InfiniBox measures the latency associated with write and read requests as follows:
- Total latency
This is the total time the I/O operation takes, from the time of arrival of the first SCSI command (CDB) to the InfiniBox to the time the final response sent to the host.
The total latency may increase due to host or fabric delays, QoS-added latency and internal latency.
- Internal latency
This is the time InfiniBox spends handling the I/O operation, i.e. the time the I/O operation would take in a "perfect" world where there is no latency on the network, hosts are immediately responsive and no QoS is applied.
Internal latency is affected by system resource use, and can increase due to synchronous replication, but is neither affected by host and fabric delays nor by QoS added latency,
What is monitored by InfiniMetrics
As of InfiniMetrics 4.2, we present graphs for both local total latency and local internal latency:
- SAN total latency by I/O type - the total time the I/O operation takes, broken down by I/O operations
- SAN internal latency vs. total latency - the difference between internal and local latency only for volumes that are not synchronously replicated