[clug] how to use traceroute (now: iinet performance)

[Stephen Hocking]

Given that I've not come anywhere near exhausting my quota, if I'm being
shaped, wrath will be expressed.


On Thu, Jun 26, 2014 at 6:58 PM, Simon Oxwell <soxwell at gmail.com> wrote:

> Apologies for the top reply and possibly sounding daft,  but fast after
> midnight and slow in the morning prompts the question: are you being shaped
> - peak vs off peak quota and all ?
>
> Simon
>  On 25 Jun 2014 13:32, "Eyal Lebedinsky" <eyal at eyal.emu.id.au> wrote:
>
> >
> >
> > On 06/25/14 13:21, Scott Ferguson wrote:
> >
> >> On 25/06/14 13:05, Eyal Lebedinsky wrote:
> >>
> >>> Scott, how did you copy the results from the ndt flash panel? my
> firefox
> >>> won't let me.
> >>>
> >>
> >> (from memory) Clicked on the Details button, then clicked on the start
> >> of the text, dragged the mouse to select all, Ctrl+C.
> >>
> >
> > Not here. Not with Ctrl+C and not with RightClick/copy.
> > This is firefox 30.0 on f19.
> > Maybe you installed some extension?
> >
> >  I also note you appear to have your DNS less than optimally configured -
> >> though you may have (local) DNS caching configured. DNS (using
> >>
> >
> > Yes, have a local DNS. However, for one (long) download, how much does
> the
> > DNS degrade the result?
> > No other activity on the line at the same time.
> >
> >  workbench) is normally the first place I start tuning broadband
> >> connections (ISP supplied DNS is generally not optimal).
> >>
> >> Then I check MTU/MRU and MSS on my router (Debian box).
> >>
> >> After that I tune my network buffers based on the results from NDT:-
> >> scott at work:~$ cat /proc/sys/net/core/rmem_default
> >> 163840
> >> scott at work:~$ cat /proc/sys/net/core/rmem_max
> >> 131071
> >> scott at work:~$ cat /proc/sys/net/core/wmem_default
> >> 163840
> >> scott at work:~$ cat /proc/sys/net/core/wmem_max
> >> 131071
> >> scott at work:~$ cat /proc/sys/net/ipv4/tcp_mem #NOTE: this is only for
> IPV4
> >> 21180   28243   42360
> >> scott at work:~$ cat /proc/sys/net/core/optmem_max
> >> 10240
> >>
> >> I 'suspect' your slow connection is due to the common poor performance
> >> we get in Canberra when Telstra's cables get wet (many have poor joints
> >> and lots of cables are so old they've lost their oil - you'll see it as
> >> "loss to earth on one leg" which degrades sync).
> >>
> >
> > Probably, but how come the whole problem went away after midnight, and is
> > back this morning?
> > I suspect lack of bandwidth capacity (network overloaded big time).
> >
> > Eyal
> >
> >  Eyal
> >>>
> >>>
> >>> On 06/25/14 12:41, Scott Ferguson wrote:
> >>>
> >>> [trim]
> >>>
> >>>  NDT test run towards M-Lab server
> >>>> ndt.iupui.mlab2.syd02.measurement-lab.org
> >>>>
> >>>> RTT between client and M-Lab server
> >>>> 49 ms
> >>>>
> >>>> DOWNLOAD SPEED
> >>>> 3.7 Mbps
> >>>>
> >>>> UPLOAD SPEED
> >>>> 1.7 Mbps
> >>>>
> >>>> As you can see my network settings could be optimised further:-
> >>>>
> >>>
> >>> [trim]
> >>>
> >>>  Kind regards
> >>>>
> >>>
> >> Kind regards
> >>
> >
> >
> > --
> > Eyal Lebedinsky (eyal at eyal.emu.id.au)
> > --
> > linux mailing list
> > linux at lists.samba.org
> > https://lists.samba.org/mailman/listinfo/linux
> >
> --
> linux mailing list
> linux at lists.samba.org
> https://lists.samba.org/mailman/listinfo/linux
>