2.3 Product Expectations

2.3.1 RTO

RTO (Recovery Time Objective) is the time it takes to failover a workload in Forge and typically occurs in 10 to 30 minutes. This duration includes:

  • booting up the workload

  • configuring the failover

The boot up time is typically affected by the number of services running on the workload. The more services that need initialization, the longer it takes the workload to completely boot up. Certain failover configurations (e.g. joining the domain) may also require extra reboots.

Forge use also plays a factor in the RTO. The more operations (i.e. replications or other failovers) running on Forge, the longer it takes to failover the workload. To get an average failover time for workloads in your environment, perform test failovers at various times and run the Forge reports.

NOTE:Domain controllers typically take longer to failover as they require more configuration and reboots. The duration may be as much as double the time it takes to perform a typical failover.

2.3.2 RPO

RPO (Recovery Point Objective) is the time between incremental replications. The minimum achieveable RPO depends on the current utilization of Forge, the number and scope of changes on the workload and the network speed. Typically, the overhead time for a replication is between 10 and 20 minutes, depending on how many other operations are currently running on Forge. Therefore, the time it takes to complete a replication is 10 to 20 minutes plus the amount of time it takes to transfer the changes.

2.3.3 Failover

At the time of a disaster, if a single workload has to be failed over, typically there will be enough resources (memory and CPU) on Forge to allow it to perform as well as the production workload. In the case where multiple workloads have to be failed over concurrently, the resources will have to be shared among these workloads and the performance may not be on par with production workloads. This should be taken into consideration when designing the disaster recovery plan. You may run through different scenarios using the test failover functionality to be better prepared for such scenarios.

NOTE:You may suspend other operations, such as replications, to free up additional resources, as required. You can also adjust the amount of memory assigned to each workload in the failover settings.

2.3.4 Recovery Points

By enabling recovery points, Forge keeps track of changes between replications for the applicable workloads. In order to store this information, each recovery point requires extra disk space on Forge. The amount of required disk space is directly tied to the amount of changes on the workload. The more changes, the more disk space that is required. This should be taken into consideration when choosing the number of recovery points. You may consult the Forge reports to estimate the average amount of change for each workload.