Can we have an objective measurement of resource usage in 6.1 compared to 5.47? Or can we have an assurance that all v5 sites will find V6 acceptable?
Without this information or assurance, are we meant to take a risk with out sites on shared servers being pulled for excess resource usage?
The way I would test it is to write a driver script in say Perl, which does a typical mix of http requests per minute (in v5 about 55% cgi and 45% html hits), run this script against another server running UBB on NT. Windows NT has a very nice builtin performance monitor which will measure CPU and RAM loading.
Either compare the CPU/RAM usage against a fixed number of http hits per minute or alternatively run it up to saturation and measure the throughput. Then can you tell us what the difference is in resource usage between V5 and V6. Thanks.
