You could shape outbound on each side. If you do police the customer could just shape outbound from their end which would prevent drops
Thanks Darren http://www.mellowd.co.uk/ccie > Date: Fri, 29 Nov 2013 12:58:39 +0000 > From: [email protected] > To: [email protected] > CC: [email protected] > Subject: Re: [j-nsp] EoMPLS data rate > > Have you thought about traffic shaping instead? > > > On 29 November 2013 10:54, Vincent <[email protected]> wrote: > > > Hello all, > > > > we are trying to build for a customer an EoMPLS circuit of 10 Mbps over a > > GE interface. > > > > Configuration looks like this at this moment: > > > > dude@LAB-MX960> show configuration protocols l2circuit > > neighbor 172.16.50.1 { > > interface ge-0/1/1.0 { > > virtual-circuit-id 1; > > } > > } > > > > ... with something similar at the other end of the network. This is easy > > and works quite well. > > > > Next step is to limit the traffic to 10 Mbps. > > > > We were thinking of using a simple policer like: > > > > dude@LAB-MX960> show configuration firewall policer 10M_fwp > > if-exceeding { > > bandwidth-limit 10m; > > burst-size-limit 15k; > > } > > then discard; > > > > ... and apply it on the ingress interface, but maybe that's a little bit > > too rough (in terms of packet drops and TCP backoff)? Furthermore we are > > not sure on how to select the burst parameter. > > > > Is there any better way to achieve this ? Note that the customer should not > > be able to burst for a long time, and have this traffic above 10 Mbps go to > > a best-effort QoS: we just want to limit its traffic to a certain level > > without completely breaking its TCP connections. > > > > Does this all make sense? > > > > Thanks, > > > > Vincent > > _______________________________________________ > > juniper-nsp mailing list [email protected] > > https://puck.nether.net/mailman/listinfo/juniper-nsp > > > _______________________________________________ > juniper-nsp mailing list [email protected] > https://puck.nether.net/mailman/listinfo/juniper-nsp _______________________________________________ juniper-nsp mailing list [email protected] https://puck.nether.net/mailman/listinfo/juniper-nsp
