How does it work?
The basic principle behind Load Impact is quite simple:
The user (you) logon to our web frontend system, configure a test and press the start button. This
places your test into a system-wide test queue in our master database.
The test engine program regularly polls the database to see if there are any tests queued
there for execution. It finds your test, removes it from the queue and launches a test engine subprocess to handle the
test execution.
The test engine subprocess now contacts the load generator node where the test is to be
executed. It transfers the test configuration to the load generator node and commands it to start the test.
The load generator node executes the load generator program and provides it with the
list of URLs (the load script) that are to be loaded during the test by all the simulated clients.
The load generator program performs the actual test, connecting to the target website with many simultaneous
connections, downloading all the URLs specified, and gathering statistics on load times (delays) and response codes. The statistics are
updated continuously throughout the test, and sent in real time back to the test engine subprocess.
The test engine subprocess use the statistics from the load generator program to perform
live updates of the master database, continuously inserting test results there while the test is in progress.
Your web browser uses Ajax requests to poll the latest test results and statistics from the database while the
test is running. This dynamic information is what enables us to make a running test more interesting to watch. A running
test will display its current status, its progress (how far it has come), and statistics on the amount of data downloaded
from the website, the number of HTTP requests, etc. The test result graph will also be updated dynamically while the test
is running.
Once the test is finished, all live updates cease and only the test results remain on the page. These results stay
in our database until you delete them, and can be viewed by you at any time. You can also export them as .csv files (easy to
import into Excel or OpenOffice).
If you want to know the basics - What is Load Impact and why should you use it? - look at
What is Load Impact?
You can also go to our online forum for more information, or directly to the
FAQ section there
The user (you) logon to our web frontend system, configure a test and press the start button. This places your test into a system-wide test queue in our master database.
The test engine program regularly polls the database to see if there are any tests queued there for execution. It finds your test, removes it from the queue and launches a test engine subprocess to handle the test execution.
The test engine subprocess now contacts the load generator node where the test is to be executed. It transfers the test configuration to the load generator node and commands it to start the test.
The load generator node executes the load generator program and provides it with the list of URLs (the load script) that are to be loaded during the test by all the simulated clients.
The load generator program performs the actual test, connecting to the target website with many simultaneous connections, downloading all the URLs specified, and gathering statistics on load times (delays) and response codes. The statistics are updated continuously throughout the test, and sent in real time back to the test engine subprocess.
The test engine subprocess use the statistics from the load generator program to perform live updates of the master database, continuously inserting test results there while the test is in progress.
Your web browser uses Ajax requests to poll the latest test results and statistics from the database while the test is running. This dynamic information is what enables us to make a running test more interesting to watch. A running test will display its current status, its progress (how far it has come), and statistics on the amount of data downloaded from the website, the number of HTTP requests, etc. The test result graph will also be updated dynamically while the test is running.
Once the test is finished, all live updates cease and only the test results remain on the page. These results stay in our database until you delete them, and can be viewed by you at any time. You can also export them as .csv files (easy to import into Excel or OpenOffice).