Trunking VLANS/OSPF

[ant0879]

Why would you trunk VLANS in between core nodes?  Doesn't OSPF handle routing for IPs behind the vlans from one layer 3 to another?

[Sergey Lisitsin]

ant0879,

That would depend on a particular case. If you have Layer 3 in your core layer, then you won't need to trunk VLANs. But some designs define Layer 2 in the core layer, so they would need VLANs trunking. There is no one size fits all kind of solution. Some networks run OSPF, some run EIGRP. Some just have static routes. All depends on a particular design choice.

[ant0879]

That would also apply to layer 3 switches correct.  Because layer 3 switches has the capability to OSPF.  

[Georg Pauwen]

Hello,

in addition to the other post, one reason could simply be speed. Trunks mean layer 2 connectivity, which in general is a whole lot faster than layer 3 connectivity (OSPF/EIGRP/RIP)...

[Joseph W. Doherty]

"one reason could simply be speed. Trunks mean layer 2 connectivity, which in general is a whole lot faster than layer 3 connectivity (OSPF/EIGRP/RIP)..."

On a L3 switch, there's often very little, if any, PPS difference in L2 frame forwarding vs. L3 packet forwarding. The one possibly notable performance difference, between L2 and L3 PPS, would be for the first packet if the L3 is using a flow based hardware architecture (which the later hardware supported CEF based L3 switches are not). (Also BTW, one of the most important reasons for tagged packet switching [now known as MPLS] was to increase L3 forwarding performance on early hardware, but again, on modern L3 switches, MPLS generally doesn't offer better performance [but MPLS is still used for other reasons.)

[paul driver]

Hello

---

@ant0879 wrote:

Why would you trunk VLANS in between core nodes?  Doesn't OSPF handle routing for IPs behind the vlans from one layer 3 to another?

---

I am assuming you mean a collapsed core/distribution design in relation the L2 extended vlans across that core/distribution - then reason why would be to extend the L2 capability and the only way to do that would is by  l2 trunks.

The way to negate this would mean having a L3 access design, However that would mean you wouldn't be able to have multiple vlans on the access switches, they will need to be single vlan specific will dual l3 interfaces into the core/distribution advertised within the IGP you have chosen.