Home > Storage > PowerFlex > White Papers > CockroachDB Deployment on Dell PowerFlex with Kubernetes > Test results
The first use case is for the tpcc workload, where linear scaling of CRDB pods are carried out, we observe the IOPS and latency with every increment of CRDB pods.
Initially, the CockroachDB cluster is configured with 8 pods, and then the pods are scaled to 12.16, and so on, till 40. As seen from the graph, the IOPS gradually increases as the pods scale up. It is also observed that, the operations/second for CockroachDB also increase linearly as the pods scale up.
The following graph shows the linear scaling of the tpcc workload:
The above graph clearly indicates that CockroachDB is highly scalable with more IOPS and ops/s seen when the number of pods is increased. The latency for all the runs, at the storage side is well below 0.7 milliseconds.
Note: During the tests, the CPU utilization was about 79 percent for all the pods. No PowerFlex storage bottlenecks were seen anywhere during the testing.