httpd-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Injam, Nishanth" <nin...@tribpub.com>
Subject [users@httpd] Strange results with using Apache to compress dynamic content - Seeking explanations
Date Thu, 26 Mar 2015 20:04:53 GMT
Hi folks,

I am writing to seek explanations about some strange things we have been seeing. In our attempt
to get Apache to compress dynamic content (not static), we turned off compression at the loadbalancer
level and got dynamic compression working with deflate module.

We’ve been testing this change using Jmeter, without and with deflate module config, and
found some interesting results with regards to latency. Latency decreased instead of increasing
slightly when we used deflate to compress dynamic content. We noticed no significant differences
in load averages in all the multiple iterations of testing we did. Often, they were slightly
higher or about the same, and had little to zero adverse impact. We are not worried about
CPU utilization. But since apache is doing the extra work of compressing on the fly with deflate
module now, we thought latency would increase slightly. The average response times went down
because we are serving much smaller files now. But latency(the time before first byte is sent)
should have increased because it has to do all the extra work of compressing which it didn’t
have to do before. But we see that latency went down too. It’s a good side-effect but we
are not sure what is causing it.

I am clueless why something like this could happen. Would really appreciate any answers.

Thanks,
Nishanth


Mime
View raw message