Climate Smart backup
We are going to talk about how to become a Green IT environmental energy efficient data center providor.
Traditionally backups of applications are performed by scheduling periodically full copies typically once per week, and the rest of days by sending incremental changes.
When the data is small, this is not of an issue, and takes several minutes to few hours to complete.
But what happens if the data grows? and the size becomes larger?
The backups will no longer take minutes to hours, but instead hours to days to complete the backup of the data.
System administrators has demand from business to secure the data as fast as possible, with as little impact as possible.
This normally means that the backup process has to be taken outside normal business hours, when all other maintenance activities also should be done. For example update statistics, rebuilding indexes, reorganization, etc.
The system administrator has to add more resources, such as memory, cpu, faster disks, faster network etc in order to complete the backups within the backup window.
This consumes a lot of resources, and is not energy efficient.
So, is there another way to save energy?
Climate Smart Backup
Taking application consistent snapshots, and by reading the built in operating system journal that keeps track of all block changes, one can perform progressive block level incremental backups (allways incremental backups).
Why backing up all data again, if the data has not been changed?
The fastest way to backup data is to not back it up (again)
The more often a backup is performed, the less amount of data to be sent.
Isn’t this what you want?
Energy Efficient Storage
Saving data on the media that meets the requirements.
If the data is not frequently used, why keeping it electified?
Storing data on tape, or CD WORM disks for data that are not frequently used saves energy.
Reusable Storage
Restore of data are often performed in several steps. First restoring last full copy, than one or several incremental copies, and lastly a restore and recovery of transactional data which can be very time consuming.
This results in that there are more data to be copied back than the actual size of the application data (full + incremental + transactional data) > full copy.
This is not energy efficient, and business often wants this to be performed as fast as possible, which also includes more resources, faster network, more memory, more cpu, faster media etc.
Restoring just a portion often requires to restore all data (full + incremental + transactional data) to a new server, where export can be done to the server with the missing data.
Is there a better way to restore data, which only requires one full copy, and with as little transactional data as possible?
The technique is called Instant Recovery
With instant recovery, the backup is provisioned from the backup system as it looked at that particular point in time of interest. The backup volume is connected as a snapshot, and the operating system starts immediately to revert to this snapshot in background while the volume is fully accessible.
This shortens the restore time to a few minutes.
After all, why waiting for all the data to be restored if you can use it immediately?
It is like a coffe machine that brews coffe, while the coffe is brewed, one can drink it.
The application can be started immediately, and if a recovery is needed, than the recovery can also be performed while restoration are done in background.
This speeds up restoration and recovery process, as it is done at the same time.
It also speeds up restoration of small portion of data, as one can start exporting the data of interest, and cancel the restoration before all data has been fully restored.
Improving data protection
Backing up transactional data are often performed on schedule basis, typically once per day, or every 4 hours…
This means that the maximum data loss can be up to 4 hours, if the backups are taken every 4 hour.
Is there a better way to improve this?
The technique is called contineous backup
By letting the backup to copy transactional data immediately will improve the data protection.
This will reduce the maximum data loss
If you add both of these solutions, than you can improve your data protection strategies, save energy resources, become a Green IT environmental friend, save money and time.
Comment if you like.
Here is how spictera position us on the market.
Contact us if you want to know more…
Follow us on Twitter: http://twitter.com/spictera
Follow us on LinkedIn: http://LinkedIn.com/company/spictera
