05-12-2020 10:38 AM
hi guys
In my toplogy, i want to connect my 5 thirdparty switches in a ring Topology with single uplink to to my Core switch Pair (VSS)
I want to use osfp protocol between my thirdparty switches ring and Core. i dont want to use layer 2 because of spanning-tree loops. i want to keep one primary route from a switches ring to my Core and keep secondary route operational incase of primary route is down. in this design OSPF can prioritize the route base on Cost. i want to ask is there any chance of Loop in the network by using OSPF or we can have anothero protocol to work better in this design.
Topology is attached for reference. Thanks in advance.
05-12-2020 01:53 PM
05-12-2020 02:12 PM
Hi joseph,
Thanks for your reply,
I aggree with you
Actully this is a typicall Layer 2 ring topology between two different vendors, e.g. Cisco and Extreme.
They have different Ring protection protocol support which is incompatible for both. it is not good to run a spanning tree in ring topology VSS/stackwise virtual environment i suppose.
so we have layer 3 capable switches and choice is to run routing protocol to establish loop free the comuunication
Active switch will Synchronize all the configuration with standby switch So Switch E uses the backup routes towards Core B.
05-12-2020 03:37 PM
05-12-2020 04:02 PM
Actually, I don't recall any issues using STP with VSS, however Cisco's PVSTP does often have issues with other vendors that support a single (for all VLANs) STP. When that's the issue, sometimes it can be addressed, at L2, using MSTP.
====================
Thanks for reply
So in this case the communication will be redirect from Core-B to Switch E in case of link single link failure between Switch A and Core A. CORE B and Switch E will communicate on the same vlans configured between Core A and Switch A or there will additional configuration on Switch E and Core B?
Thanks in advance
05-13-2020 07:51 AM - edited 05-13-2020 07:53 AM
For either a L2 solution (e.g. using MSTP) or a L3 solution (using a dynamic routeing protocol) all the devices will need the proper configuration. For either, switch A <> core A Giuseppebe made the primary path which will fail over to switch E <> core B, which when switch A <> core A is restore, will again become the primary path.
PS:
BTW, it's also possible to mix a L3 solution with L2. For example, only switches A and E and cores A and B would need to be L3, for the L3 solution. The other switches could remain L2.
05-12-2020 03:20 PM
The original poster has an interesting question. We don't know much about the thirdparty switches other than that they have their own layer 2 loop prevention mechanism and that they will run OSPF among themselves in area 1 and 2 of them will have layer 3 connections to Cisco switches running OSPF. The concern is the possibility of loops. My response is that the possibility of a layer 2 loop is confined to the thirdparty switches and their loop prevention should be effective. And that OSPF has mechanisms to prevent layer 3 loops.
As @Joseph W. Doherty has pointed out there may be some impact as OSPF converges to react to failure of the primary link. But loops will not be a problem.
Discover and save your favorite ideas. Come back to expert answers, step-by-step guides, recent topics, and more.
New here? Get started with these tips. How to use Community New member guide