cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
940
Views
20
Helpful
8
Replies

EIGRP doubts

Manu Shankar
Level 1
Level 1

I have few doubts here..

What is the difference between wild card mask and subnet mask in Network command when configuring eigrp ?

How this command " ip default-network" works ?

8 Replies 8

Peter Paluch
Cisco Employee
Cisco Employee

Hi Manu,

What is the difference between wild card mask and subnet mask in Network command when configuring eigrp ?

There is no difference between using the wildcard mask and the subnet mask in the network command in the EIGRP configuration. You may use any of them. However, the router will always convert the subnet mask into the wildcard mask format. (Why is the subnet mask used in the wildcard format, that remains a mystery to me and I find that utterly confusing and illogical.)

So, you may use any format you like the most. The router will always convert it into the wildcard format.

How this command " ip default-network" works ?

Try reading the following document first, it is a very nice description. If any doubts remain then please come back and let us know!

http://www.cisco.com/en/US/tech/tk365/technologies_tech_note09186a0080094374.shtml

Best regards,

Peter

Thank you, i will go through on this. But the first one is still mystery ..:-)

I  went through one pdf, to check the wild card mask vs subnet mask. but  there also its not differentiating the concept. this pdf saying the use  of wild card mask. if we are using subnet mask in that place, we will  get the same result. I have attached the same here.

But still I am confused, both are doing the same function ?

manu_nambu wrote:

Thank you, i will go through on this. But the first one is still mystery ..:-)

I  went through one pdf, to check the wild card mask vs subnet mask. but  there also its not differentiating the concept. this pdf saying the use  of wild card mask. if we are using subnet mask in that place, we will  get the same result. I have attached the same here.

But still I am confused, both are doing the same function ?

I suspect there may be a couple of things going on here -

1) the subnet mask is often easier for people to understand and so it is an ease of use function that Cisco have put into the IOS. I also suspect there are many lines of EIGRP code that use the wildcard mask and expect to find a wildcard mask. So rather than have to translate the subnet mask to a wildcard mask every time the EIGRP code needs to deal with the "network" entries, a one off conversion is done when it is originally entered into the configuration.

This would then mean that all the other parts of the EIGRP code could simply use the mask involved without having to worry whether it was subnet or wildcard mask format.

2) Generally speaking a subnet mask must be contiguous whereas a wildcard mask does not have to be so you can be far more flexible with wildcard masks in specifying networks. However with EIGRP and OSPF the wildcard mask must also be contiguous when using it under the "router config. However i believe, although i could be wrong, that with OSPF at least, earlier versions of IOS did not require the wildcard mask to be contiguous which gave you far greater flexibility in specifying the networks. This may also have applied to EIGRP.

Both of the above are best guesses.

Jon

Hi Jon,

So rather than have to translate the subnet mask to a wildcard mask
every time the EIGRP code needs to deal with the "network" entries, a
one off conversion is done when it is originally entered into the
configuration.

I would say that this is a matter of implementation. The entire network command in EIGRP is not related to directly advertised networks, it is merely used to define a set of interfaces whose IP networks shall be advertised into EIGRP. Thus, evaluation of this command can be considered a singular (i.e. initiatory) operation whose complexity is not so important (and in fact, it cannot be complex at all considering its straightforward meaning). If you evalute the command argument once and store them in an internal data structure already formatted in your most appropriate format and use it for all other operations inside the EIGRP code then it is irrelevant what format does the network command use. Yes, this is the common programming dilemma - trading memory efficiency for speed.

And this is also my doubt about the true reason why the network command accepts wildcard-formatted mask. As far as I know, it always has to be a wildcard mask that corresponds to a valid subnet mask, i.e. they must be equivalent which effectively "downgrades" the meaning of the wildcard mask here to a mere differently-looking-subnet-mask-notation and only causes confusion.

However i believe, although i could be wrong, that with OSPF at least, earlier versions of IOS did not require the wildcard mask to be contiguous which gave you far greater flexibility in specifying the networks

This would be the only reasonable explanation for me. But then again... the most logical way of activating a routing protocol is the IS-IS way for me: having the routing protocol activated on a particular interface - because that is what you essentially always want! The entire dancing around with network commands is introducing a layer of indirection that does not, in my opinion, have any significant advantages. I've seen the horror in the eyes of my students so many times when I told them that "the network and mask you are writing in the routing protocol configuration is not necessarily the network/mask which really gets advertised".

I am not bashing I'm just discussing and I like it very much!

Best regards,

Peter

Hi Peter

I am not bashing

I know your not. It's good to discuss these things as you say. I agree with the fact that if the wildcard mask has been "downgraded" to use your words then there seems little point in it's use. Perhaps a more interesting point would be why it was downgraded in the first place. Admittedly it's not often that you would actually need to use a true wildcard mask but i could see a situation where you wanted to start the routing protocol on a certain number of interfaces and the only way to use one "network" statement would be with a wildcard mask.

Still, i suppose it's not a huge problem to have more than one "network" statement. And having the requirement for the wildcard mask to be contiguous does indeed simply add to the confusion especially as i find most people have a much better understanding of subnet masks than wildcards.

Point taken on the programming aspect that i was talking about.

Jon

Hello Jon,

Thank you very much for replying.

I agree - if the wildcard mask used in the network command worked as true wildcard mask then I would understand its presence there. However, the router, at least not with recent IOS, won't accept a wildcard mask resulting in discontiguous address ranges so it seems to me to be more of an annoyance than anything useful.

The wildcard mask came actually with some of the first classless protocols implemented in IOS. It surely wasn't RIP or IGRP - it must have been either OSPF or EIGRP. Would be interesting to see if there is an old IOS still available for download that accepts discontiguous wildcard masks.

The IS-IS (or the IPv6) way of configuring routing protocols directly on interfaces is still the best for me

Best regards,

Peter

Peter

The IS-IS (or the IPv6) way of configuring routing protocols directly on interfaces is still the best for me

I shall happily take your word on that as i have absolutely no experience with IS-IS whatsoever

Jon

Jon,

No experiences with IS-IS whatsoever? Why do I find myself not believing that?

In any case, should you really have been saved from the IS-IS so far, I am absolutely sure you would find it a marvellous protocol. Most people are appalled at first because the IS-IS provokes an "otherworldly" impression (as pretty much anything that comes into the IP world from ISO or ITU-T). Once a person gets over the different perspective in OSI concepts, terminology and addressing which is necessary to run the IS-IS, the clean-and-lean design of the IS-IS starts revealing itself. I am very indecisive myself which is my favourite - OSPF or IS-IS

Best regards,

Peter

Getting Started

Find answers to your questions by entering keywords or phrases in the Search bar above. New here? Use these resources to familiarize yourself with the community:

Review Cisco Networking products for a $25 gift card