Discovery storage rates
You can use the discovery storage rates that are listed to help you determine the number of components that are required to meet your objectives for discovery.
In a synchronization server deployment, storage to the database
is typically the main bottleneck. In a streaming server deployment,
any bottleneck in discovery throughput has moved to the sensors that
are waiting for the data to be stored.
The following table lists typical discovery storage rates.
The number of imports that is listed is the number of discovery servers that are able to
send enough data to the storage servers so that they are not waiting on sensor results, which means
that they are storing at their maximum rate. As the number of storage servers increases, there is an
increase in the CI storage rate.
| Number of storage servers | Number of imports (discovery servers) | CI rate per second | Percent improvement | Number of threads used for persisting discovery results in the database (topopumpcount) |
|---|---|---|---|---|
| 1 | 2 | 144 | 16 | |
| 2 | 2 | 246 | 101.83 | 16 |
| 3 | 2 | 280 | 33.91 | 16 |