LARTC - May 2005 - how to guarantee 1/numflows bandwidth to each flow dynamically

I am looking for a simple way to guarantee to each flow
going through my traffic control point 1/numflows of
bandwidth.  I thought using SFQ would do this effectively
but it appears to be quite unfair: a very high speed
download that fills the pipe easily starves smaller flows to
the point where it becomes unusable (especially if they are
at all interactive)

Because numflows is dynamic, I''m not sure how I would have
the bandwidth allocated to each flow change dynamically and
automatically as flows are added and removed.

Anyone have an idea how to implement this? Thanks.

if u find sfq unfair u can try WRR.
but if for example one flow is eating all the badwidth and a new flow 
starts, the new has to "claim" he's part by going from slow
transfer to fast
one. 
u can try making normal QOS based on classes and give one class to every 
customer. should work best. remember to do the same on both upload and 
download

On 5/9/05, Scott Mcdermott <lartc@vaxerdec.net>
wrote:
> 
> I am looking for a simple way to guarantee to each flow
> going through my traffic control point 1/numflows of
> bandwidth. I thought using SFQ would do this effectively
> but it appears to be quite unfair: a very high speed
> download that fills the pipe easily starves smaller flows to
> the point where it becomes unusable (especially if they are
> at all interactive)
> 
> Because numflows is dynamic, I'm not sure how I would have
> the bandwidth allocated to each flow change dynamically and
> automatically as flows are added and removed.
> 
> Anyone have an idea how to implement this? Thanks.
> 
> _______________________________________________
> LARTC mailing list
> LARTC@mailman.ds9a.nl
> http://mailman.ds9a.nl/cgi-bin/mailman/listinfo/lartc
> 


-- 
Miłego Dnia
Krystian Antoni


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

Scott Mcdermott wrote:
> 
> I am looking for a simple way to guarantee to each flow
> going through my traffic control point 1/numflows of
> bandwidth.  I thought using SFQ would do this effectively
> but it appears to be quite unfair: a very high speed
> download that fills the pipe easily starves smaller flows to
> the point where it becomes unusable (especially if they are
> at all interactive)
> 
> Because numflows is dynamic, I''m not sure how I would have
> the bandwidth allocated to each flow change dynamically and
> automatically as flows are added and removed.
> 
> Anyone have an idea how to implement this? Thanks.
Perhaps you should look at esfq rather than sfq.  esfq lets you specify
quantum in bytes, depth in flows, limit in packets, and a source or dest
hash.  Don''t apply any sfq to what Wondershaper calls
"interactive".

You need to match esfq to the kernel version; 2.6 uses a different esfq
than 2.4.
--
buck

Scott Mcdermott wrote:
> I am looking for a simple way to guarantee to each flow
> going through my traffic control point 1/numflows of
> bandwidth.  I thought using SFQ would do this effectively
> but it appears to be quite unfair: a very high speed
> download that fills the pipe easily starves smaller flows to
> the point where it becomes unusable (especially if they are
> at all interactive)
If you are shaping from the "wrong" end of the bottleneck then you
need
to back off from link speed to get control - and it still won''t be
perfect.

It''s best to priorotise interactive traffic and not mix it with bulk 
traffic.

Andy.
> 
> Because numflows is dynamic, I''m not sure how I would have
> the bandwidth allocated to each flow change dynamically and
> automatically as flows are added and removed.
> 
> Anyone have an idea how to implement this? Thanks.
> 
> _______________________________________________
> LARTC mailing list
> LARTC@mailman.ds9a.nl
> http://mailman.ds9a.nl/cgi-bin/mailman/listinfo/lartc
>