Why to use Ospf Point to Point for Core to Distro Switches

[syedfahadahmed]

Dear All,

I have recently joined in a new company. Here they have configured "OSPF Point-to-point" between 6500 Core switches and 4500 Distribution Switches for 3 buildings on Fiber links. May I know why would we do that  and why not as transit ?

[Mark Malone]

Hi

One of the reasons is there are only 2 links so there's no requirement for OSPF to go into a calculation and try chose a DR and BDR if the network was broadcast , setting as p-p prevents this

OSPF as well uses the /32 for the loopback in advertisements , if your creating your neighbour relationship through loopbacks you would use the point-point statement otherwise it will set them as /32 instead of say your mgmt. range which could be /24 or /23 etc  

[syedfahadahmed]

Hello Mark,

Thank you for the response. 

Here we don't have point to point configured on loopback. If it is configured on Loopback I can understand. But here they configured "Point to Point Network" from Core Switches to DownStream Distribution switches fiber interfaces. May I know is it a good design practice ? 

[Mark Malone · Accepted solution]

If its not 2 loopbacks but 2 physical interfaces its still good practice to use it as there's no requirement for OSPF to use anything else but p-p as that's what it is 2 directly connected point-point links so yes I would still use it even if not loopbacks , loopbacks is just another known reason people use it as its can change the subnet mask but if you have 2 direct physical links as neighbours it can be used , its an optional command it wont make or break your design but it is helpful fro the reasons mentioned and probably more

[syedfahadahmed]

Thank you for the detail explanation Mark.

It was indeed very helpful.