From: ccie2be (ccie2be@nyc.rr.com)
Date: Tue Apr 05 2005 - 19:17:25 GMT-3
Simon,
How interesting.
Unfortunately, but as usual, the Doc-CD sheds no light on this. Did you
discover this by testing?
I'm still trying to think of a scenario where using the offset command is
the only way to achieve the required result, but can't think of any.
Have you come across any eigrp scenario's which would require this command
to be used?
Tim
-----Original Message-----
From: simon hart [mailto:simon.hart@btinternet.com]
Sent: Tuesday, April 05, 2005 6:05 PM
To: ccie2be; 'John Matus'; ccielab@groupstudy.com
Subject: RE: blocking eigrp routes
Tim,
I believe the offset is added to the delay component. The offset cannot be
the composite metric as the TLV's that eigrp advertised do not include the
composite value, but the delay, bandwidth, MTU, Load and Reliability (MTU
not used though!!). This is why the K values have to be the same througout
the domain.
Now what I said below is was slightly incorrect (whoops!!). If the
originating route has a delay function of 1000, an offset list of 255 will
in fact add 255 to that delay function, thus the delay will now be
advertised as 1255 to the next hop.
Simon
-----Original Message-----
From: ccie2be [mailto:ccie2be@nyc.rr.com]
Sent: 05 April 2005 22:50
To: 'simon hart'; 'John Matus'; ccielab@groupstudy.com
Subject: RE: blocking eigrp routes
Hi Simon,
Excellent post!!!!
But, you raise an interesting question.
Given that eigrp has 2 types of metrics: bandwidth, delay, etc, and
composite, I assume that the metric used with the offset command is the
composite version, right?
Now, normally with eigrp, whenever you manipulate metric values you're
manipulating the component (b/w, delay, etc,) values of the metric, not the
composite itself. So, what happens to the component values when you
manipulate the composite instead of the component values?
Does eigrp use it's formula for computing the composite metric in reverse
when the composite itself is changed?
Also, given what you've said, it seems like there's no good reason for using
the offset-list command with eigrp.
If you wanted or needed to change Eigrp's metric, for example, for the
purpose of load-balancing, you would change one of the component metrics
such as bandwidth or delay on the interfaces to do so rather than manipulate
the composite metric with the offset command.
So, what reason might there be for using the offset command?
TIA, Tim
-----Original Message-----
From: nobody@groupstudy.com [mailto:nobody@groupstudy.com] On Behalf Of
simon hart
Sent: Tuesday, April 05, 2005 5:04 PM
To: John Matus; ccielab@groupstudy.com
Subject: RE: blocking eigrp routes
Hi John,
You cannot use the offset list to make a route(s) unreachable with eigrp,
although you can do this with rip.
Rip has the philosophy that an infinite metric......anything over 15..... is
unreachable, therefore making the metric 16 effectively makes those routes
effected by the offset list unreachable.
Eigrp does not have the same philosophy, the composite metric in eigrp can
be anything and as far as the protocol is concerned it is still reachable
(within the confines fo FD's and FS's etc. etc.)
The example you have given in RIP environment would dictate that every route
advertised out of that interface would unreachable (bearing in mind that you
could not use 255 in RIP, it would have to be 16). This really would seem a
little pointless, may as well make the interface passive.
If you used the example below for eigrp you would be advertising eigrp with
a metric of 255, and probably making those routes the most favourable.
Eigrp uses a composite metric, derived by default as function of bandwidth
and delay. The metrics are normally quite high, in fact a lot higher than
255 (for information the highest eigrp figure you can use on the offset list
is 2147483647).
In order not to advertise to the next hop router you could use:
router eigrp 100
distribute-list 10 out s0/0
access-list 10 deny 0.0.0.0 255.255.255.255
Change the access list if you wish to deny a subset of routes.
HTH
Simon
-----Original Message-----
From: nobody@groupstudy.com [mailto:nobody@groupstudy.com]On Behalf Of
John Matus
Sent: 05 April 2005 21:30
To: ccielab@groupstudy.com
Subject: blocking eigrp routes
if i wanted to block all eigrp routes from exiting to the next hop would it
be:
offset-list 1 out 255 s0/0
access-l 1 permit 0.0.0.0 255.255.255.255 ?
i wasn't sure since rip's metric would be '16' for offset but i'm not sure
about eigrp.
thanks in advance
This archive was generated by hypermail 2.1.4 : Tue May 03 2005 - 07:54:53 GMT-3