On Sun 17 Aug 2014 at 01:10:22PM +0800, Peter Bell wrote: > On 17/08/14 02:30, guns wrote: > > >On Fri 15 Aug 2014 at 06:17:37PM +0800, Peter Bell wrote: > > > >>I have been experiencing extremely slow browsing performance - as > >>an example, a 'first' google search takes, typically, 20 seconds to > >>display the results page - the slowest I've recorded has been 30 > >>seconds. Subsequent google searches complete within 2 seconds. > > > >Firefox (and Chromium IIRC) will break down individual HTTP request > >timings into sections, the first of which is DNS resolution. You can > >use this to at least confirm that DNS is really the issue while web > >browsing. > > > > https://developer.mozilla.org/en-US/docs/Tools/Network_Monitor > > https://mdn.mozillademos.org/files/7417/nm-timings.png > > Wow, that is a useful tool, of which I was not aware. > > The results I am seeing show that my problem has little to do with > name resolution, but to some considerable delays in connecting and, > sometimes, sending a request. I'm guessing that these will be delays > at my ISP, or beyond.
traceroute will usually identify the slow link.
I have also read that CDNs offer different IP addresses based on the
geographical region of your nameserver:
http://blogs.gartner.com/lydia_leong/2008/10/15/the-nameserver-as-cdn-vantage-point/
Might not be your problem, but it's worth thinking about if the
webservers you frequent are geographically remote from your upstream
nameservers.
> However, since this appears to exonorate dnsmasq as the major culprit,
> I will need to find another forum for this discussion ... any
> recommendations?
If you find an obvious bottleneck with traceroute, a solution or
workaround will probably become obvious.
Good luck!
guns
pgpNUissy3tjR.pgp
Description: PGP signature
_______________________________________________ Dnsmasq-discuss mailing list [email protected] http://lists.thekelleys.org.uk/mailman/listinfo/dnsmasq-discuss
