Deployment for the use case
Before we can start walking through the deployment process, we need to answer a number of questions:
- How many nodes should we have per gateway?
- How many logical gateways do we need?
- How much storage will the database need to account for details such as the volume of analytics needing to be stored?
How many nodes
Stepping back to our use case, MRA identified that they wanted to start small so that they can build experience and confidence. So, we can assume that we are not going to see more than 10 API calls per minute over a 24/7 x 365 period from external sources or internal sources. While workload volumes won't be large, to start the environment needs a reasonable level of robustness so that the external consumers are not impacted by maintenance processes. This points to deploying gateways as two or three nodes depending upon the maintenance strategy.
While the performance of a gateway is somewhat dependent upon the configuration of the hardware it is running on and...
