No, I will not pretend that the backup is dead! It is just becoming something else.
The old days:
Traditionally, a backup solution involved some kind of media (tapes, disks) where you copy the important Prod data and send the media for safe-keeping in an off-site vault. The restore process, which is the reverse of the backup process, meant recalling a tape or a set of tapes, and restore the data on the original place or an alternate restore point. Restoring large amount of data was not a quick procedure, several hours if you were lucky. It is one of the reasons applications owners and DBAs were keeping several sets of Prod data on the Prod storage, just in case!
The amount of data grew exponentially in the last five years, and, thankfully the technology has advanced in a similar fashion. The backup concept is gradually replaced with the total data protection and management concept meaning that the data will be analyzed once and it will be determined which data to backup, which to archive and what amount of indexing is needed for search purposes. The majority of the storage arrays have snap shot capabilities and with an intelligent backup solution integrating and controlling the snap shot process, the backup and restore are much faster and require less space on the arrays. The synthetic backups and “incremental forever” method together with deduplication and global deduplication, are massively reducing the size of the backups as well.
On top of this, the backup to the cloud as a service is offered by all major cloud providers, although if your backup is a few hundred terabytes in size, the bandwidth between the Prod site and the cloud can be prohibitive.
Based on the above facts, an enterprise backup solution (when properly implemented) is a robust and very dynamic data protection and management solution, with very fast backup and restore Server Level Agreements (SLAs), with load based variable backup windows and eliminating the duplicated data at an enterprise/global level, with an archiving and indexing component smartly allowing the archiving of legal documents and messages for the 7-10 years’ period, etc. because remember, backup is for operational recovery not long term retention of data!
The logical step in data protection will be the Data Protection Architecture concept allowing the data centers do deliver data protection as a service with the flexibility to allow different modules to be plugged in at the appropriate time to address specific user or application requirements. With a data protection architecture there will be no lock to a single application or storage device, but a mixture of various data protection tools from stand-alone software, to storage systems and dedicated backup appliances to be used as a service and managed as a single entity.
A Data Protection Architecture solution should be flexible, with data protection utilities being able to write direct to the storage, should keep the data in its native format and have a governing body to make sure all the required components work together to simplify management. The goal will be to eliminate the user interface from the data protection tools and applications and let them embed into the central data protection architecture framework.
As stated at the beginning of this article, backup is not dead and will not be for a long time, it is just that backup as we know it is evolving into a data protection architecture.