4.7.1.2 Defining an Appropriate RPO
The Recovery Point Objective set by the business has a direct impact on the bandwidth available for replication traffic. For example, if you have an RPO set to one hour, and have an hourly data change rate of 20 GB across your protected virtual machines, that means that vSphere Replication will need to be able to transmit 20 GB of changed data within each one-hour period.
A connection between the end customer and the VMware Cloud Provider of 100 Mbps will allow for transmission of 100 GB each hour, assuming the link is dedicated for replication traffic. If the network is leveraged for additional services, you must take in to account the amount of saturation of the link for other services.
With this RPO example, we could have 80 percent saturation for other services and still meet the RPO of transmitting 20 GB within the hour period.
The following table highlights how the data change rate and the network link speed have a direct impact on the potential RPO that could be offered to end customers.
Table 9. Data Change Rate Compared with Link Speed and Potential RPO
Hourly Data Change Rate | Link Speed (No saturation) | Potential RPO |
5 GB | 100 Mbps | 15 mins (7 mins)* |
10 GB | 100 Mbps | 15 mins (13 mins) |
20 GB | 100 Mbps | 30 mins (25.5 mins) |
40 GB | 100 Mbps | 1 hour (50.1 mins) |
80 GB | 100 Mbps | 2 hours (101.8 mins) |
Note These calculations round up and do not take in to account any saturation on the links. Take this in to consideration when making your calculations.
*With vSAN as the storage layer, it is possible to configure an RPO of 5 minutes where applicable.
It is important to set the correct RPO because violations will create a snowball effect where the RPO falls further and further behind the schedule and this will have an impact on the ability to perform recovery.