Modern enterprises know that disasters are a fact of life. They also understand that disasters come in many shapes and sizes and lead to varying degrees of availability loss and impact on business continuity. Because major wide-scale disasters occur way less often than smaller disasters from within a data center, it’s also important to plan and test cloud disaster recovery models for smaller disasters that can happen at VM, guest OS or application levels.
Because disasters can lead to revenue, productivity and reputation loss, it’s important to understand that having cloud-based backup is just one piece of the business continuity and availability puzzle. Here at Veeam, we empower our cloud and service providers to offer services based on Veeam Cloud Connect backup and replication. However, the planning and testing of what happens once disaster strikes is ultimately up to either the organizations purchasing the services or the services company offering DRaaS (Disaster Recovery as a Service) and business continuity wrapped around backup and replication offerings.
Why it’s important to plan
In theory, planning for a disaster should be completed before selecting a product or solution. In reality, it’s common for organizations to purchase cloud DR (disaster recovery) services without an understanding of what needs to be put in place prior to workloads being backed up or replicated to a cloud provider. Concepts like recovery time and recovery point objectives (RTPO) need to be understood and planned so, if a disaster strikes and failover has occurred, applications will not only be recovered within SLAs, but also that data on those recovered workloads will be useful in terms of its age.
Smaller RTPO values go hand-in-hand with increased complexity and administrative services overhead. When planning ahead, it’s important to size your cloud disaster platform and build the right disaster recovery model that’s tailored to your needs. When designing your DR plan, you will want to target strategies that relate to your core line of business applications and data.
A staged approach to recovery means that you recover tier-one applications first so the business can still function. A common tier-one application example is the mail server. Another is payroll systems, which could result in an organization being unable to pay its suppliers. Once your key applications and services are recovered, you can move on to recovering data, while keeping mind that archival data generally doesn’t need to be brought online first. Again, being able to categorize systems where your data sits and then working those categories into your recovery plan is important.
Planning should also include specific tasks and controls that need to be followed up on and adhered to during a disaster. It’s important to have specific run books executed by specific people for a smoother failover. Finally, it is critical to make sure that all IT staff know how to accessing applications and services after failover.
Why it’s important to test
When talking about a cloud disaster recovery model, there are a number of factors to consider before a final sign-off and the validation testing process. Once your plan is in place, test it regularly and make adjustments if issues arise from your tests. Partial failover testing should be treated with the same level of criticality as full failover testing.
Testing your DR plan ensures that business continuity can be achieved in a partial or full disaster. Beyond core backup and replication services testing, you should also test networking, server and application performances. Testing should even include situational testing with staff to be sure that they are able to efficiently access key business applications.
Cloud disaster recovery models
There are a number of different cloud disaster recovery models, which can be broken down into three main categories:
- Private cloud
- Hybrid cloud
- Public cloud
Veeam Cloud Connect technology works for hybrid and public cloud models, while Veeam Backup & Replication works across all three models. The Veeam Cloud & Service Provider (VCSP) program offers Veeam Cloud Connect backup and replication classified as hybrid clouds offering RaaS (recovery-as-a-service). Public clouds, such as Azure, can be used with Veeam Backup & Replication to restore VM workloads. Private clouds are generally internal to organizations and leverage Veeam Backup & Replication to replicate or back up or for a backup copy of VMs between data center locations.
The ultimate goal here is to choose a cloud recovery model that best suits your organization. Each of the models above offer technological diversity and different price points. They also plan and test differently in order to, ultimately, execute a disaster plan. Each model comes with its own issues that need to be considered, including:
- The replicated or copied data type
- The SLAs offered by the company offering services for each model
When a partial or full disaster strikes, a thoroughly planned and well-tested DR plan, backed by the right disaster recovery model, will help you avoid a negative impact on your organization’s bottom line. Veeam and its cloud partners, service-provider partners and public cloud partners can help you build a solution that’s right for you.
The post Cloud disaster recovery models: planning and testing appeared first on Veeam Software Official Blog.