If we want to pause the execution of the action we can have think time at the end of the action. or we can control by setting the pacing option in the run time settings.Wat is the difference between setting the pacing option and having think time at the end of the action.?
thanks for your responses
Think time is time between keyboard and chair. This is something that occurs naturally throughout a business process as users read screens, type at various rates to fill in forms, take breaks for coffee and phone, etc... These inter-request delays are windows of opportunity for other users to make requests of the system and have their requests satisfied. This chaotic arrival and departure of requests is a foundation of the client-server model of computing.
As to controlling the pacing with think time there may be models where you need precise control which is not afforded by the tool, but I would recommend a different option. If your tool allows, consider sleep() instead of a think time, as think time can uisually be altered by your tool's run time settings to allow for variance in the behavior whereas a sleep() type solution will be unaltered by run time information.
Here is an example of a type of test where you want to adjust a time at the end of a set of actions, a decaying interval between iterations to increase load with a fixed number of sessions. First you ramp up to a static interval, run for a period of time representing a fixed number of iterations per user, then after a specific iteration number you begin to adjust a sleep interval at the end of your actions by sleep()/(actual iteration number - fixed iteration offset for decay to begin). With each new iteration after your mark the delay between iterations decreases, first by half, then to a third, then a fourth, until eveneutally you consume the inter-iteration delay, but still leaving in place the think tiem delays for natual pacing within the business process.
Such a fine degree of control is typically not offered in existing test tools, but the tools do offer more pacing options than just offset at the end of the iteration (as would be suggested by just having think time or sleep at the end of a set of actions), such as pacing from the start of one iteration to the next, pacing set off the end of the previous transaction, fixed or random values within a range for both of the above. This is all with the goal of better allowing you to represent the chaotic arrival and departure within a controllable, repeatable range and to hit your transaction targets.