[LARTC] wrr question

[LARTC] wrr question

[Kenneth Kalmer]

Guys

All the recent discussions recently, and the knowledge of a 2.6 port,
of WRR has made me very keen on trying it. I had a look at the docs
and examples know but my mind is not in a very receptive state.

Take this simple example.

Incoming internet connection of 1mbps. Shared between up to 25 users
simultaneously.

I know that WRR can fairly distribute the traffic amongst the
currently connected clients at any specific time. I'd like to know how
can I restrict any client from getting more than 256kbps (or 25%) of
the total link speed, even when they are the only users.

Kind regards

-- 

Kenneth Kalmer
kenneth.kalmer@gmail.com
http://opensourcery.blogspot.com
_______________________________________________
LARTC mailing list
LARTC@mailman.ds9a.nl
http://mailman.ds9a.nl/cgi-bin/mailman/listinfo/lartc

Re: [LARTC] wrr question

[Peter Surda]

On Tue, 7 Jun 2005 21:35:03 +0200 Kenneth Kalmer <kenneth.kalmer@gmail.com>
wrote:

>Guys
hi

>I know that WRR can fairly distribute the traffic amongst the
>currently connected clients at any specific time. I'd like to know how
>can I restrict any client from getting more than 256kbps (or 25%) of
>the total link speed, even when they are the only users.
You can't, at least not easily. However, unless you have specific reasons to do
so, it doesn't matter.

>Kind regards
Yours sincerely,
Peter
_______________________________________________
LARTC mailing list
LARTC@mailman.ds9a.nl
http://mailman.ds9a.nl/cgi-bin/mailman/listinfo/lartc

Re: [LARTC] wrr question

[Jonathan Day]

Hi,

For something like this, where you're wanting to do
bandwidth capping, you're probably better off with
something like CBQ, which supports limits.

It sounds like you want soft limits of 4% (a fair
slice, when 25 users are present) and hard limits of
25%.

Another option would be to use WRR and then use
pattern-matching in Netfilter to set the hard limit.

Part of the problem is that there are a very large
number of "Quality of Service" protocols, of which
Linux supports some, but that there is no really clear
cheat-sheet on what to use when, what works well with
what, and what capabilities each QoS method has.

Jonathan

--- Kenneth Kalmer <kenneth.kalmer@gmail.com> wrote:

> Guys
> 
> All the recent discussions recently, and the
> knowledge of a 2.6 port,
> of WRR has made me very keen on trying it. I had a
> look at the docs
> and examples know but my mind is not in a very
> receptive state.
> 
> Take this simple example.
> 
> Incoming internet connection of 1mbps. Shared
> between up to 25 users
> simultaneously.
> 
> I know that WRR can fairly distribute the traffic
> amongst the
> currently connected clients at any specific time.
> I'd like to know how
> can I restrict any client from getting more than
> 256kbps (or 25%) of
> the total link speed, even when they are the only
> users.
> 
> Kind regards
> 
> -- 
> 
> Kenneth Kalmer
> kenneth.kalmer@gmail.com
> http://opensourcery.blogspot.com
> _______________________________________________
> LARTC mailing list
> LARTC@mailman.ds9a.nl
>
http://mailman.ds9a.nl/cgi-bin/mailman/listinfo/lartc
> 


__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com 
_______________________________________________
LARTC mailing list
LARTC@mailman.ds9a.nl
http://mailman.ds9a.nl/cgi-bin/mailman/listinfo/lartc

Re: [LARTC] wrr question

[Kenneth Kalmer]

Jonathan

I've got a full HTB setup currently, and I'm just looking at other
options for greener pastures. The thing at the moment is that we offer
internet to students, and South African bandwidth is probably the most
expensive in the world... We want to limit things to stop abusive
behaviour even before it begins...

My main concern is that since we do overschedule the available
bandwidth, I need everyone's bandwidth to degrade equally, depending
on the amount of users. Currently I basically set my rules like this:

rate = total rate / num of users
ceil = total rate / 8

I'm just worried about the equal degrading of everyones links..

Anycase, I just thought that WRR might offer a smoother experience for
the users, but I'll stick with my HTB setup for now since it's working
beautifully. If anybody has an alternative suggestion, please shout.

Thanks for the reply though.

On 6/7/05, Jonathan Day <imipak@yahoo.com> wrote:
> Hi,
> 
> For something like this, where you're wanting to do
> bandwidth capping, you're probably better off with
> something like CBQ, which supports limits.
> 
> It sounds like you want soft limits of 4% (a fair
> slice, when 25 users are present) and hard limits of
> 25%.
> 
> Another option would be to use WRR and then use
> pattern-matching in Netfilter to set the hard limit.
> 
> Part of the problem is that there are a very large
> number of "Quality of Service" protocols, of which
> Linux supports some, but that there is no really clear
> cheat-sheet on what to use when, what works well with
> what, and what capabilities each QoS method has.
> 
> Jonathan
> 
> --- Kenneth Kalmer <kenneth.kalmer@gmail.com> wrote:
> 
> > Guys
> >
> > All the recent discussions recently, and the
> > knowledge of a 2.6 port,
> > of WRR has made me very keen on trying it. I had a
> > look at the docs
> > and examples know but my mind is not in a very
> > receptive state.
> >
> > Take this simple example.
> >
> > Incoming internet connection of 1mbps. Shared
> > between up to 25 users
> > simultaneously.
> >
> > I know that WRR can fairly distribute the traffic
> > amongst the
> > currently connected clients at any specific time.
> > I'd like to know how
> > can I restrict any client from getting more than
> > 256kbps (or 25%) of
> > the total link speed, even when they are the only
> > users.
> >
> > Kind regards
> >
> > --
> >
> > Kenneth Kalmer
> > kenneth.kalmer@gmail.com
> > http://opensourcery.blogspot.com
> > _______________________________________________
> > LARTC mailing list
> > LARTC@mailman.ds9a.nl
> >
> http://mailman.ds9a.nl/cgi-bin/mailman/listinfo/lartc
> >
> 
> 
> __________________________________________________
> Do You Yahoo!?
> Tired of spam?  Yahoo! Mail has the best spam protection around
> http://mail.yahoo.com
> 


-- 

Kenneth Kalmer
kenneth.kalmer@gmail.com
http://opensourcery.blogspot.com
_______________________________________________
LARTC mailing list
LARTC@mailman.ds9a.nl
http://mailman.ds9a.nl/cgi-bin/mailman/listinfo/lartc

Re: [LARTC] wrr question

[Peter Surda]

On Tue, 7 Jun 2005 23:33:20 +0200 Kenneth Kalmer <kenneth.kalmer@gmail.com>
wrote:

>Anycase, I just thought that WRR might offer a smoother experience for
>the users, but I'll stick with my HTB setup for now since it's working
>beautifully. If anybody has an alternative suggestion, please shout.
In case you pay by the amount of transferred data, WRR is probably not the right
choice.

Yours sincerely,
Peter
_______________________________________________
LARTC mailing list
LARTC@mailman.ds9a.nl
http://mailman.ds9a.nl/cgi-bin/mailman/listinfo/lartc

Re: [LARTC] wrr question

[Kenneth Kalmer]

Hmm, just has an idea, dunno if this will work...

Can I use WRR on an IMQ disc to make sure that the incoming traffic is
not saturated by a single squid request? Squid runs transparently, and
I've noticed that it downloads the file faster than the client gets it
from squid, so big downloads can very easily congest the link...

Possible?

On 6/7/05, Kenneth Kalmer <kenneth.kalmer@gmail.com> wrote:
> Jonathan
> 
> I've got a full HTB setup currently, and I'm just looking at other
> options for greener pastures. The thing at the moment is that we offer
> internet to students, and South African bandwidth is probably the most
> expensive in the world... We want to limit things to stop abusive
> behaviour even before it begins...
> 
> My main concern is that since we do overschedule the available
> bandwidth, I need everyone's bandwidth to degrade equally, depending
> on the amount of users. Currently I basically set my rules like this:
> 
> rate = total rate / num of users
> ceil = total rate / 8
> 
> I'm just worried about the equal degrading of everyones links..
> 
> Anycase, I just thought that WRR might offer a smoother experience for
> the users, but I'll stick with my HTB setup for now since it's working
> beautifully. If anybody has an alternative suggestion, please shout.
> 
> Thanks for the reply though.
> 
> On 6/7/05, Jonathan Day <imipak@yahoo.com> wrote:
> > Hi,
> >
> > For something like this, where you're wanting to do
> > bandwidth capping, you're probably better off with
> > something like CBQ, which supports limits.
> >
> > It sounds like you want soft limits of 4% (a fair
> > slice, when 25 users are present) and hard limits of
> > 25%.
> >
> > Another option would be to use WRR and then use
> > pattern-matching in Netfilter to set the hard limit.
> >
> > Part of the problem is that there are a very large
> > number of "Quality of Service" protocols, of which
> > Linux supports some, but that there is no really clear
> > cheat-sheet on what to use when, what works well with
> > what, and what capabilities each QoS method has.
> >
> > Jonathan
> >
> > --- Kenneth Kalmer <kenneth.kalmer@gmail.com> wrote:
> >
> > > Guys
> > >
> > > All the recent discussions recently, and the
> > > knowledge of a 2.6 port,
> > > of WRR has made me very keen on trying it. I had a
> > > look at the docs
> > > and examples know but my mind is not in a very
> > > receptive state.
> > >
> > > Take this simple example.
> > >
> > > Incoming internet connection of 1mbps. Shared
> > > between up to 25 users
> > > simultaneously.
> > >
> > > I know that WRR can fairly distribute the traffic
> > > amongst the
> > > currently connected clients at any specific time.
> > > I'd like to know how
> > > can I restrict any client from getting more than
> > > 256kbps (or 25%) of
> > > the total link speed, even when they are the only
> > > users.
> > >
> > > Kind regards
> > >
> > > --
> > >
> > > Kenneth Kalmer
> > > kenneth.kalmer@gmail.com
> > > http://opensourcery.blogspot.com
> > > _______________________________________________
> > > LARTC mailing list
> > > LARTC@mailman.ds9a.nl
> > >
> > http://mailman.ds9a.nl/cgi-bin/mailman/listinfo/lartc
> > >
> >
> >
> > __________________________________________________
> > Do You Yahoo!?
> > Tired of spam?  Yahoo! Mail has the best spam protection around
> > http://mail.yahoo.com
> >
> 
> 
> --
> 
> Kenneth Kalmer
> kenneth.kalmer@gmail.com
> http://opensourcery.blogspot.com
> 


-- 

Kenneth Kalmer
kenneth.kalmer@gmail.com
http://opensourcery.blogspot.com
_______________________________________________
LARTC mailing list
LARTC@mailman.ds9a.nl
http://mailman.ds9a.nl/cgi-bin/mailman/listinfo/lartc

Re: [LARTC] wrr question

[Peter Surda]

On Wed, 8 Jun 2005 22:01:42 +0200 Kenneth Kalmer <kenneth.kalmer@gmail.com>
wrote:

>Hmm, just has an idea, dunno if this will work...
>
>Can I use WRR on an IMQ disc
Yes, this works without problems (Route Hat's tc script does it).

>to make sure that the incoming traffic is
>not saturated by a single squid request? Squid runs transparently, and
>I've noticed that it downloads the file faster than the client gets it
>from squid, so big downloads can very easily congest the link...
This indicates an incorrect setup. Limiting squid's connections to clients won't
have the expected effect, the connections are handled asynchronously.

>Possible?
Now, combining WRR and Squid is another topic. It is possible, but difficult.

On the WRR website, there is a program called "proxyremap" that should solve
this in userspace. I never tried it, but it is supposed to work.

The other option is to use squid with tproxy patch. This requires a
rearrangement of the network setup though (tproxy and NAT don't work on the same
machine, and because tproxy mangles IP but not MAC, you have to put it on a
separate segment or use it as a next hop or a bridge, or use other tricks, such
as arptables' MAC mangling).

I only tried the last one (tproxy + mac mangling), for about 10 days, on a
network with about 60 local computers, and it was a horrible hack, but worked.
The reason that I only used it for such a short time wasn't that there were
problems, I just wanted to test it with Route Hat in case some customers request
it. 

Unfortunately, TPROXY mailing list mentions a couple of times that rewriting
TPROXY so that it works with NAT isn't easy, so for foreseeable future we're
stuck with the above "solutions" (well, most of them are workarounds).

Other than that, you can play with squid's delay pools. Unless however you
fine-tune the proxies priority (in my experience very difficult) you're still
screwed.

Yours sincerely,
Peter
_______________________________________________
LARTC mailing list
LARTC@mailman.ds9a.nl
http://mailman.ds9a.nl/cgi-bin/mailman/listinfo/lartc

Re: [LARTC] wrr question

[Kenneth Kalmer]

Thanks for the blunted honesty... I'll give the delay pools a shot
rather than having multiple boxes doing the work...

Regards,

On 6/8/05, Peter Surda <shurdeek@routehat.org> wrote:
> On Wed, 8 Jun 2005 22:01:42 +0200 Kenneth Kalmer <kenneth.kalmer@gmail.com>
> wrote:
> 
> >Hmm, just has an idea, dunno if this will work...
> >
> >Can I use WRR on an IMQ disc
> Yes, this works without problems (Route Hat's tc script does it).
> 
> >to make sure that the incoming traffic is
> >not saturated by a single squid request? Squid runs transparently, and
> >I've noticed that it downloads the file faster than the client gets it
> >from squid, so big downloads can very easily congest the link...
> This indicates an incorrect setup. Limiting squid's connections to clients won't
> have the expected effect, the connections are handled asynchronously.
> 
> >Possible?
> Now, combining WRR and Squid is another topic. It is possible, but difficult.
> 
> On the WRR website, there is a program called "proxyremap" that should solve
> this in userspace. I never tried it, but it is supposed to work.
> 
> The other option is to use squid with tproxy patch. This requires a
> rearrangement of the network setup though (tproxy and NAT don't work on the same
> machine, and because tproxy mangles IP but not MAC, you have to put it on a
> separate segment or use it as a next hop or a bridge, or use other tricks, such
> as arptables' MAC mangling).
> 
> I only tried the last one (tproxy + mac mangling), for about 10 days, on a
> network with about 60 local computers, and it was a horrible hack, but worked.
> The reason that I only used it for such a short time wasn't that there were
> problems, I just wanted to test it with Route Hat in case some customers request
> it.
> 
> Unfortunately, TPROXY mailing list mentions a couple of times that rewriting
> TPROXY so that it works with NAT isn't easy, so for foreseeable future we're
> stuck with the above "solutions" (well, most of them are workarounds).
> 
> Other than that, you can play with squid's delay pools. Unless however you
> fine-tune the proxies priority (in my experience very difficult) you're still
> screwed.
> 
> Yours sincerely,
> Peter
> _______________________________________________
> LARTC mailing list
> LARTC@mailman.ds9a.nl
> http://mailman.ds9a.nl/cgi-bin/mailman/listinfo/lartc
> 


-- 

Kenneth Kalmer
kenneth.kalmer@gmail.com
http://opensourcery.blogspot.com
_______________________________________________
LARTC mailing list
LARTC@mailman.ds9a.nl
http://mailman.ds9a.nl/cgi-bin/mailman/listinfo/lartc