cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
1778
Views
5
Helpful
10
Replies

WAN addressing

alliasneo1
Level 1
Level 1

Hi Everyone,

     I understand (from my limited Cisco knowledge)  that every WAN connection on routers have to be on the same subnet? Is this correct? from my picture below if Router 1-  serial 0/0 is 172.10.10.1/30 and Router 2- serial 0/1 is 172.10.10.1/30 are setup as you get more and more routers going on surely evertually you have to get on to another subnet. I was just curious as to how that happens if each serial interface is on the same subnet?

1 Accepted Solution

Accepted Solutions

Daryl

Exactly right. From a physical point of view you can go through many frame-relay switches, you don't care and don't need to know because that is the SPs job. But from a logical point of view you have a L3 connection from R1 to R4.

Jon

View solution in original post

10 Replies 10

Collin Clark
VIP Alumni
VIP Alumni

The address space between 2 routers needs to be different than any other address space. In your example above you're using a /30 address space which follows best practices. It provides 2 hosts per subnet. The link between your next set of routers would be the next subnet. The serial interfaces cannot have the same IP address. Here's an example of subnet usage:

So the serial interfaces between the routers would fall under the host range above. One router would have .1 and the other .2. The next pair of routers would use .5 and .6.

Jon Marshall
Hall of Fame
Hall of Fame

Daryl

No, there is no requirement for Wan connections to use addressing from the same subnet. Not sure if you have read something somewhere that would lead you to believe that ?

You could if you wanted use a separate subnet per WAN connection if you wanted.

Edit - you may have got this impression because a common practice is to use a /24 subnet and then split it up into multiple /30s for the point to point links. But it's just a practice, there is no requirement for it.

Jon

Hi,

Thanks for the quick responses. Ahh I see. Well so far in the labs I've done etc we have always configured the WAN connections for the same subnet. I think I just need to take my thinking a bit broader and think outside the box a bit more.

That's pretty clear what you have said Jon. so I could use for example, a 172 address on point-to-point between two routers and the next WAN interface off one of those could be a 180 address?

This really leads me on to my next question as well. When I had a lecture on WAN earlier this year he said that between two WAN routers you could have tens, if not hundreds of other routers and I was just wondering how you then configure those two routers to be point-to-point if they are then going through lots of other routers on different subnets?

I may be getting ahead of my studies here though.

Daryl

You wouldn't configure them as point to point because you couldn't. So if you had -

R1 -> R2 -> R3 -> R4

all connected via point to point links then each link would be it's own subnet and they could each be from a different network. But you couldn't have R1 and R4 on the same subnet. Each L3 hop, and a router represents a L3 hop, means you cannot use the same subnet on either side if you see what i mean.

However R1 could still send packets to R4 as long as you had routes on all the routers either by using a routing protocol such as EIGRP/OSPF or using static routes.

Jon

right I think I'm with you. so, taking your example. If you had R1 on one campus and R4 on another that was miles away.

Presumably you would connect R1 to your ISP connection and those packets would then be fired off to the ISP routers which would route the packets over to R4 on the other side of town?

How would the ISP learn of the routes you are using though? So you would set say OSPF on your side which would then communicate out to the ISP and update them accordingly? If this is the case though, how do you know they are using OSPF?

thanks

Daryl

It depends on what connectivity you are using from you service provide eg.

1) frame-relay. In this case the frame-relay network is only L2 so in actual fact you would be able to use a single point to point subnet between the routers R1 and R4 because there are no actual routers or more specifcally L3 hops in the SP network ie. the equipment used in the frame-relay network are L2 switches.

2) ATM - again you could create a point to point subnet between R1 and R4 because there are no routers in between

3) MPLS - L3 MPLS then yes you need to peer with your SPs routers. How do they know what routing protocol you are using. Well you need to agree this with them. Actually they will probably tell you what routing protocols they allow you to use.

So a general answer to your question is if the SP network provides L2 connectivity such as frame-relay and ATM then you can connect your sites with point to point subnets because there are no L3 hops in between your routers. If the SP network provides L3 connectivity then you peer with the SP routers and you use different point to points at each site.

Jon

Hi Jon,

You're good at this! One last question. When we are talking about point-to-point in frame relay and ATM this wouldn't be physical though, as the routers could be hundreds of miles apart? but logically they would be point-to-point because they would connect to this L2 switch which connects to this one and this one etc until they pop out the other side to your other router?

Thanks again for all your help, I really appreciate it.

Daryl

Exactly right. From a physical point of view you can go through many frame-relay switches, you don't care and don't need to know because that is the SPs job. But from a logical point of view you have a L3 connection from R1 to R4.

Jon

Thats great, thanks very much. Definitely cleared some things up!

until next time.....

Daryl

No problem, and thanks for the ratings. The CSC forums are a great place to clarify concepts so feel free to post whenever you need to. There are lots of experienced people here who will be able to help you out.

Just one last clarification in case you come across this and perhaps get confused after this discussion.

When i referred to ATM being a L2 switched network that is true of an end to end ATM infrastructure. However you could also used ATM connections to connect to an MPLS network. In this case although you are using ATM you are only going as far as the first SP router so in this case you would be peering with the SP and using a point to point subnet to the SP router.

Hope i haven't added confusion

Jon