A data backup solution for any business organization is one of the most important components of IT infrastructure to protect the most vital asset—data. Backing up of data affects application performance. To avoid this adverse impact, enterprises choose data backup windows in such a manner that backups have minimal impact on the application’s performance. They try to keep the backup window to the minimal possible duration, that too in off-production hours. Beyond this common practice, data backup windows can be kept to the minimal possible duration in the following ways.
Tip 1: Categorize backup clients according to the criticality
Backup clients should be categorized according to their criticality in terms of application performance requirements. SAN-based backups should be implemented for the critical servers to minimize the data backup windows for critical servers.
Tip 2: Deploy a dedicated backup LAN
A dedicated backup LAN should be deployed for LAN-based backups, so that the backup data travels through a dedicated LAN to the backup device. This can be achieved by installing a separate network interface card (NIC) on the backup clients.
Tip 3: Maintain consistency of flow control settings
All members of the backup infrastructure—backup clients and switches—should be either kept at full duplex, half duplex or auto settings. Maintaining consistency of the flow control setting across backup infrastructure can considerably reduce data backup windows.
Tip 4: Choose proper backup schedules
Backup schedules should be chosen in such a way that on weekdays you back up only the data which has been changed since the last backup. This is due to the fact that you can’t afford long data backup windows during production hours (as you can on weekends). The most effective schedule is to opt for a weekly full daily incremental, that is, full backup once a week and incremental on weekdays. Similarly, you can opt for weekly full daily differential.
Tip 5: Archive data
It has been found that organizations back up the same data again and again, even data that has not been changed since long. Hence the backup window keeps increasing with addition of new data. An archiving solution should therefore be implemented so that it archives data which has not been changed for a certain amount of time to a secondary tier of storage. This reduces the amount of data to back up, and thus directly affects the duration of your data backup windows.
Tip 6: Use storage array-based snapshot technologies
Storage array products have a snapshot feature which can be used to create an instantaneous point in time (PIT) copy, that is, backup of the source logical unit number for which we need to take the backup. We can assign the PIT copy to a backup host and take the backup on tapes (if needed), without any effect on the application’s performance.
Tip 7: Deploy de-duplication-enabled backup software solutions
Software solutions that perform de-duplication reduce the data backup windows by a significant duration (because they back up only the changed data blocks). Consequently, all the data need not be backed up again; also, the data need not travel on the LAN to the backup device. Software which supports variable-length de-duplication reduces backup windows considerably because the probability of finding common data segments is more.
Tip 8: Backups on disk
Tiered backup infrastructure should be deployed if possible. By tiered backup infrastructure we are referring to the practice of backup performed on a disk-based device. This practice is advantageous for reduction of data backup windows since higher write throughputs are achieved on such devices. Later, the data should be tiered to tapes during production hours. This will reduce the time required for backups; the data can be vaulted to tapes later and the disk storage can be freed for the next backup cycle.
Tip 9: Integrate VTL into your existing backup infrastructure
The virtual tape library (VTL) is a disk-based library which emulates a physical tape library to the backup software. Due to this characteristic, you can use a VTL without making significant changes on the backup software. Since the VTL uses disk drives as storage media, backup data is written to disk (that too on multiple disks at the same time). So the write throughput is higher and backup time comes down in a noticeable manner. Most VTL vendors have integrated target-based de-duplication in the VTL software code; this reduces the backup time further. Opt for a VTL which performs the inline de-duplication—that is, it de-duplicates data on the fly in the RAM. You should also look for variable length de-duplication features in the VTL under evaluation.
About the author: Anuj Sharma is an EMC Certified and NetApp accredited professional. Sharma works for an EMC Velocity Partner, and in his current role handles implementation projects related to SAN, NAS and BURA. One of his articles was published globally by EMC, and titled the Best of EMC Networker during last year’s EMC World held at Orlando, US.