eValid LoadTest scenarios can be played back with varying fidelity, as outlined in the table below. The full-browser modes -- in the first three columns -- require use of one or more eValid browsers running on the driver machine.
Agent-based playbacks with the eVlite playback engine process back any eValid script but only performs navigation events. Each eVlite instance can have a replication count R of up to 1000. medium-size desktop machine can handle 100+ eValid's, or about N = 10 eVlite's, or some combination of these that doesn't consume all of the memory.
|Playback Script Feature||eValid Browser Based Playback||eVlite Agent Based Playback|
N = 1
R = 1000
N = 10
R = 1000
|Machine Capacity Issues|
|Typical Machine Capacity||100+ simultaneous full browsers, any mix of scripts||100+ simultaneous full browsers, any mix of scripts||100+ simultaneous full browsers, any mix of scripts||1 eVlite Playback Agent, one script.||Up to 1000 eVlite Playback Agent Replications each running the same script.|
|Maximum Simulated Users||100||100||100||1000||10,000|
|What's Included In Playback||Everything||Everything except images||URL navigation||URL navigation||URL navigation|
|What's Excluded In Playback||Nothing||Images||User Interactions Involving Screen||User Interactions Involving Screen||User Interactions Involving Screen|
|Relative Download Volume||100%||~33%||0.1%-33%||0.1%-33%||0.1%-33%|
|Relative Speedup Ratio||1:1||3:1||1000:1 to 3:1||1000:1 to 3:1||1000:1 to 3:1|
|Sophistication of Script|
|Secure ID's or any automatically generated URLs.||Yes||Yes||No||No||No|
|Text Validation on Playback||Yes||Yes||No||No||No|
|Image Validation on Playback||Yes||No||No||No||No|
For comparison an eValid full browser evaluation of a script that does a single navigation to one URL in Serve URL mode is exactly the same work to the server as a single eVlite playback agent with N = 1 for that same script. In both cases only the named URL base page is downloaded.
The 100+ indication makes assumptions about the driver machine and is based on
our internal experimentation.
If the eValid script being played back is partcularly long and complex
this capacity may not be obtained.