Crossing the Zetabyte Boundary

A zettabyte is the really large measure of data which represents 1 trillion gigabytes, or 1 billion terabytes.

And, the world's online data is crossing the ZB mark for the first time this year.

According to IDC's annual survey of global digital output (published in May 2010 and sponsored by EMC), the current total amount of global data is expected to pass 1.2 ZBs sometime during 2010. To put this in context, IDC compares this as being equivalent to the amount of data that would be generated by every individual person in the world posting messages on Twitter continuously for a century.

IDC also reports that the world's digital information will grow by a factor of 44, between 2009 and 2020, at which time it is expected to reach 35 ZBs. IDC calls this an incomprehensibly large number, characterized as a perpetual data tsunami.

John Gantz, the chief research officer working at IDC on the survey also says "Good luck, all you CIOs out there!"

I found this all rather interesting, in that IDC also reports that the gap in protection will also grow over this time: there is just too much data - and too much growth of data - to protect it well using current backup and recovery methods. The storage costs alone will overrun global IT budgets. In the same period in which data is expected to grow by a factor of 44, IT staffing and investment to manage data is expected to grow by only a factor of 1.4.

Gantz ends his analysis, after sharing these facts, by stating: "Have courage!"

A bit over-the-top for my taste, to be honest. But the point is made: we need new protection methods or IT will be swamped, and critical data will not be protected.

How can Image-Based Methods Assist?

Image-based data protection can help to solve this problem, by eliminating the need for extra backup passes of the same data. Our recent introduction of vConverter 5.0 makes it possible to apply image-based protection methods not only to virtual systems, but also to physical systems. Performing image-based backup enables all types of recovery - including:

  • FLR - file level restore, meaning that individual files can be selected and restored from the backup image
  • OLR - object level restore, meaning that individual application objects like Exchange email messages and folders, can be selected and restored from the backup image
  • FlashRestore - which is instant recovery of full virtual machines by booting them from the backup image in the repository rather than waiting for a restore process to occur

If all of these recovery types can be gained from the same backup image copy, then various types of backup copies can be eliminated altogether - including so-called brick-level backup of email databases, file system backup, bare metal recovery tools, and more - all without compromising the recovery levels. Eliminating these 'extra' backup copies helps bring more of the world's data under management, by speeding the backup process, using networks and storage more efficiently, and by lowering the administrative burden.

About the Author