ok so it's not like the cef switched R2 is marking them just passing them
through and yes same de-list on all 3 routers

would this not mean that packets coming into an enterprise router from say
a LAN switch be processed switched therefore no real need to have cef
transit routers to initiate the marking, please correct me..

thanks Bob


On 8 October 2012 01:55, Bob McCouch <[email protected]> wrote:

> The counters show the number of frames that have been seen with the
> marking, not just the number marked by the local router.
>
> Do you have the same de-list on both R5 and R6?
>
> Bob
> --
> Sent from my iPhone, please excuse any typos.
>
> On Oct 7, 2012, at 8:45 PM, Tony Singh <[email protected]> wrote:
>
>
> Hi Bob
>
> Thanks for the awesome reply really appreciate it.
>
> here's what I see, if R2 is purely transit (HUB & cef switched) should I
> be seeing the de packets incrementing on R2? or is it a case that the
> marking has already been done by R5 and R2 is simply incrementing the
> counters?
>
> *Ping from R5 to R6 via R2 HUB*
>
> *R5#ping 150.100.100.6*
>
> Type escape sequence to abort.
> Sending 5, 100-byte ICMP Echos to 150.100.100.6, timeout is 2 seconds:
> !!!!!
> Success rate is 100 percent (5/5), round-trip min/avg/max = 112/114/124 ms
> R5#ping 150.100.100.6
>
> Prior to ping
> *R5#show frame-relay pvc 502*
> in DE pkts 59            out DE pkts 79
> Post
> in DE pkts 60            out DE pkts 84
>
> Prior to ping
> *R2#show frame-relay pvc 205*
> in DE pkts 79            out DE pkts 59
> Post
> in DE pkts 84            out DE pkts 60
>
> Prior to ping
> *R2#show frame-relay pvc 206*
> in DE pkts 63            out DE pkts 23
> Post
> in DE pkts 68            out DE pkts 24
>
> Prior to ping
> *R6#show frame-relay pvc 602*
> in DE pkts 23            out DE pkts 63
> Post
> in DE pkts 24            out DE pkts 68
>
>
> Thanks in advance
>
> Tony
>
>
>
>
>
> On 8 October 2012 01:10, Bob McCouch <[email protected]> wrote:
>
>> Hi Tony,
>>
>> It was me that mentioned the de-list vs. MQC behavior on the list.
>> Remember that regardless of CEF settings on an interface, traffic to/from
>> the CPU will be process switched. So if you're pinging from R5 and the
>> de-list is set up on R5, those packets are being process switched.
>>
>> See the documentation on the feature for confirmation:
>>
>>
>> http://www.cisco.com/en/US/docs/ios-xml/ios/wan/command/wan-f1.html#GUID-6160F4F7-0B57-4EF9-9AE8-F18B4FCF8751
>>
>> Specifically, it states:
>>
>> "Frame Relay DE group functionality is supported on process-switched
>> packets only"
>>
>> When I tested, my network looked like this:
>>
>> R1--eth--R2--FR--R3
>>
>> When pings or HTTP traffic were originated from R2 to R3, the DE bit was
>> set. For the same traffic originated from R1 (and this CEF switched, pure
>> transit traffic to R2) the DE list did not take effect.
>>
>> Regardling the ACL match, you must remember that in many places IOS uses
>> an ACL not as a traffic filter, but as a selector. Something that matches a
>> permit statement is selected, and something that hits a deny statement is
>> not selected, but not filtered. This is done in QOS class maps, route-maps
>> for things like policy-routing, etc. So in your example, the "deny"
>> statements simply stop HTTP traffic from being selected, and as they are
>> not permitted by the ACL and thus nor matched by the DE-list.
>>
>> The two HTTP statements catch traffic that might be originating from a
>> client or a server in either direction. After all, "permit tcp any any eq
>> 80" would only match traffic with a TCP source port of anything, and a TCP
>> destination port of 80. In other words, traffic from client to server. To
>> catch return traffic from server to client, you need to select traffic that
>> has a TCP source port of 80.
>>
>> Hope that helps,
>> Bob
>>
>>
>>
>> On Sun, Oct 7, 2012 at 7:28 PM, Tony Singh <[email protected]> wrote:
>>
>>> Hi George
>>>
>>> I've labbed this with a hub and spoke topology with cef enabled by
>>> default on all frame routers and I'm getting the icmp traffic de marked ok
>>> both inbound and outbound as it first passes from R5 DLCI 502 > R2 DLCI 205
>>> repackaged to DLCI 206 > R6 DLCI 602
>>>
>>> I can see the ip cef interface mappings on all 3 routers, my local icmp
>>> pings seem to use cef although no confirmation of routed via FIB (debug ip
>>> packet) but the replies back using the source ip address of the HUB R2
>>> shows routed via RIB, is this expected?
>>>
>>> The HUB R2 in the debug just shows these ip packets being redirected as
>>> it uses the same interface for receiving and sending the traffic.
>>>
>>> Thanks in advance..
>>>
>>> --
>>> BR
>>>
>>> Tony
>>>
>>> Sent from my iPhone on 3
>>>
>>> On 7 Oct 2012, at 17:13, George Leslie <[email protected]>
>>> wrote:
>>>
>>> >
>>> >  Hi Tony,
>>> > Yes I believe that is correct, but advise to lab it up.  On the other
>>> side of the frame DLCI, put a policy map on there to match on the de-bit.
>>> >
>>> > This is from memory and it was a long time ago, so I hope I've not
>>> misled you!! It could also be IOS dependent.
>>> >
>>> > As for the access-list, a packet that is a www packet will first hit
>>> one of the two denys and will therefore, be denied from the de-list and
>>> hence from marking on DLCIs 205 and 206.  Other TCP or UDP or ICMP or IP
>>> traffic, will not match the first two denies, and will fall through to the
>>> permit, so WILL get de-marked (in theory, remembering that it might not due
>>> to the process vs. CEF switching thing).
>>> >
>>> > I think it needs to be labbed up to verify.
>>> >
>>> > G.
>>> > Subject: Re: [OSL | CCIE_RS] V1 task 6.16 FR
>>> > From: [email protected]
>>> > Date: Sun, 7 Oct 2012 17:02:06 +0100
>>> > To: [email protected]; [email protected]
>>> >
>>> > Hi George
>>> >
>>> > Thanks for the detailed reply so traffic generated from the router
>>> will always get marked but traffic going through (cef) the router would
>>> not? Will lab it later...
>>> >
>>> > WRT defining the de-list where I say match "ip" from access list "101"
>>> how does this ignore the "deny www" traffic would matching "ip" not
>>> automatically match any www traffic too? Or would the "deny www" explicitly
>>> not set the de bit for this traffic , struggling to understand this bit
>>> (pardon pun)
>>> >
>>> > --
>>> > BR
>>> >
>>> > Tony
>>> >
>>> > Sent from my iPhone on 3
>>> >
>>> > On 7 Oct 2012, at 16:22, George Leslie <[email protected]>
>>> wrote:
>>> >
>>> > Hi Tony,
>>> > I don't have the lab book in front of me, so this is from memory!!
>>> >
>>> > Your config will set the DE bit in the frame relay frame, for all IP
>>> packets, other than HTTP packets, sent out DLCI 205 or 206.
>>> >
>>> > You need the second deny as the HTTP packets could either be heading
>>> out towards a server, or coming back from a server.  If the task does not
>>> state where the http server is, then you need to accound for both.
>>> >
>>> > First deny would be outbound towards a server that lives the other
>>> side of the frame network; second deny would mean the server is on your
>>> side of the frame network.
>>> >
>>> > There is a slight gotcha with this command, that came up on this forum
>>> last year.  This command actually only works for packets that are process
>>> switched, not CEF or fast switches.  So, in the real lab, this is an "ask
>>> the proctor moment", as disabling CEF on the interface may disable other
>>> necessary functions.
>>> >
>>> > So, you'd have to ask the proctor if they are only looking for the
>>> command in the config, or if you should be using MQC syntax to achieve the
>>> same thing.
>>> >
>>> > HTH, George.
>>> >
>>> > > Date: Sun, 7 Oct 2012 04:06:59 +0100
>>> > > From: [email protected]
>>> > > To: [email protected]
>>> > > Subject: [OSL | CCIE_RS] V1 task 6.16 FR
>>> > >
>>> > > Hi
>>> > >
>>> > > Sorry for the dumb question but I'm struggling to understand the
>>> logic when
>>> > > de bit is set to on for the following traffic....
>>> > >
>>> > > frame-relay de-list 1 protocol ip list 101
>>> > >
>>> > > frame-relay de-group 1 205
>>> > > frame-relay de-group 1 206
>>> > >
>>> > > access-list 101 deny tcp any any eq www
>>> > > access-list 101 deny tcp any eq www any
>>> > > access-list 101 permit ip any any
>>> > >
>>> > > i.e de bit set to on, on the frame-relay frame > we mark the ip
>>> packets for
>>> > > de > passed to service provider who can discard at times of
>>> congestion
>>> > >
>>> > > how is deny tcp www "allowed on the CIR" against this de-list using
>>> this
>>> > > reverse logic, how does this work technically? is it that de-list
>>> looks for
>>> > > only a permit statement in the referenced acl & any deny it ignores?
>>> > >
>>> > > and why do we use the second acl statement, is this not covered by
>>> the
>>> > > first acl statement?
>>> > >
>>> > > BR
>>> > >
>>> > > Tony
>>> > > _______________________________________________
>>> > > For more information regarding industry leading CCIE Lab training,
>>> please visit www.ipexpert.com
>>> > >
>>> > > Are you a CCNP or CCIE and looking for a job? Check out
>>> www.PlatinumPlacement.com
>>> > >
>>> > > http://onlinestudylist.com/mailman/listinfo/ccie_rs
>>> _______________________________________________
>>> For more information regarding industry leading CCIE Lab training,
>>> please visit www.ipexpert.com
>>>
>>> Are you a CCNP or CCIE and looking for a job? Check out
>>> www.PlatinumPlacement.com
>>>
>>> http://onlinestudylist.com/mailman/listinfo/ccie_rs
>>>
>>
>>
>
_______________________________________________
For more information regarding industry leading CCIE Lab training, please visit 
www.ipexpert.com

Are you a CCNP or CCIE and looking for a job? Check out 
www.PlatinumPlacement.com

http://onlinestudylist.com/mailman/listinfo/ccie_rs

Reply via email to