After you've defined recovery point objectives and recovery time objectives and tested your disaster recovery methods, maintaining cloud-based DR requires a considerable amount of management -- even if a disaster never strikes.
It is the DR manager's responsibility to ensure that provisioning is properly configured.
The key to successful disaster recovery management involves bringing systems online and restoring them according to service-level agreements (SLAs), meeting RPO and RTO goals and mitigating results of your business impact analysis (BIA).
But before disaster actually hits, it's important to know where -- physically and virtually -- underlying systems such as servers are located, which end users and IT staff can and need to access them to restore service, and what other components, such as network switching, routing, firewalls, etc., are central to the disaster recovery plan. For example, you should have a directory system in place such as Active Directory or OpenLDAP to manage user authentication, database systems to store data and end-user access points to obtain data.
The easiest way to manage these aspects of DR is to have a three-pronged provisioning protocol in place. First, you need to bring the underlying directory system online. Next, bring up database systems and then bring back application servers. Finally, you are ready to restore data.
In a manual disaster recovery process, the process of restoring data can take hours per system. With virtual machine-based backup to the cloud, it's as simple as taking a VM that's located in the cloud, making it available and then powering it on. Point-in-time, snapshot-based VMs that power up and take over for downed servers have taken the place of tape backups.
It is the DR manager's responsibility to ensure that, in addition to testing the validity of the VM-based recovery systems, provisioning is properly configured. Automated provisioning, or having VMs start automatically, allows you to meet SLA-based RPO/RTO metrics. You then can use the central log interface built into most cloud and virtualization platforms to document the testing process and results to verify you've met SLAs.
While this sounds very complex -- and it can be in large environments -- cloud computing offers small and medium businesses a level of efficiency that traditional backups cannot match. The typical SMB environment often is a flat network with minimal complexity (virtual LANs, etc.). Using a third-party cloud provider that specializes in recovery greatly reduces the time and effort involved in setting up and testing a product.
DR in the cloud is flexible, so IT teams can monitor, test and adjust it without affecting production systems. This makes disaster recovery in SMBs much less labor intensive, a welcome benefit in businesses that can scarcely afford a fully staffed IT department, not to mention a dedicated DR group.
Cost shouldn't prevent good disaster recovery
The high cost involved in implementing a solid disaster recovery plan often forces companies to stick with traditional backup and restore methods. While having high RPO and RTO numbers, which indicate a need for quick recovery, is a perfectly valid justification for keeping the old systems in place, cost shouldn't be an issue.
As the cost-per-GB of cloud storage continues to decrease, the cost of media for tape backups remains roughly constant. Backup devices, such as disk-to-disk-to-tape systems, robotic libraries and others backup products, are expensive to replace. When the time comes to replace them, SMBs often need to examine the cost of storing data in the cloud for backups. That may also be the time to consider implementing full DR in the cloud. This shift in the cost structure, the need to perform a BIA and the relative ease of moving to cloud-based DR make the change an attractive, even compelling, option.
ABOUT THE AUTHOR
Joseph Foran is the IT director for Bridgeport, Conn.-based FSW, Inc., and principal at Foran Media, LLC. He has been in IT since 1995, specializing in infrastructure, and involved in virtualization and cloud computing since 2002. Email Joe at firstname.lastname@example.org or follow him on Twitter (@joseph_foran).