How the objectives were met
Vendors of server based Internet content filters were invited to participate in the testing, which was conducted at the Telstra Broadband eLab in Launceston, Tasmania.
Vendors were requested to submit their own hardware system(s) with their application pre-installed. The reason for this request was to ensure that no vendor was disadvantaged by having to comply and run on hardware that was specified by either NetAlert or the TestLab Each vendor could therefore submit the platform of their choice so their application would run at its best performance.
To test the network performance of the content filters, and any performance degradation, automatic test tools were utilised to provide a range of virtual loads that were able to simulate a large number of concurrent users.
Once preliminary baseline performance testing had been carried out, each vendor was invited to then send one of their engineers to the test site and have the opportunity to tune the performance of the filtering system. At the same time the vendor’s engineers were given the opportunity to tune their products for the accuracy testing.
When tuning of the filters was complete the automatic test system was run without a filter connected to determine its baseline performance. Once this was achieved the system was run again, once with the filter connected and disabled and then a second time with the filter enabled. This process was repeated for each filter and the results correlated.
Automatic tools were again used to run basic accuracy tests. This testing automatically ran through two lists of URLs and collected results on whether the site was blocked or not.
Finally, a further test system was installed in the RMIT TestLab at Bundoora to conduct user based testing and collect subjective feedback from a number of typical Internet users regarding their impression of the system performance with filters connected and not connected.
Back to Objectives of the filter trial | Table of contents | Forward to Findings of the report

