EIGRP - Configuring a Router

The first specifies the entire classfull range of 10.0.0.0

So does the second, but with a Wildcard mask.

You could have used a different wildcard mask to cover a smaller piece of the entire classfull portion.

I have noticed in a lab environment that EIGRP doesn't like wildcard masks very much. Had some problems creating adjancenies with the wildcard masks.

Well, EIGRP is a classless protocol, so I don't have to put classfull networks.
So, the first command is in case I want to put a classfull network and the second when I want to put a classless one?

sorry I didn't mean you had to use classfull networks. I was just explaining that
network 10.0.0.0 would cover all interfaces that were in the range of 10.0.0.0 - 10.255.255.255 .
Network 10.0.0.0 0.255.255.255 does the same thing.

It doesn't matter too much whether you are using classfull or classless addressing. This only tells which interfaces to start sending out EIGRP updates.

keenon

gabrielbtoledo wrote:

Well, EIGRP is a classless protocol, so I don't have to put classfull networks.
So, the first command is in case I want to put a classfull network and the second when I want to put a classless one?

well eigrp will only act classless with the "no auto-summary" command under the routing process, you can add the wild card masks if you only want to run it specifically from a single interface and not all the interfaces on that router.

your first command tells the router which interfaces to include in the routing process and to share from.

normally you would use the "no auto-summary" first before adding networks under the process

Yes, I know about the "no auto-summary" command, but I am still a little confused.
Could somebody write down how would you configure a router using EIGRP? Just the part where you enter in global configuration mode. I want to visualize what commands you use and which order. Thanks.

EDIT:
Would it be:
router eigrp 100
no auto-summary
eigrp log-neighbor-changes
network 172.16.0.0

dtlokee

The problem you may encounter when configuring EIGRP is you may have 2 interfaces that fall in the same Classful network, but you only want to configure one of them to run EIGRP.

10.1.0.0/24

10.2.0.0/24

Let's say we only wanted to run EIGRP on the 10.1.0.0/24 subnet and not the 10.2.0.0/24 subnet. The "network 10.0.0.0" command won't do this for us, we need to use a wildcard mask to accomplish the goal. "network 10.1.0.0 0.0.0.255" will get the job done, basically it says to match the first 2 octets of the provided network address and ignore the last 2 (the ones with the 255 in them).

Also keep in mind the network statement does not control the subnet address that is advertised by the router, it only tells the router what interfaces to run the routing protocol on. In the previous example I could have used the "network 10.1.0.0 0.0.255.255" command to enable EIGRP on the interface connected to the 10.1.0.0/24 subnet but it would still advertise the subnet as 10.1.0.0/24 regardless of the network statement provided. (also we would need the no auto-summary command if we were crossing classful networks in out IP addressing space)

enable
conf t
router eigrp 10
no auto-sum
network 10.0.0.0

If you had IP's 10.1.1.1/24 10.1.2.1/24 10.1.3.1/24 and you only wanted EIGRP to advertise on one of those interfaces and not all of them.

enable
conf t
router eigrp 10
no auto-sum
Network 10.1.2.0 0.0.0.255

keenon wrote:

your first command tells the router which interfaces to include in the routing process and to share from.

So, EIGRP is not like RIP when you put the network you want to advertise instead of putting the interface that will advertise in IEGRP?

gabrielbtoledo wrote:

keenon wrote:

your first command tells the router which interfaces to include in the routing process and to share from.

So, EIGRP is not like RIP when you put the network you want to advertise instead of putting the interface that will advertise in IEGRP?

Remember RIP is classful...well, at least v1 is, so when the command is applied it will encompass all the subnets in the classful range...with EIGRP there's more flexibility...

r_durant: Thanks, but my question now is...

RIP - input networks that you want to be advertised.
EIGRP - input interfaces that will advertise.

That is my question.

NeonNoodle

gabrielbtoledo wrote:

r_durant: Thanks, but my question now is...

RIP - input networks that you want to be advertised.
EIGRP - input interfaces that will advertise.

That is my question.

Let me take a stab (in the dark) at this. Let's say you have two interfaces configured with the following IP addresses: 10.1.2.1/24 and 10.1.1.1/24. Suppose you only want to advertise one of the routes, say, 10.1.1.0/24, and you don't want to advertise 10.1.2.0/24. If you are using RIPv1 to route, well you have no choice but to advertise both because RIP is classful. (Type in 'network 10.1.1.0' and then do a 'show running-config' and see what's there.) If you are routing using EIGRP, then you can advertise the 10.1.1.0/24 network only--by using the 'network 10.1.1.0 0.0.0.255' statement since the wildcard mask indicates that 10.1.2.1/24 won't be a match.

To answer your question more directly: you don't advertise interfaces, you advertise routes. CIDR allows you to be more specific in which routes you want to advertise. That's why RIPv1 and IGRP are 'going the way of the dodo'.

Well, this is how I look at it...and it seems similar to how some of the others explained

I personally, wouldn't say that with EIGRP you input interfaces that would advertise, but if you have two or more interfaces (physical or otherwise) that would fall in the same classful network, but are subnetted and you only wanted to advertise one subnet, I would use Netstudent's example....

If you had IP's 10.1.1.1/24 10.1.2.1/24 10.1.3.1/24 and you only wanted EIGRP to advertise on one of those interfaces and not all of them.

enable
conf t
router eigrp 10
no auto-sum
Network 10.1.2.0 0.0.0.255

And this would only advertise the 10.1.2.0 subnet and not the other 2.

I guess it maybe a matter of interpretation...

What i'm saying is that i wouldn't refer to it by the interface, because you can also have sub-interfaces on one physical interface...so i look at it from the subnet point of view

Your interpretation may not be wrong, and i'm not saying it is, but i'd say be careful with how u refer to them...

I'm not sure if you understand where i'm coming from....

dtlokee

The behavior of the network command for all IGPs (RIP, EIGRP, OSPF, et al) is to specify what interfaces we want to send and receive updates on. It has no bearing on what address and mask is advertised regardless of the protocol.

Interface E0
 ip address 10.1.1.1 255.255.255.0
!
Interface E1
 ip address 10.2.1.1 255.255.255.0

Assuming that is the configuration for the router interfaces, there are a number of combinations that we can use to tell the router to send updates through the interface.

router eigrp 10
 network 10.0.0.0

--- or ---

router eigrp 10
 network 0.0.0.0 255.255.255.255

--- or ----

router eigrp 10
 network 10.1.1.0 0.0.0.255
 network 10.2.1.0 0.0.0.255

---- or (my favorite) ----

rotuer eigrp 10
 network 10.1.1.1 0.0.0.0
 network 10.2.1.1 0.0.0.0

So which one is correct? they all are. They all do the same thing, tell the router to send updates through both int E0 and E1, and to advertise the subnets 10.1.1.0/24 and 10.2.1.0/24. Keep in mind the network command does not tellthe router what to advertise, just which interfaces to advertise.

Also there's some confusion about auto-summary here. Since all router interfaces are int the major network 10.0.0.0 it is not needed in this example. It's generally a good idea to disable it anyhow, but you must understand why. If the router had interfaces connected to more than one major network it would be needed, but in this case it's not.

dtlokee wrote:

The behavior of the network command for all IGPs (RIP, EIGRP, OSPF, et al) is to specify what interfaces we want to send and receive updates on. It has no bearing on what address and mask is advertised regardless of the protocol.

Thanks, that was the answer I was looking for. So, I wasn't too crazy about my theory.
When I said to input the interfaces that will advertise, I meant to say... input the interfaces (networks=ip addresses configured to the interface) that will send and receive advertisement (updates). I just didn't know how to express myself.

dtlokee wrote:
Interface E0
 ip address 10.1.1.1 255.255.255.0
!
Interface E1
 ip address 10.2.1.1 255.255.255.0
Assuming that is the configuration for the router interfaces, there are a number of combinations that we can use to tell the router to send updates through the interface.
router eigrp 10
 network 10.0.0.0

--- or ---

router eigrp 10
 network 0.0.0.0 255.255.255.255

--- or ----

router eigrp 10
 network 10.1.1.0 0.0.0.255
 network 10.2.1.0 0.0.0.255

---- or (my favorite) ----

rotuer eigrp 10
 network 10.1.1.1 0.0.0.0
 network 10.2.1.1 0.0.0.0
So which one is correct? they all are. They all do the same thing, tell the router to send updates through both int E0 and E1, and to advertise the subnets 10.1.1.0/24 and 10.2.1.0/24. Keep in mind the network command does not tellthe router what to advertise, just which interfaces to advertise.

Also there's some confusion about auto-summary here. Since all router interfaces are int the major network 10.0.0.0 it is not needed in this example. It's generally a good idea to disable it anyhow, but you must understand why. If the router had interfaces connected to more than one major network it would be needed, but in this case it's not.

So, when you input the network address (we talking about EIGRP) to be advertised and you don't specify the mask or wild card mask, we are meaning the network as a whole (classful), however if we want to advertise a subnetwork, then we would input the mask or wild card mask, right?
Also, in your last example:

network 10.1.1.1 0.0.0.0
network 10.2.1.1 0.0.0.0

We use this way when we want an exactly match, right?

dtlokee

gabrielbtoledo wrote:

So, when you input the network address (we talking about EIGRP) to be advertised and you don't specify the mask or wild card mask, we are meaning the network as a whole (classful), however if we want to advertise a subnetwork, then we would input the mask or wild card mask, right?

When you leave out the wildcard mask it will default to using the major classful network to identify what interfaces will send updates. This may match more interfaces than you require, so you have 2 options. Option 1 is to use a wildcard mask to limit the match to only the required interfaces, or disable EIGRP packets on the ones you don't need EIGRP using the passive-interface command. Again, the network command has no bearing on what is advertised, major network, subnetwork... what is advertised comes from the interface IP and subnet mask. The network command only tells the router what interfaces to include in EIGRP

gabrielbtoledo wrote:

Also, in your last example:

network 10.1.1.1 0.0.0.0
network 10.2.1.1 0.0.0.0

We use this way when we want an exactly match, right?

[/quote]

Yes this will match the only the interfaces with the IP address 10.1.1.1 and 10.2.1.1, all other interfaces will not be matched.

dtlokee...so my question now is...

Am I wrong in the way i look at it? I understand fully what you explained, but i just look at it from the subnet and not the interface.

The way i see it, an interface will have an address, so that goes without saying...it's tied to the interface already...so if i want to advertise subnet x, it's already on that specific interface??

Maybe it's just how i interpret it or try to explain it, but i understand how and why it's done...

rjbarlow

For my learning in order to "cut" one interface from advertise the routes in EIGRP is possible to issue the "router(config-router)#passive-interface <interface>", that in EIGRP means that the specified interface not send and not learn routes.
But if is possible to use wildcard mask with the "network" command in EIGRP I think it follow the rules of the wildcardmask too.

r_durant wrote:

dtlokee...so my question now is...

Am I wrong in the way i look at it? I understand fully what you explained, but i just look at it from the subnet and not the interface.

The way i see it, an interface will have an address, so that goes without saying...it's tied to the interface already...so if i want to advertise subnet x, it's already on that specific interface??

Maybe it's just how i interpret it or try to explain it, but i understand how and why it's done...

I just kept hitting the same key because my book says interface and not network. That is why I want to understand the official concept.
My mistake what the concept of advertising was different from RIP to EIGRP.
In RIP I thought you would input the networks you want to advertise and with EIGRP you would input the interfaces you want to send and receive the advertisements (not sure if is the right terminology).
Anyway, now I understand the concept and it doesn't matter the protocol, you will always input the interfaces you want to advertise.
My other mistake was not know that you can configure EIGRP with the network command without mask or wild card mask.

Thanks a bunch.

rjbarlow wrote:

For my learning in order to "cut" one interface from advertise the routes in EIGRP is possible to issue the "router(config-router)#passive-interface <interface>", that in EIGRP means that the specified interface not send and not learn routes.
But if is possible to use wildcard mask with the "network" command in EIGRP I think it follow the rules of the wildcardmask too.

Isn't there something about the passive-interface still being learnt by the other routers, even though the interface itself does not advertise or learn routes??

IN RIP a passive interface revieves routes but does not advertise them. In EIGRP, a passive interface niether sends or recieves updates. An Adjacency cannot be formed in EIGRP with passive interfaces.