs. However, that doesn't mean that backup technology as a whole is no longer relevant.

    At the same time, backup vendors trumpet how wonderful each new technology, or even each release, is. If you don't roll out the forklifts, you're falling behind.

    We wanted to cut through the hype and overused phrases of the vendors and pundits to see how real IT teams are protecting their data. We polled 437 IT pros in our third InformationWeek Backup Technologies Survey; all of those surveyed are involved with the evaluation, selection, or use of backup technologies.

    Contrary to the popular meme, the survey showed that backup is far from broken. IT pros are pleased with their backup systems even while they might not be diligent in protecting their virtual machines and remote offices or performing enough tests. Let's take a look at the survey findings.

    (Image: geralt/Pixabay)

  • Backup satisfaction

    Our survey found a surprising level of satisfaction among respondents in the state of their backups, with 84% reporting that they're very (36%) or somewhat (48%) satisfied. Refuting the "backup is broken" meme, only 6% of those polled expressed dissatisfaction. With that level of satisfaction, you'd think these IT pros would be very confident in their ability to get their data centers back up and running after a disaster, but only 65% said they were extremely confident or very confident in their recovery capabilities.

  • Backup applications

    Most IT shops prefer to use a single application and method to back up all their data. Our respondents agreed, with almost half using one backup application and 42% using two. Also, 59% of those surveyed backup directly to tape, down from 66% in 2013, and 54% expect to still be backing up direct to tape in 18 months.

  • Virtual server backups

    New computing platforms are generally accompanied by new backup applications. As virtualization emerged as a new platform, a new class of backup applications arrived to support it. Just 25% of respondents said they use different backup software for their virtual and physical servers, down two points from last year's survey. It's not just the shift to VM-specific backup applications that's sluggish; the shift to VM-specific backup techniques also is on the slow track. Forty-eight percent continue to back up their VMs as if they were physical servers, with agents in each VM. Less than 20% use more sophisticated and specialized techniques such as software with changed block tracking or source deduplication capabilities.

  • Backup frequency

    When we first started doing these backup surveys, some of our respondents were definitely treating their virtual servers as second-class citizens. In 2011, only 8% of respondents said they backed up less than half of their physical servers weekly, while 22% said they left half of their virtual servers at best unprotected. In this year's survey, that percentage has fallen to 14% for VMs while the physical server rate remained 8%. While this may seem lax, survey participants had pretty good reasons for not backing up those VMs. After all, if you're using VMs for testing, like 44% of those with unprotected servers say they are, you can always copy the production data again to reset the test systems.

  • Remote & branch office backup

    While backup administrators have started treating their virtual servers more like first-class citizens, the same can't be said for servers in remote and branch offices. In fact, the percentage of respondents saying they don't back up remote office data has grown from 43% in 2013 to 46% this year. It's become easy to set up a centrally managed backup infrastructure for remote and office branch data, so we have to assume politics rather than technology has led us to the point where almost half of our respondents think it's acceptable to leave remote office data unprotected. One other explanation is that some respondents have adopted cloud services to the point that their remote offices don't contain any unique data.

  • Testing

    A truism of coaching football, along with "Winning isn't everything, it's the only thing," is "You play the way you practice." That's also true of the backup world. Despite all the technological advances, it's frequently true that restoring a major application involves multiple interactions and tasks, the proper order of which you may not discover until you actually attempt a restore. Consequently, we were somewhat disappointed that the percentage of respondents reporting that they test their backup processes at least once a year fell from 44% last year to 41% this year.

  • Encrypted backup tapes

    Likewise, we're disappointed in how much data survey participants reported storing on removable media without the protection of encryption, despite all the reports of data breaches caused by lost or stolen backup tapes. Just 31% encrypt all removable media; 42% encrypt more than half. A lot of grief and expense after a data breach could be avoided by following one simple rule: Any media leaving the premises is encrypted, period. State breach notification laws explicitly exclude data stored on encrypted media from their notification requirements. Plus, it's not like encryption is that difficult; all but the most basic backup software has the ability to encrypt the data it writes to tape or disk.