[Rpm] actively measuring owd successfully using icmp

[Rpm] actively measuring owd successfully using icmp

[Dave Taht]

Somewhere in this thread the actual working method is buried, but:

https://forum.openwrt.org/t/cake-w-adaptive-bandwidth/108848/944

-- 
I tried to build a better future, a few times:
https://wayforward.archive.org/?site=https%3A%2F%2Fwww.icei.org

Dave Täht CEO, TekLibre, LLC

Re: [Rpm] actively measuring owd successfully using icmp

[Sebastian Moeller]

Two notes.
1) Technically it is not OWD that is measured, but "OWD + unknown_offset", but since the goal is to evaluate the delta between the current delay and a history delay aggregate, the unknown offset does not matter too much, and as it turns out quite a lot of the tested reflectors are reasonably well synchronized already (to my utter, utter surprise).
2) It appears that ICMPv6 removed the timestamp option, does anybody here have a link to the discussions that lead to this somewhat unfortunate decision?

Best Regards
	Sebastian


> On Dec 11, 2021, at 13:59, Dave Taht via Rpm <rpm@lists.bufferbloat.net> wrote:
> 
> Somewhere in this thread the actual working method is buried, but:
> 
> https://forum.openwrt.org/t/cake-w-adaptive-bandwidth/108848/944
> 
> -- 
> I tried to build a better future, a few times:
> https://wayforward.archive.org/?site=https%3A%2F%2Fwww.icei.org
> 
> Dave Täht CEO, TekLibre, LLC
> _______________________________________________
> Rpm mailing list
> Rpm@lists.bufferbloat.net
> https://lists.bufferbloat.net/listinfo/rpm

Re: [Rpm] actively measuring owd successfully using icmp

[Matt Mathis]

[-- Attachment #1: Type: text/plain, Size: 2150 bytes --]

icmp_timestamp has not been removed, and in Linux it is still present in
the shared icmp code used by both IPv4 and IPv6.

More likely a bug that has never been noticed, because nobody bothered to
test it.  I note that most icmp documentation doesn't even mention it's
existence.

If a kernel bug, most likely a corrupted checksum.  But my bet would be
code outright missing from the library.

Good luck!

Thanks,
--MM--
The best way to predict the future is to create it.  - Alan Kay

We must not tolerate intolerance;
       however our response must be carefully measured:
            too strong would be hypocritical and risks spiraling out of
control;
            too weak risks being mistaken for tacit approval.


On Sat, Dec 11, 2021 at 5:20 AM Sebastian Moeller via Rpm <
rpm@lists.bufferbloat.net> wrote:

> Two notes.
> 1) Technically it is not OWD that is measured, but "OWD + unknown_offset",
> but since the goal is to evaluate the delta between the current delay and a
> history delay aggregate, the unknown offset does not matter too much, and
> as it turns out quite a lot of the tested reflectors are reasonably well
> synchronized already (to my utter, utter surprise).
> 2) It appears that ICMPv6 removed the timestamp option, does anybody here
> have a link to the discussions that lead to this somewhat unfortunate
> decision?
>
> Best Regards
>         Sebastian
>
>
> > On Dec 11, 2021, at 13:59, Dave Taht via Rpm <rpm@lists.bufferbloat.net>
> wrote:
> >
> > Somewhere in this thread the actual working method is buried, but:
> >
> > https://forum.openwrt.org/t/cake-w-adaptive-bandwidth/108848/944
> >
> > --
> > I tried to build a better future, a few times:
> > https://wayforward.archive.org/?site=https%3A%2F%2Fwww.icei.org
> >
> > Dave Täht CEO, TekLibre, LLC
> > _______________________________________________
> > Rpm mailing list
> > Rpm@lists.bufferbloat.net
> > https://lists.bufferbloat.net/listinfo/rpm
>
> _______________________________________________
> Rpm mailing list
> Rpm@lists.bufferbloat.net
> https://lists.bufferbloat.net/listinfo/rpm
>

[-- Attachment #2: Type: text/html, Size: 3447 bytes --]

Re: [Rpm] actively measuring owd successfully using icmp

[Sebastian Moeller]

Hi Matt,

thanks a lot.

> On Dec 11, 2021, at 17:25, Matt Mathis <mattmathis@google.com> wrote:
> 
> icmp_timestamp has not been removed, and in Linux it is still present in the shared icmp code used by both IPv4 and IPv6.

	Ah thanks, serves me well for not actually testing; I did however look into the ICMPv6 RFC and obviously did not notice that this describes changes from ICMPv4 and that hence the absence of type 13/14 timestamps in RFC does not mean they are not supported/intended to work.

> 
> More likely a bug that has never been noticed, because nobody bothered to test it.  I note that most icmp documentation doesn't even mention it's existence.

	Yes, that is what made me come to the wrong conclusion, thanks for clearing that up.

> 
> If a kernel bug, most likely a corrupted checksum.  But my bet would be code outright missing from the library.

	Since I asked before testing, maybe I should test this now, maybe it just works....

Thanks again & Best Regards
	Sebastian


> 
> Good luck!
> 
> Thanks,
> --MM--
> The best way to predict the future is to create it.  - Alan Kay
> 
> We must not tolerate intolerance;
>        however our response must be carefully measured: 
>             too strong would be hypocritical and risks spiraling out of control;
>             too weak risks being mistaken for tacit approval.
> 
> 
> On Sat, Dec 11, 2021 at 5:20 AM Sebastian Moeller via Rpm <rpm@lists.bufferbloat.net> wrote:
> Two notes.
> 1) Technically it is not OWD that is measured, but "OWD + unknown_offset", but since the goal is to evaluate the delta between the current delay and a history delay aggregate, the unknown offset does not matter too much, and as it turns out quite a lot of the tested reflectors are reasonably well synchronized already (to my utter, utter surprise).
> 2) It appears that ICMPv6 removed the timestamp option, does anybody here have a link to the discussions that lead to this somewhat unfortunate decision?
> 
> Best Regards
>         Sebastian
> 
> 
> > On Dec 11, 2021, at 13:59, Dave Taht via Rpm <rpm@lists.bufferbloat.net> wrote:
> > 
> > Somewhere in this thread the actual working method is buried, but:
> > 
> > https://forum.openwrt.org/t/cake-w-adaptive-bandwidth/108848/944
> > 
> > -- 
> > I tried to build a better future, a few times:
> > https://wayforward.archive.org/?site=https%3A%2F%2Fwww.icei.org
> > 
> > Dave Täht CEO, TekLibre, LLC
> > _______________________________________________
> > Rpm mailing list
> > Rpm@lists.bufferbloat.net
> > https://lists.bufferbloat.net/listinfo/rpm
> 
> _______________________________________________
> Rpm mailing list
> Rpm@lists.bufferbloat.net
> https://lists.bufferbloat.net/listinfo/rpm