Since the Lab Server runs the BLL, are the resources required on client PCs reduced?
All of the testing was done using the API and an older Solaris SunFire 100.
For any single command, the LS would finish faster. That means that if the single command takes a long time to complete (perhaps an apply),
then the LS would be faster. However, the LS also has more overhead in communication, so if you are issuing a lot of commands that don’t
require a lot of processing, then the LS is going to be slower. If you have a slower network link…it’s going to be a LOT slower.
Solaris Solaris LabServer Windows Windows LabServer
Action Time Action Time Action Time Action Time
Config 28 Config 33 Config 9 Config 16
Apply 9 Apply 6 Apply 8 Apply 6
Traffic Start 47 Traffic Start 19 Traffic Start 23 Traffic Start 19
Total 84 Total 58 Total 40 Total 41
Here are some of the numbers that we have received. These are times (in seconds) to accomplish the same tasks using Solaris local, Solaris with LS,
Windows local and Windows with LS.
The “Apply” stage is actually connect, reserve, map and apply.
The Traffic start, for some reason, takes forever.