Re: [reSIProcate] Followup on c-ares support
Brad Spencer wrote:
On Wed, Nov 26, 2008 at 05:33:08PM -0600, Adam Roach wrote:
During our testing, we did note that c-ares is substantially
faster than the resip version of ares.
Interesting. How much faster are we talking about?
I don't have the actual benchmarks at hand, but Byron and I
discussed first-cut performance results. Doing some very DNS-heavy
processing (i.e., each request was associated with a different DNS
record, so no caching was possible), I think he said that we were
able to push almost twice as many SIP transactions through repro
with c-ares than with the built-in ares.
No SIP processing was done. The load-test was relying on globbing
ENUM NAPTR records; so it was an initial NAPTR query that would miss
cache, which fed into another NAPTR query that did cache. c-ares was
about twice as fast at this as resip-ares.
Best regards,
Byron Campen
Normal disclaimers apply (we may have overlooked something, YMMV,
etc).
/a
_______________________________________________
resiprocate-devel mailing list
resiprocate-devel@xxxxxxxxxxxxxxx
https://list.resiprocate.org/mailman/listinfo/resiprocate-devel