The ability to access meaningful data in a timely, efficient manner is essential where access to customer information is critical. Of equal importance is the ability to move and share data throughout the organization - among departments, offices, and business partners. The time it takes to recover from a failure has to be shorter than ever to avoid the very serious consequences of impacted productivity and profitability, damaged current and future revenues, and lost confidence of customers.

One way of protecting data is to duplicate it in real time to another system - using Data Replication software. In the event of one system failing, the standby system can come online ( automatically if LifeKeeper cluster manager is used) to provide continuation of service, with minimal (if any) data loss occuring.

Data replication can either occur locally over a LAN (removing the requirement for expensive shared storage) or remotely over a WAN (so forming part of a disaster recovery solution).

Data Replication can be used in one of three ways:

• To back up data real time to another server
• With LifeKeeper, to provide failover when no shared storage is available. Use of data replication software can drastically reduce the cost of a highly available solution by eliminating the need for shared storage.
• As an addition to LifeKeeper cluster manager to provide an offsite copy of the data for disaster recovery.

Data replication features

Block level replication takes place over a TCP/IP link to a second server. The servers can be connected via a LAN or WAN link. As the replication is at block level, only changed data is replicated over to the second server. Avoiding the need for the additional traffic of file level replication.

LifeKeeper Data Replication integrates fully with LifeKeeper clustering so it can be managed through the GUI together with any other application software that is being protected. Use of data replication eliminates the need for shared storage as data is always up-to-date on the backup server.

Mirroring over LAN and WAN to multiple targets

It is possible to extend the mirror to replicate to more than one target. The benefit of this is that a local and remote replica of critical data can be produced which will further extend the resilience. A local replication (and failover) for use when there is a local disruption to the active server, for example, if there is a failure of the service or the server. A remote, off-site replication (and failover) for site disasters.

Behaviour of the mirror

Mirroring is from source to target and can be paused by the administrator for a number of reasons:

• To perform a backup or other operations of the data that would require the data but would affect the performance of the active system.
• To alleviate traffic in periods of peak usage of the network.

The mirrors can be re-activated at a convenient time and the data will automatically resynchronize. LifeKeeper keeps a bitmap of changed data on the mirror. If the mirror is broken and resynchronization is required, then only the changed data is resynchronized.

Connection

Replication takes place over a standard TCP/IP link, and so can easily be tunneled through a VPN to provide security and encryption and can be used across a firewall simply by opening up the correct ports. Because, after its initial syncronisation with the backup server, LifeKeeper Data Replication only replicates the blocks that have changed on the disk the bandwidth it requires is proportional to the amount of data that changes on the source system, however if the required bandwidth is unavailable for a period the mirror will pause and can be restarted easily when the network quietens down.

Disk to disk backup

Data replication can be used as a stand-alone product to produce a real-time block level backup of a volume or partition. The backup is to another server on a LAN or off-site over a WAN. Disk to disk backup is lower cost than a combining data replication and failover, but still leaves the option of adding failover at a later date. It can also be used as a means of consolidating backups of multiple servers onto one node.

Other Uses

While disk to disaster recovery is the most obvious use of disk to disk backup it can be used with off-host processing - movement of data to a seperate server from where it can be backed up to tape, have reports run on it, tested in a variety of ways, or any other task that requires production data but that would effect the performance of the production system.

Original Source: www.openminds.co.uk