distance vs metric with int/ext EIGRP

[tato386]

I was under the impression that the admin distance was only used to make routing decisions between different types of routing protocols like RIP vs EIGRP, IGRP vs BGP and so forth.  Howerver I have included an example printout in which apparenly the admin distance is used to make a routing decision between internal and external EIGRP routes.  It seems like if internal and external EIGRP is treated like two different routing protocols.  Is this correct?

In the example below the router is configured so that internal EIGRP has a higher admin distance than external.  Specifically, the command "distance eigrp 180 175" is configured on this particular router.  In example #1 the topology shows two routes to 172.30.6.0 with the external route having a much higer metric BUT the route is external with a distance of 175 so it gets installed in the routing table.  So apparently the metric is ignored and the admin distance is used regardless of the metric.  Example #2 shows the other, internal route with admin distance 180 being installed once the external route is removed.

Am I interpreting this correctly?  Is there a way of modifying this behavior to rely metrics rather than admin distance?

Thanks,

Diego

[dslice]

Who is to say the comparison is within a single process?  This is a generic solution using interaction with the rib to resolve the installation of a prefix learned from multiple sources.  This could also be a case where an internal in one EIGRP AS is being compared to an external in another EIGRP AS.  The logic still applies.  EIGRP internal is more believable than EIGRP external, regardless of the AS.

I agree that we haven't always made decisions that are obvious (static to interface mentioned earlier in the thread as one example, which pre-dates my time at Cisco) but I don't believe the AD comparison is one of those.  Feel free to disagree.

[Peter Paluch]

Hello Donald,

Who is to say the comparison is within a single process?

I am not sure if I understand you correctly here. But this entire thread revolves around an issue that clearly indicates that the AD is used within a single EIGRP process.

Diego came here with a scenario where a single EIGRP process learns about the same prefix from two neighbors - one of them is advertising that prefix as internal, another is advertising the prefix as external. The EIGRP metrics for the external path are set to be orders of magnitude higher than the metrics of the internal route. Diego has also modified the AD for EIGRP so that the internal routes have AD set to 180 while external routes have AD set to 175.

My logic would dictate here that the EIGRP would still choose the better path  based on its metric (which is the internal path in this scenario), and  offer it to the RIB, obviously with the AD set to 180. This is how OSPF and BGP would work, too. However, Diego observed a different phenomenon: the best path inserted into the routing table was the external route which had substantially larger metric, yet its AD was lower. Hence, obviously, it was the AD that forced a single EIGRP process to prefer a route with a higher metric. In other words, the comparison is in a single process - using your own words.

I confirmed this behavior in my lab afterwards - Diego is correct, and so is Jon who also commented on this issue and confirmed Diego's original findings.

I find this EIGRP behavior inconsistent with the behavior of all other routing protocols that have a concept of internal/external route types, and that it my point throughout this thread.

Best regards,

Peter

[dslice]

My last comment on this.

The metric of an external route is unreliable due to the reasons I stated earlier.  Using this as the differentiator within EIGRP makes no sense when comparing to an internal route, where the metric value is meaningful.  The AD mechanism is a generic mechanism that works whether within a single AS/Process or between AS/Processes and the operation of the AD seems fairly simple and clear (with the exception of EIGRP PE-CE support, which has it's own strangeness.)  Agreed that different protocols have different requirements and operations in many areas, including how to deal with external routes.

[Peter Paluch]

Donald,

Thank you for responding. Please do not take my disagreements personally and please do continue debating things with us. There's no need to be categorical to the point of trying to end the discussion ("my last comment..."), especially considering the fact you just barely joined this thread and now it seems like you are trying to present a definitive resolution with no more debate about that. The CSC is a place where discussions are not and should not be ended abruptly like that. You may find more often than not that the most pleasant and fruitful discussions here on CSC are those where we - in a friendly way - disagree with each other to a great length, allowing each of us to present his reasoning.

The metric of an external route is unreliable due to the reasons I  stated earlier.  Using this as the differentiator within EIGRP makes no  sense when comparing to an internal route, where the metric value is  meaningful.

Such a statement is arguable. The metric, be it external or internal, in essence describes the distance to the originator of the routing information within the particular routing protocol. The external metric is just additionally increased by the seed metric. It is up to the network administrator to decide whether the seed metrics are set in a sensible way, and thus whether the metrics of external routes are trustworthy and in some way comparable to internal metrics. We can easily set up a scenario where it would be absolutely legal to compare external and internal routes and decide solely basing on their EIGRP metric.

After all, if it was true as you stated that "using the external metric as a differentiator within EIGRP when comparing to an internal route makes no sense", why would then EIGRP increase the external metrics using the internal metric components? Does it make sense, then, to mix the unreliable external metric information with the increments derived from the internal routing?

What I am pointing at is that it is too general and imprecise to say that external metrics are not comparable in a sensible way to internal metrics, and that the usage of AD in EIGRP bestpath algorithm is therefore justified. This always depends on the individual case and needs.

Once again, please do not consider my post here as asserting myself. I am merely pointing out different views to avoid seeing things in their traditional, often steretypical ways. I am thankful for any idea.

Best regards,

Peter

[tato386]

Thank you sir.  I wish the documentation would be more clear on that.  Seems like all the docs that I read concentrate on the metrics as being used within a routing protocol and the AD for choosing routes from different routing protocols.

Rgds,

Diego