Multiple WAN Interfaces for one LAN

[DV_Philip]

Hi to all Cisco Community members,

I have a question about the use of having multiple WAN interfaces on a routing device. (Firewall with routing capabilities, traditional router, Layer 3 switches[?])
Admittedly, I don't know what I'm looking for in specific, but I was hoping that someone could give me a few examples.

Q1: I'm sure I can find a reason why you would want to have multiple WAN interfaces in a network that has two physically separate LANs with the router being an "edge router". However, when it comes to having it being for 1 single LAN, I start to get a bit confused. Only one would be used at a time, right? Is this for redundancy/bandwidth inflation?

Q2: Is there a reason why you would have two WAN interfaces for a network that is logically segmented via subnetting? Logically speaking, NAT would be able to do it's work regardless of what subnet the original source is on, but I feel like there's something conflicting with the ideology of having a segmented network if we're using the same WAN port for both of them. 

Q3: Also, can someone give me an example of how you can achieve a funnel into 1 WAN interface if you have multiple VLANs? The switch that has the uplink to the router has multiple VLANs configured on it, but how would you need to configure your uplink port to have all VLANS talk to it? Can you set the VLANs of that port to be all of the VLANs?

Bear with me, my experience in networking is little to none. I've been working as a entry-level network analyst for two months now with no former experience and this is also my first post here, so if I shouldn't have posted here/if I broke any community rules, by all means, let me know.

Thank you very much for your patience, I seem to have a bit of missing knowledge on WAN interfaces. Not sure what they do aside from connect a LAN device to the internet, and I still fail to understand how WAN is different from a large LAN and why we need a modem for it when the rest of our network transfers data without one. Maybe that can be a bonus question for anyone willing. :)

-Phil from Datavalet

[Georg Pauwen]

Hello,

Q1: Two (or more) WAN interfaces, even for a single LAN, are indeed typically used for redundancy and/or load balancing. You can (and should actually, since you are paying for it) use both connections simultaneously. How the redundancy is set up depends on whether you use the same or different ISPs, which routing protocol you use, and what the bandwidth is of both links.

Q2: Local subnetting is done to have separate logical networks, which usually correspond to local departments, or specific user groups. Let's say you have three local subnets - the redundant WAN connections would still make sense, since all three subnets need outside connectivity, and, as you already stated, they would typically be NATted to a single address anyway, so from the outside, it wouldn't really matter what your internal structure looks like.

Q3: In order for all Vlans to be able to talk to each other you typically create Vlan interfaces (SVIs, e.g. interface Vlan 10 to which you can assign an IP address, and which acts the same as a physical Layer 3 interface). The WAN interface and the SVIs are configured on the same device, so they can automatically talk to each other, since they all are directly connected interfaces. 

There also is the concept of 'router-on-a-stick', which basically achieves the same thing, which is to create one layer 3 interface per Vlan.

HTH

[paul driver]

Hello Phil
Welcome to CSC!

---

@DV_Philip wrote:

Hi to all Cisco Community members,

I have a question about the use of having multiple WAN interfaces on a routing device. (Firewall with routing capabilities, traditional router, Layer 3 switches[?])
Admittedly, I don't know what I'm looking for in specific, but I was hoping that someone could give me a few examples.

Q1: I'm sure I can find a reason why you would want to have multiple WAN interfaces in a network that has two physically separate LANs with the router being an "edge router". However, when it comes to having it being for 1 single LAN, I start to get a bit confused. Only one would be used at a time, right? Is this for redundancy/bandwidth inflation?

Yes and NO you can if you desire just to use one and have the other as a backup however as you paying for this service it would be a waste of money just to have maybe an expense second link idle, So you can if you desire utilise both- In fact it’s preferable to do so and cisco has certain features that can assist in doing so and static or dynamically choose the best wan path based on certain criteria like load/bandwidth application flow source/destination address etc..

Q2: Is there a reason why you would have two WAN interfaces for a network that is logically segmented via subnetting? Logically speaking, NAT would be able to do it's work regardless of what subnet the original source is on, but I feel like there's something conflicting with the ideology of having a segmented network if we're using the same WAN port for both of them.

Nat is mainly used to allow private non routable subnets (192.168.0.0/16,172.16.0.0/12 10.0.0.0/8) to be translated so this users on these subnets are able to communicate to routable internet addressing like on your WAN links.
It can also be used to translate between each other within a LAN, MAN environment but eventually there will be an exit point onto the WAN that needs to be translated.
 

Q3: Also, can someone give me an example of how you can achieve a funnel into 1 WAN interface if you have multiple VLANs? The switch that has the uplink to the router has multiple VLANs configured on it, but how would you need to configure your uplink port to have all VLANS talk to it? Can you set the VLANs of that port to be all of the VLANs?

Usually you would have a L3 device (routing) like a Router or Intelligent switch that would be able to route all you internal vlans between each other ( inter-vlan routing) with each vlan having its own L3 interface.

If a router was designated to perform you inter-vlan routing (router-on stick scenario as stated by georg) then there would be usually an LAN interface, logically split into sub-interfaces for each internal vlan and a then its physical interface would have trunk connection towards an internal switch for your users to connect through.

It would also have an additional interface (WAN ) that connects to your ISP or firewall and this is probably where Network  Address Translation would occur between the LAN/WAN interfaces

For a L3 switch, this would have probably multiple Switch Virtual interfaces (SVI) designated for each vlan that would allow any users of these vlans to be able to communicate over these interfaces.

The connection between the edges switches and your L3 switch would usually be trunks links that allow all the necessary vlans to traverse these trunks so the edge ports on your edge switches can be allocated their own necessary vlan port.

Lastly this L3 switch would usually have an access port assigned to maybe a SVI management L3 vlan connected to the upstream router which would forward any designated wan traffic to your wan rtrs for offsite connectivity.

Bear with me, my experience in networking is little to none. I've been working as a entry-level network analyst for two months now with no former experience and this is also my first post here, so if I shouldn't have posted here/if I broke any community rules, by all means, let me know.

Thank you very much for your patience, I seem to have a bit of missing knowledge on WAN interfaces. Not sure what they do aside from connect a LAN device to the internet, and I still fail to understand how WAN is different from a large LAN and why we need a modem for it when the rest of our network transfers data without one. Maybe that can be a bonus question for anyone willing. :)

-Phil from Datavalet

---

 Lastly if you have any other queries please dont hesitate to ask them , I am sure they would be gratefully reviewed by myself -Georg and other CSC forum members.