Rising numbers of organisations are recognising that backing up their valuable corporate data on tape is no longer adequate and are turning to disk-to-disk backup.
Uptake of disk has jumped to 21 per cent of the total market this year from a mere eight per cent last year and is tipped to rise.
Although disk-based systems were once considered too expensive, prices have fallen to such an extent that disk is on a par with tape. While an LTO-4 cassette holding 800GB of uncompressed data will cost you about £110, today you can buy a disk holding 1TB of compressed data for that.
Also, disk-based data is encrypted automatically, without the punishing performance overhead sustained by tape. In addition, information is sent digitally to a remote vault over a broadband connection so there is no danger of cassettes being lost or mislaid in transit.
The NHS, for example, is in the enviable position of having its own N3 fast broadband connection, which means that it can exploit the advantages of disk-to-disk systems with no additional outlay.
Tape also requires a lot of expensive manual intervention. Some businesses are forced to hire dedicated employees, who do nothing else all day but change, manage, index, catalogue, store and retrieve tapes.
Storing and retrieving data on disk-based systems, however, can be automated in line with pre-set policies. Administration and labour costs are reduced, not least because valuable resource can be deployed elsewhere.
Moreover, because information is stored in blocks, you don’t need to back up entire files whenever small changes are made. Instead, only the changes themselves are backed up, saving a lot of time.
Tape-based systems also can be unreliable. The disaster recovery industry has claimed that 40 per cent of attempts to recover data from tape fail. Research company Gartner has put the figure as high as 72 per cent.
Furthermore, many organisations do not adhere to manufacturer recommendati ons to replace their tapes every three months. Magnetic media such as tapes degrade as a result of exposure to normal atmospheric conditions.
Tapes also have a habit of stretching over time.
With disk-to-disk environments, reliability can be measured in years rather than months, particularly if systems are set up in resilient RAID 5 or 6 configurations.
Also, because data is written to a hard drive, it can be recovered even after accidental overwrites, as an imprint remains - and such recovery occurs at much quicker speeds than was traditionally the case.
This is because, like CDs, disk drives write and read data in a linear fashion, which makes information quicker to retrieve than if held on tape in a typically random format.
You also do not need to wait impatiently to get hold of tape catalogues that are both held in distant locations and likely to contain data from only as far back as the night before disaster struck anyway.
This situation, combined with the hours required to copy information from tape to repaired or replaced servers, means that a restore typically takes between two and five days. Such timescales are no longer acceptable.
Because disk-based backups can be done as often as desired, the recovery point objective can be pre-set to hours or even minutes before trouble hit.
Restoration of data takes minutes as it is simply sent from a remote vault over a broadband connection and so recovery times can be shortened.
Jon Leary is a consultant at CSA Waverley
HPE CEO talks up RedPixie acquisition and indicates DRAM crisis is over on Q2 results call
Reseller expecting to beat expectations this financial year
Data storage vendor announces former Commvault executive as European channel director
Chinese parent reportedly looking to reduce debt after mega acquisition spree