cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
372
Views
0
Helpful
4
Replies

Redundant WAN link question design question.

Eric Hansen
Level 1
Level 1

Got a dumb quesiton, We have several WAN links with Century Link and the normal operating procedure is they drop in a layer 2 switch at the WAN site with one at the home office and they pass layer 2 ethernet to us.  We throw a vlan on the link with IP's on the virtual interfaces and off we go.  no routers, no /30 networks with IP'd interfaces on both sides.  Its worked great in the past.

We we brought in some new circuits and they dropped in two switches, both passing layer 2.  And they said they need us to fire up both because when we dont they call saying one is down.  So one port on each switch going to our layer 2 network, on the other side we plug them into a 6500 VSS.  My first thought was that STP should take over an block one link, assume somewhere in the cloud the two switches can see each other.  Not the case.  Both links fire up and forward and we lose all contact with the far side.

At this point I am tempted to blow the whole thing down and IP one interface on the 65k with one network and the other interface on the other 65k with a seperate network, then IP the WAN side with interfaces and sub interfaces running two networks over the link and let a routing protocol figure it out.  I am going to assume that MPLS is at work in the cloud.

I am curious which is the best way to go forward?  Right now one link on one 65k is shut, things work,  ISP is grumpy and we have no redundancy.

4 Replies 4

rais
Level 7
Level 7

Are both the links at 65K site suppose to connect to all three sites and vice versa? They are point to multipoint circuits?

What's the meaning of: "Both links fire up and forward and we lose all contact with the far side". Only one link can be used at a time?

Thanks.

One circuit as of now, and yes both links ont he 65k should see all sites.

"Both links fire up", so when I plug in both ports on the 65K they both come up.  I do a sh span on the switch and the ports both show FWD, when I expected one to be FWD and one BLK.  So because they both go to FWD I am effectively looping, and all hell breaks loose.  So I shutdown one port.

e-

Usually, ISP would drop your BPDUs, so you won't see BLK state.

Some ISP wouldn't allow to connect both CEs to the same device. If the working link goes down, the other should work.

HTH.

Well based on that and some physical limitations on my end I like the sub interface idea and may give that a go.

I've never studied any of the ccda/dp stuff, is there a best practice for this type of design?

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