Helios, LLC

May 1st 2012 - To begin, data replication in any form is primarily for the purpose of disaster recovery. Secondary to this purpose may be the desire to utilize copies of the data for various other purposes or sharing among multiple business units. Therefore, the most important factor in determining the best data replication strategy for your business must be its ability to protect you from data loss in the event of the unexpected and then secondary considerations can include functionality of the data for alternative reasons. The most common experience most computer users have with this concept is simply the performing of a routine backup (although I suspect most users don’t even take this basic precaution) which protects against the potential for a computer failure and specifically a hard disk failure. An alternative method which is growing in popularity is disk mirroring in which one drive is routinely copied to another drive in order to create a functional copy of the drive. External backup hard drives have made this very common with the cost of storage at an all-time low. Of course there are drawbacks and risks in either of these methods. One drawback is that backups and disk mirroring can cause performance degradation while in progress and depending on the amount of time required this can be unproductive. Both are also susceptible to viruses and malware and the addition of virus scanning during the backup process can further degrade performance to the point that the computer is practically unusable during the backup or mirroring routine. This makes incremental backups attractive because the only data needing to be copied is that which has changed since the last backup. The more frequently the backup is performed, the faster it becomes because of less opportunity for data to have been changed. Taken to the extreme, real-time or near-real-time backup schemes essentially replicate every data change locally and on the backup media as the changes occur. This is what is known as a synchronous operation. It is the Rolls-Royce of data replication solutions but also comes with a Rolls-Royce sized price tag. The importance then is to determine the level of protection you require which is cost-justified for your business needs. Cloud solutions can aid in reducing the investment cost of a synchronous or near-synchronous backup strategy but in reality the cost savings on the front end are simply deferred into ongoing operating costs. Cloud hosting services are priced according to capacity and reliability; add in the required speed and reliability of an internet connection for each location and the ongoing subscription costs will quickly rival the capital costs of an on-site solution. An added benefit of an off-site cloud solution is that it removes the IT demands of hardware and software administration, which when factored into the overall operational costs can make the cloud solution a better overall value. So how valuable is a truly synchronous data replication solution to your business? As an insurance policy against catastrophe - probably not as much as you may think. In a multi-location operation, the solution may not be in a cloud replication solution but instead in a cloud hosting solution. Both are equally dependent on a reliable internet connection, but whereas cloud replication has a slight advantage in that it can offer redundancy against internet failure with its local data source, it is also plagued with data integrity concerns in which the copied data and the source data can fall out of sync. When this occurs, there is no easy solution to reconcile which data source is the most trustworthy (known as the data of record or system of record) and must rely on predefined rules (rules of record) to resolve which data to trust as it replicates itself between the local data and the cloud data. When non-trustworthy data is acceptable then this is an acceptable scheme. However, very few businesses are willing to accept or tolerate unreliable data and even more so if that data is the sole source for disaster recovery. There would be no less acceptable scenario than to put your faith in a disaster recovery solution that proves itself unreliable right at the time when you must rely on it the most. A cloud hosted data solution doesn’t rely on data replication at all; rather, all users and locations simply access a single data source to ensure that the data is always current. This method is highly desirable but also elevates the risk of an internet failure because there is no local copy of data to fall back upon. Conclusion: There is no single ‘right’ solution. Varying factors will determine which solution is best for your business needs.

Share This with a Friend

Marketing Hub

Integrated Credit Card
Processing

Marketing Hub

Integrated Credit CardProcessing

Integrated Credit Card
Processing