Re: Qos over GRE using 'pre-classify' command

From: Roman Rodichev (roman@iementor.com)
Date: Thu Mar 12 2009 - 11:09:22 ARST


you need preclassify when your classmap is matching traffic based on
src/dst address, protocol or port. If you only matching based on tos
byte, then you don't need to preclassify, because the tos byte is
always copied from the original header to the encapsulating header

On Mar 12, 2009, at 7:50 AM, "Hunt, Richard (LDN-GIS)" <Richard.Hunt@interpublic.com
> wrote:

> Thanks Lejoe,
>
> I just realised that in my case ToS is copied as you say to the GRE
> header and
> each hop (which is providers MPLS cloud)does nothing with it to
> prioritise it
> because I haven't asked them to. Therefore unless I engage the
> provider, I
> get no guarantees , doh!
>
> I see my markings at both ends so the pre-classify part isn't needed
> as you
> rightly say.
>
> The question for me remains about the when to use pre-classify
> exactly. The
> link that you sent just re-gurgitates the Cisco link I sent before.
> If ToS is
> Ok and passed on to the new GRE header, then what sort of markings
> are not OK
> ?, and therefore need this feature?
>
> Quote:
> Apply the policy to a physical interface and enable qos-preclassify
> on a
> tunnel interface when you want to classify packets based on the pre-
> tunnel
> header.
>
> Thanks,
> Rich.
>
> From: Styran [mailto:styran@gmail.com]
> Sent: 12 March 2009 12:06
> To: Hunt, Richard (LDN-GIS)
> Cc: ccielab@groupstudy.com
> Subject: Re: Qos over GRE using 'pre-classify' command
>
> Hi Richard,
>
> If you're applying the service-policy command under the tunnel
> interface, you
> dont need to add the qos-preclassify under the tunnel interface. IOS
> copies
> the DS-Field or TOS byte of the original IP header to the new the IP
> header (
> eg: in case of GRE tunnel, when the original packet is
> encapsulated), so
> intermediate nodes can examine this field to apply appropriate QOS
> policies.
>
> As to when to use the qos-preclassify, I believe this might help
>
> http://ccieeverest.blogspot.com/2008/12/qos-pre-classify.html
>
> Lejoe
> On Thu, Mar 12, 2009 at 9:52 PM, Hunt, Richard (LDN-GIS)
> <Richard.Hunt@interpublic.com<mailto:Richard.Hunt@interpublic.com>>
> wrote:
> Dear All,
>
> I have backup WAN over a provider MPLS that uses GRE tunnels (but
> it's just a
> network with GREs nothing complex). It's hub and spoke in essence.
> I have a
> service-policy that matches my VoIP traffic and is applied to each
> Tunnel
> interface. As far as I could determine it was all marking and
> sending the
> VoIP using LLQ ( I used Netflow to verify this). Now I read about the
> 'pre-classify' interface command and I'm a little confused as to
> whether my
> receiving tunnel interfaces are able to pass on the original QoS
> marking
> (based on DSCP) and therefore if the rest of the LAN behind can
> prioritise
> this traffic.
>
> My GRE tunnel service-policy is applied outbound on all sites
> including the
> hub site and that seems to give priority based on my match criteria.
> Here's the CCO link that has me confused.
> http://www.cisco.com/en/US/tech/tk543/tk545/technologies_tech_note09186a00801
> 7405e.shtml#qoscomm<http://www.cisco.com/en/US/tech/tk543/tk545/technologies_
> tech_note09186a00801%0A7405e.shtml#qoscomm>
>
> Can anyone shed any light, has anyone sniffed this on say a LAN
> Ethernet of a
> router that has a GRE to another site? Who knows when you'd use
> 'pre-classify'?
>
> Thanks in advance.
> Richard.
>
>
>
>
> ________________________________
> Interpublic GIS (UK) Limited | Registered office: Ground Floor, 84
> Eccleston
> Square, London SW1V 1PX, England | Registered number: 2724363
> This message contains information which may be confidential and
> privileged.
> Unless you are the intended recipient (or authorized to receive
> this message for the intended recipient), you may not use, copy,
> disseminate
> or disclose to anyone the message or any information contained
> in the message. If you have received the message in error, please
> advise the
> sender by reply e-mail, and delete the message.
> Thank you very much.
>
>
> Blogs and organic groups at http://www.ccie.net
>
> _______________________________________________________________________
 

> Subscription information may be found at:
> http://www.groupstudy.com/list/CCIELab.html
>
>
> Blogs and organic groups at http://www.ccie.net
>
> _______________________________________________________________________
 

> Subscription information may be found at:
> http://www.groupstudy.com/list/CCIELab.html

Blogs and organic groups at http://www.ccie.net



This archive was generated by hypermail 2.1.4 : Mon Apr 06 2009 - 06:44:04 ART