Data replication provides a continuously updated copy of critical data at a remote site. In the event of a disaster, users are rapidly failed over to the live backup systems at the remote site, minimising data loss and greatly speeding recovery.
Disk-based recovery is more reliable, less complex and takes less time than tape-based backup, improving the RPO and RTO of the disaster recovery solution.
Data replication has long been considered an impractical solution for data protection and recovery, especially for small or medium-sized business with limited resources.
Historically, it required expensive hardware and a large investment in bandwidth to replicate and protect data in real-time. However, the recent evolution of replication solutions is starting to dispel this belief.
Within software-based data replication, there are two major technologies: synchronous and asynchronous replication. It’s important to understand the benefits and drawbacks of each.
With synchronous replication, the software intercepts data being written to disk and sends it to both the primary and secondary disk arrays at the same time.
Only when both arrays confirm receipt of the data does the software accept another write.
With synchronous replication, data loss approaches zero because both the primary and secondary disk arrays must contain the same data.
However, the confirmation process can cause performance problems, especially in applications that process lots of transactions.
Acceptable performance of synchronous solutions often requires connecting the arrays with high-bandwidth fibre channel, which is very expensive and has an effective range of about ten miles.
As a result, synchronous replication is most often used to create a local backup of data in situations where having an exact copy of the data is essential.
With asynchronous replication, the replication software grabs data once it is written to disk, and rewrites it to a second array.
The application doesn’t have to wait for confirmation so it can operate with little or no impact on application performance.
It can also use low a bandwidth connection over any distance.
Asynchronous replication provides enterprises with a very high degree of data protection and makes the most cost-effective use of existing infrastructure.
While asynchronous replication can’t deliver the zero data loss available through synchronous replication, it can be configured to deliver RPOs and RTOs that are more than acceptable for most business uses. The combination of excellent data protection, minimal performance impact, long-distance effectiveness and low-cost deployment makes asynchronous replication an ideal solution for backing up data to a remote recovery site.
There was a time when tape-based backup was widely believed to be the only feasible backup solution for remote offices.
However, relying on tape backup is expensive, complex and cause a company to incur additional expenses related to recreating lost data.
In the worse case, relying on tape backup can significantly impact the ability of a business to continue trading after a disaster.
Advances in technology have made centralised backup through data replication easier to manage, taking the responsibility for backups out of the hands of non-technical resources in remote offices, and giving it to the experts back at a central data centre.
Replicating data to a central location from branch offices can reduce the per-location costs associated with tape-only solutions and provide a higher level of recoverability for business-critical systems and data.
Augmenting or even replacing a tape-centric data protection scheme with continuous data replication, can have a significant positive impact on protection budgets at the same time as providing a far higher level of protection and recoverability than tape.