Flat backup has major advantages over traditional backups, namely simplicity and efficiency, according to storage analyst W. Curtis Preston.
Flat data backup uses storage snapshots to protect data without using traditional backup software. The snapshots are copied to another location, while the originals still exist in their original location and native format. The replicas serve as backups in the event that the originals get lost or damaged.
Flat backups, snapshots and replication are near-continuous data protection, said Preston, of analyst firm Storage Switzerland, at the Storage Decisions conference in Chicago.
"The big thing about flat backups is that they are stored in the same format as the original," Preston said. So in a recovery situation, an organization can mount that backup as production data.
One of the major issues with traditional backup is in the restore process, and the time and energy it takes. Flat backup simplifies and quickens restores.
"That's what we call instant recovery. That's how you do a restore without doing a restore," Preston said. "I see it as the single most important feature of a backup product for the 21st century."
Another major problem with traditional backup is the creation of duplicate data, Preston said. That data adds up and can become unwieldy. In recent years, technologies and processes have materialized to combat the growth of duplicate data, such as copy data management and deduplication. But a process like deduplication isn't necessary if an organization is performing a flat backup.
"The only duplicate data to get rid of is duplicate data within the data set itself," Preston said.
Unlike traditional backups, a flat data backup does not require a backup server or a media server, which means it could cost less to implement. In this video, Preston explains how the technology can simplify your storage environment and make the recovery process a little less stressful.
Transcript - How flat backup can simplify your data storage environment
Editor's note: The following is a transcript of a video clip from W. Curtis Preston's presentation on flat backup at Storage Decisions in Chicago. The transcript has been edited for clarity.
Flat backup is not a new concept, it's just a new term. Flat backups, snapshots and replication are also known as near-continuous data protection.
CDP is continuous data protection, and CDP is truly continuous: Every byte is sent to the data protection system the second that it's created. With a near-CDP system, the idea is that, every once in a while, typically every hour, I'm going to create a snapshot on the source, and then I'm going to replicate to the destination the bits that have changed in between the last snapshot and this snapshot.
The big thing about flat backups is that they are stored in the same format as the original. There isn't a container. There isn't a transformation. With traditional backup, the data is transformed into something else, like TAR or CPIO or a proprietary format most likely from a system like NetWorker or Commvault. A flat backup is stored in the same format, which means that, in a disaster, I can simply mount that backup as my production. I can point my file server clients over to an IP address.
That's what we call instant recovery. That's how you do a restore without doing a restore. And this concept is really important because I see it as the single most important feature of a backup product for the 21st century. If you're a backup product and don't have this feature, it's time to either get out of the business or get this feature.
Flat backups are also more efficient at using storage. We are only, each time, sending the incremental bytes that have changed since the last snapshot. It's block-level, incremental forever, and if I don't need that snapshot anymore, I can delete that snapshot and the blocks that were necessary only for that snapshot to go away.
If I want to do hourly snapshots for the first 24 hours, and then delete 23 of those and keep just a daily snapshot, it's just an extremely efficient use of storage. If I turn on deduplication, I don't get much benefit because there isn't any duplicate data to get rid of. The only duplicate data to get rid of is duplicate data within the data set itself.