Re: [LARTC] Life sucks with a saturated link.

["Christian Worm Mortensen" <worm@dkik.dk>]

Hi,

> One way to obtain control over incoming traffic might be to place a
> virtual entity between the incoming interface and the local network.

I think we are many who wants to do that..

> [ Sanity check:  Will our traffic policies for the virtual interface ]
> [ be reflected back through the modem interface in a useful way?  Or ]
> [ will excess traffic just end up piling up in the virtual host's    ]
> [ outgoing queue, not affecting upstream's behaviour at all?         ] 

No. In theory it will be possible to do this, but I don't know if it is possible in Linux. 

I have for almost 2 years been running a seperate Linux box which does incomming scheduleing for a dorm with 500 people in this way with big success. That is, a qdisc is setup on the LAN interface of the machine and all traffic from the internet is passing through this interface. Now, the bandwidth for this interface has been setup to restricting internet-to-the-lan traffic to a rate a bit lower that the capacity of the WAN link. It is described in the documentation for the WRR qdsic. I think it will be very relevant to put some of this stuff in the howto.

> Now, the problem we have here is one I have not seen addressed in
> the how-to:  The outbound queue and the inbound queue are intimately
> related to each other.  If the outbound queue is empty, the inbound
> queue is free to use as much bandwidth as it likes.

Is the an inbound queue? 

> However, if the
> outbound queue contains low-latency traffic, or packets to establish
> new connections, the inbound queue needs to back off enough to let this
> traffic out

I don't understand this. To me inbound and outbound traffic is completly unrelated. We just want to shape in both directions.

> [1] ...without randomly dropping packets:
>       I think this makes the current ingress policer a poor choice.

Yes, you won't get any benefit from randomly dropping packets - in effect that will just make your links lower.

> [2] outbound and inbound...a common queue:
>       At least, I believe this is true for modem transports.  One can
>       easily imagine this might not be true for all transports.

Hmm.. You can't control when you receive a packet. At least not on ethernet. I don't imagine you can do that on a modem either - it is up to the sender to send.


Christian


_______________________________________________
LARTC mailing list / LARTC@mailman.ds9a.nl
http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://ds9a.nl/2.4Routing/