From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail-ie0-x22a.google.com (mail-ie0-x22a.google.com [IPv6:2607:f8b0:4001:c03::22a]) (using TLSv1 with cipher RC4-SHA (128/128 bits)) (Client CN "smtp.gmail.com", Issuer "Google Internet Authority G2" (verified OK)) by huchra.bufferbloat.net (Postfix) with ESMTPS id 0420521F443 for ; Mon, 20 Apr 2015 01:11:45 -0700 (PDT) Received: by iejt8 with SMTP id t8so110604427iej.2 for ; Mon, 20 Apr 2015 01:11:45 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20120113; h=mime-version:sender:in-reply-to:references:date:message-id:subject :from:to:content-type; bh=y+2Z65VfrpSFkjqibF7DiedpToYOtwWlirBCMrUVJ9k=; b=KnuJL21GnhSH4Q8xBxUDFMKNd4PwqfRyz/SuUzadOq+8OTQliWHc7eTRNOSq59HkYQ sw0oyItGYLa8DLKWZaipoV2Buz7DTmrJ5EPAqKHdiYod0rnSsrErHMGgW2EPnVt+RrPc VE87hb1W2QndUBGBBl3R+1IzLg9HByuLlzdhgsRWYXwOcM6pioy7+aPa4ry2P+r0WdF0 NhSgDVzVtLcpIQ2yD5hJl5hZqrgJQvdZkj/9Ns/puxIuBMF6UjxYzLQEdov0dHzcTs6/ CdTNa3v6VOpoSoQIO5ppA6vqlp3BCEaVeMXoCT/70/9SDknc4W7GzKWwyILcLaJURBkg c/3g== MIME-Version: 1.0 X-Received: by 10.42.129.73 with SMTP id p9mr17084707ics.48.1429517504876; Mon, 20 Apr 2015 01:11:44 -0700 (PDT) Sender: justinbeech@gmail.com Received: by 10.50.107.42 with HTTP; Mon, 20 Apr 2015 01:11:44 -0700 (PDT) In-Reply-To: References: <87wq18jmak.fsf@toke.dk> <87oamkjfhf.fsf@toke.dk> <87k2x8jcnw.fsf@toke.dk> <87fv7wj9lh.fsf@toke.dk> <87zj64hsy9.fsf@toke.dk> Date: Mon, 20 Apr 2015 18:11:44 +1000 X-Google-Sender-Auth: IJyQ1NQO1m87JoURJpcMUaNcilM Message-ID: From: jb To: Pedro Tumusok , bloat Content-Type: multipart/alternative; boundary=20cf301af6d9142fab0514237b49 Subject: Re: [Bloat] DSLReports Speed Test has latency measurement built-in X-BeenThere: bloat@lists.bufferbloat.net X-Mailman-Version: 2.1.13 Precedence: list List-Id: General list for discussing Bufferbloat List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 20 Apr 2015 08:12:14 -0000 --20cf301af6d9142fab0514237b49 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: quoted-printable Whoops I better set that z-index correctly thanks. It is interesting you mentioned gaming because 10 of the servers are from a place that rents clan servers. They have to be in top quality data centres and not congest anything because their customers abandon them immediately and they're always watching ping time and packet loss. Always happy for long term commitments to a servers. I just need permanent root on an ubuntu 14.10 virtual or real box, or centos I guess. ipv6 dual-stack would be good. Memory cpu and disk unimportant. A 1 gig/e port, but average usage can be set to stay below whatever they prefer. I'm going to do some kind of donor recognition thing, so if a donated server is used it'll show something like a company name and URL I just haven't had to do that yet. thanks -Justin On Mon, Apr 20, 2015 at 5:28 PM, Pedro Tumusok wrote: > I noticed on my tests that the label Ping time during test, was displayed > on top of the tool tips, which meant I only had the y-axis to look at and > had to guessestimate my ping time. > > Another step to help cure the bufferbloat, maybe drawing a few vertical > threshold lines through the ping times. > Visualizing that ping over x ms will make VoIP work badly, gaming will > suffer etc. > At least VoIP have some hard numbers we can use, gaming is more dependant > upon the network code of the game and its client-side prediction I guess. > But still anything over y ms in a fps game means you're dead before you > even see your opponent. > > Are you looking for places to deploy servers? I got a couple of people > here in Norway and Sweden, I can reach out to and ask about that. If yes, > what requirements do you have? > > Pedro > > On Mon, Apr 20, 2015 at 9:00 AM, jb wrote: > >> IPv6 is now available as an option, you just select it in the preference= s >> pane. >> >> Unfortunately only one of the test servers (in Michigan) is native dual >> stack so the test is then fixed to that location. In addition the latenc= y >> pinging during test is stays as ipv4 traffic, until I setup a web socket >> server on the ipv6 server. >> >> All the amazon google and other cloud servers do not support ipv6. They >> do support it as an edge network feature, like as a load balancing front >> end, however the test needs custom server software and custom code, and >> using a cloud proxy that must then talk to an ipv4 test server inside th= e >> cloud is rather useless. It should be native all the way. So until I get >> more native ipv6 servers, one location it is. >> >> Nevertheless as a proof of concept it works. Using the hurricane electri= c >> ipv6 tunnel from my australian non ipv6 ISP, I get about 80% of the spee= d >> that the local sydney ipv4 test server would give. >> >> >> On Mon, Apr 20, 2015 at 1:15 PM, Aaron Wood wrote: >> >>> Toke, >>> >>> I actually tend to see a bit higher latency with ICMP at the higher >>> percentiles. >>> >>> >>> http://burntchrome.blogspot.com/2014/05/fixing-bufferbloat-on-comcasts-= blast.html >>> >>> http://burntchrome.blogspot.com/2014/05/measured-bufferbloat-on-orangef= r-dsl.html >>> >>> Although the biggest "boost" I've seen ICMP given was on Free.fr's >>> network: >>> >>> http://burntchrome.blogspot.com/2014/01/bufferbloat-or-lack-thereof-on-= freefr.html >>> >>> -Aaron >>> >>> On Sun, Apr 19, 2015 at 11:30 AM, Toke H=C3=B8iland-J=C3=B8rgensen >>> wrote: >>> >>>> Jonathan Morton writes: >>>> >>>> >> Why not? They can be a quite useful measure of how competing traffi= c >>>> >> performs when bulk flows congest the link. Which for many >>>> >> applications is more important then the latency experienced by the >>>> >> bulk flow itself. >>>> > >>>> > One clear objection is that ICMP is often prioritised when UDP is no= t. >>>> > So measuring with UDP gives a better indication in those cases. >>>> > Measuring with a separate TCP flow, such as HTTPing, is better still >>>> > by some measures, but most truly latency-sensitive traffic does use >>>> > UDP. >>>> >>>> Sure, well I tend to do both. Can't recall ever actually seeing any >>>> performance difference between the UDP and ICMP latency measurements, >>>> though... >>>> >>>> -Toke >>>> _______________________________________________ >>>> Bloat mailing list >>>> Bloat@lists.bufferbloat.net >>>> https://lists.bufferbloat.net/listinfo/bloat >>>> >>> >>> >>> _______________________________________________ >>> Bloat mailing list >>> Bloat@lists.bufferbloat.net >>> https://lists.bufferbloat.net/listinfo/bloat >>> >>> >> >> _______________________________________________ >> Bloat mailing list >> Bloat@lists.bufferbloat.net >> https://lists.bufferbloat.net/listinfo/bloat >> >> > > > -- > Best regards / Mvh > Jan Pedro Tumusok > > --20cf301af6d9142fab0514237b49 Content-Type: text/html; charset=UTF-8 Content-Transfer-Encoding: quoted-printable
Whoops I better set that z-index correctly thanks.

It is interesting you mentioned gaming because 10 of= the servers are from a place that rents clan servers. They have to be in t= op quality data centres and not congest anything because their customers ab= andon them immediately and they're always watching ping time and packet= loss.

Always happy for long term commitments to a serv= ers. I just need permanent root on an ubuntu 14.10 virtual or real box, or = centos I guess. ipv6 dual-stack would be good. Memory cpu and disk unimport= ant. A 1 gig/e port, but average usage can be set to stay below whatever th= ey prefer.

I'm going to do some kind of donor recogn= ition thing, so if a donated server is used it'll show something like a= company name and URL I just haven't had to do that yet.

=
thanks
-Justin


On Mon, Apr 20, 2015 at 5:28 P= M, Pedro Tumusok <pedro.tumusok@gmail.com> wrote:
<= blockquote class=3D"gmail_quote" style=3D"margin:0 0 0 .8ex;border-left:1px= #ccc solid;padding-left:1ex">
I noticed on my tests that t= he label Ping time during test, was displayed on top of the tool tips, whic= h meant I only had the y-axis to look at and had to guessestimate my ping t= ime.

Another step to help cure the bufferbloat, maybe dr= awing a few vertical threshold lines through the ping times.
Visu= alizing that ping over x ms will make VoIP work badly, gaming will suffer e= tc.=C2=A0
At least VoIP have some hard numbers we can use, gaming= is more dependant upon the network code of the game and its client-side pr= ediction I guess.
But still anything over y ms in a fps game mean= s you're dead before you even see your opponent.

Are you looking for places to deploy servers? I got a couple of people h= ere in Norway and Sweden, I can reach out to and ask about that. If yes, wh= at requirements do you have?

Pedro =C2=A0=C2=A0

On Mon, Apr 20, 2015 at 9:00 AM, jb <= justin@dslr.net>= ; wrote:
IPv= 6 is now available as an option, you just select it in the preferences pane= .

Unfortunately only one of the test servers (= in Michigan) is native dual stack so the test is then fixed to that locatio= n. In addition the latency pinging during test is stays as ipv4 traffic, un= til I setup a web socket server on the ipv6 server.

All the amazon google and other cloud servers do not support ipv6. They d= o support it as an edge network feature, like as a load balancing front end= , however the test needs custom server software and custom code, and using = a cloud proxy that must then talk to an ipv4 test server inside the cloud i= s rather useless. It should be native all the way. So until I get more nati= ve ipv6 servers, one location it is.

Nevertheless = as a proof of concept it works. Using the hurricane electric ipv6 tunnel fr= om my australian non ipv6 ISP, I get about 80% of the speed that the local = sydney ipv4 test server would give.


On Mon, Apr 20, 2015= at 1:15 PM, Aaron Wood <woody77@gmail.com> wrote:
Toke,

I actual= ly tend to see a bit higher latency with ICMP at the higher percentiles.

http://burntchr= ome.blogspot.com/2014/05/fixing-bufferbloat-on-comcasts-blast.html
<= /div>
Although the biggest "boost" I've seen ICMP given= was on Free.fr's network:
=

-Aaron

On Sun= , Apr 19, 2015 at 11:30 AM, Toke H=C3=B8iland-J=C3=B8rgensen <toke@toke.dk&= gt; wrote:
Jonathan Morton &= lt;chromatix99@g= mail.com> writes:

>> Why not? They can be a quite useful measure of how competing traff= ic
>> performs when bulk flows congest the link. Which for many
>> applications is more important then the latency experienced by the=
>> bulk flow itself.
>
> One clear objection is that ICMP is often prioritised when UDP is not.=
> So measuring with UDP gives a better indication in those cases.
> Measuring with a separate TCP flow, such as HTTPing, is better still > by some measures, but most truly latency-sensitive traffic does use > UDP.

Sure, well I tend to do both. Can't recall ever actually seeing = any
performance difference between the UDP and ICMP latency measurements,
though...

-Toke
_______________________________________________
Bloat mailing list
Bloat@list= s.bufferbloat.net
= https://lists.bufferbloat.net/listinfo/bloat


_______________________________________________
Bloat mailing list
Bloat@list= s.bufferbloat.net
= https://lists.bufferbloat.net/listinfo/bloat



_______________________________________________
Bloat mailing list
Bloat@list= s.bufferbloat.net
= https://lists.bufferbloat.net/listinfo/bloat




--
Best regards / Mvh<= br>Jan Pedro Tumusok


--20cf301af6d9142fab0514237b49--