Results

The time taken by the Failover System for RPS copy, file copy, restore, reconfigure, and Codebook computation varied with varying topology sizes. Time taken for IP Manager provides information on the time taken for the IP Manager with approximately 50,000 managed ports and interfaces.

Table 1. Time taken for IP Manager

Parameters	Active Server	Standby Server
Backup time	388 seconds	438 seconds
Repos Copy time	15 seconds	35 seconds
Restore repos time	126 seconds	104 seconds
Reconfiguration	225 seconds	510 seconds

Time taken for Service Assurance Manager provides information on the time taken for Service Assurance Manager.

Table 2. Time taken for Service Assurance Manager

Parameter	Active Server	Standby Server
Backup time	18 to 30 seconds	18 to 30 seconds

When the topology was increased by approximately 30 devices, the time taken by the Failover System to copy the RPS and other files, restore, and reconfigure the IP Manager also increased. Time taken for IP Manager after increasing topology provides more information.

Table 3. Time taken for IP Manager after increasing topology

Parameters	Active Server	Standby Server
Backup time	438 seconds	Not tested
Repos Copy time	17 seconds	Not tested
Restore repos time	135 seconds	Not tested
Reconfiguration	263 seconds	Not tested

Failover time on the Active and Standby servers provides information on the actions during a failover and, the time taken to complete the failover and bring up the new Active server.

Table 4. Failover time on the Active and Standby servers

Failover action	Time taken
Time taken to detect that the Active server is down and trigger a failover	Approximately 3 minutes
Time taken by the Failover System to assign the Standby role to the Active server which is down	10 seconds
Expected time which includes DetachInterval(300) and failover detection time from polling interval(120), retry(3), timeout(700 ms)	8 minutes
Total number of polling cycles to set up monitoring and polling on the new Active server	3 polling cycles
Total time taken to complete the failover action and bring up the new Active server	Approximately 11 minutes

Note:

The performance of the Failover System was tested in the VMware lab environment. There are external factors outside of software control, such as customer topology size, server machine resource allocation, LAN configuration, and others, that may introduce variances in the statistics.