Trends in Disaster Recovery
by Shirish Phatak on November 3, 2017
Disaster recovery plans have been a part of organizations' data management and availability strategies for years. However, disaster recovery techniques have evolved significantly. Here's a look at new disaster recovery trends.
What Is Disaster Recovery?
A disaster recovery strategy is the plan that an organization develops to ensure that business continuity is maintained when an unexpected event causes a major disruption to the organization's IT infrastructure.
The event could be a natural disaster that puts a data center offline, or it could be a cyber attack that brings down mission-critical servers. Even something as simple as a failed software upgrade could trigger downtime or data loss that, if left unchecked, will disrupt business operations.
Traditional Disaster Recovery
Traditionally, disaster recovery strategies were relatively simple. They usually involved a two-prong approach that consisted of:
- Data backups, from which data could be restored quickly following a disaster. The data backups were typically created by copying data from production servers to a backup location. After a disaster, the data would be manually restored.
- A plan for bringing up new servers in the event that a disaster disrupted production servers. Although setting up a recovery plan ahead of time helped to streamline recovery, standing up replacement servers usually required manual work.
New Trends in Disaster Recovery
Today, organizations are moving beyond this basic approach to disaster recovery. New trends are enabling more efficient, reliable, and automated disaster recovery strategies. They include:
Automated data replication
Instead of backing up data to dedicated backup servers or an offsite location, some organizations now use software-defined storage systems to replicate data automatically and directly on production servers.
Essentially, this means that multiple copies of all files exist and are spread among multiple servers or nodes, making the data more resilient to unexpected disruptions. In the event that some servers are destroyed, data is still available on others. Plus, no manual restoration of data is necessary. The storage system can recover data automatically.
Setting up new physical servers after a disaster is time-consuming. It requires you to have spare servers readily available, which is not always the case (keeping a bunch of extra servers on hand for disaster recovery purposes would not be cost-efficient).
Virtualization offers a better solution. If the servers that host your applications and data are virtualized, you can create copies of the virtual server images and use those images to spin up new instances of the servers following a disaster. Creating a new virtual server based on an existing image is much faster than setting up a physical server from scratch.
In addition, because a virtual server image can run on a wide range of physical host servers, a disaster recovery strategy based on virtual servers allows you to spin up new servers in the cloud, in the event that you don't have the requisite host infrastructure available on-premise.
Point-in-Time-Recovery, or PITR, refers to the ability to choose from multiple points in the past when restoring data and applications after a disaster. PITR provides more flexibility and safeguards than a disaster recovery strategy that is based on a single data backup performed at regular intervals.
With a traditional disaster recovery strategy, you might create data backups every twenty-four hours. The data you restore after a disaster would, therefore, be up to twenty-four hours old, even if the disaster occurred more recently. Any data created more recently would be permanently lost.
With PITR, in contrast, you can choose to restore data from the point in time just before disaster struck. This ensures that any important files created prior to the disaster can be recovered.
Cloud-Based Disaster Recovery****
Traditionally, data backups were stored on dedicated backup servers that either existed alongside an organization's main servers or were housed in a separate, off-site location. The problem with the former approach is that in the event that the entire data center was disrupted, the backup servers would go down, along with the servers they were backing up, because they were all in the same physical location. And the problem with the latter approach (off-site data backups) is that it is expensive. Maintaining a secondary data center solely for backup purposes isn't exactly cheap.
By moving data backup and recovery operations to the cloud, organizations can square this circle. They can back up their data to a remote location that will not go down if a disaster strikes their on-premise data center. At the same time, they can avoid the costs of maintaining a secondary storage site. There is a cost associated with cloud-based data backups, but, in most cases, this is likely to be a more cost-effective solution than running another data center.
These new trends in disaster recovery share a few key features. They enable a greater degree of automation, which in turn reduces costs (because you don't need as many personnel to perform disaster recovery tasks), and speeds results, because automated processes usually go faster than manual ones.
Another is more flexibility. Virtualization and cloud-based data storage provide choice regarding exactly how to backup and restore infrastructure. PITR gives you more flexibility to select which data to back up.
Finally, these new trends enable greater efficiency. They allow organizations to build more robust disaster recovery strategies while spending less on disaster recovery.
Talon provides the data integration, storage, and collaboration tools necessary to build efficiency and flexibility into your disaster recovery strategy. Click here to learn more about the TalonFAST solution.