02-09-2011 01:06 PM - edited 03-06-2019 03:27 PM
What is the maximum IE-3000-4TC switches that can be used in a REP segment?
Thanks,
David
Solved! Go to Solution.
02-09-2011 08:57 PM
Hello David,
Our development testing has tested 32 nodes in our scalability/performance testing. That being said we have several customers who have upwards of 100 devices in a ring without issue. There's no hard/fast rule on exactly what is fully supported versus not supported. More switches in the ring is going to equal slower convergence time and response to a failure. Even with 100 devices in a ring though convergence time is still significantly improved over RSTP.
-Matt Blanshard
Cisco TAC - Lan Switching
02-09-2011 08:57 PM
Hello David,
Our development testing has tested 32 nodes in our scalability/performance testing. That being said we have several customers who have upwards of 100 devices in a ring without issue. There's no hard/fast rule on exactly what is fully supported versus not supported. More switches in the ring is going to equal slower convergence time and response to a failure. Even with 100 devices in a ring though convergence time is still significantly improved over RSTP.
-Matt Blanshard
Cisco TAC - Lan Switching
02-10-2011 12:45 PM
Matt,
Thanks for the quick reply. My engineer passed along this information:
___________________________________________
"Let’s see if he can confirm that this design should work fine and not cause any loops in the system. I think it will control loops without issues if my understanding of REP is correct. But we have no control of how the fiber is being laid out so I’m making sure how I configure the network will work with this physical layout. This is a small sample design but the idea is the same no matter how many IE3000’s there are in each segment. Just will take a little longer (ms) to converge based on how many switches.
If this won’t work by using segment 2 then I’ll just have to run RSTP on that part of the ring and deal with the results of how many switches there are and hope RSTP doesn’t get angry.
___________________________________________
I've attached his example of the network. Could you take a look at it and let us know what your thoughts are?
Thanks,
David
02-11-2011 10:32 AM
Hello David,
I have looked through your design and it looks good to me. By using the two segments like you are this guarantees no loops are possible. Your list of concerns with how offices link to each other are exactly the same ones I would have (last bullet point).
-Matt
02-11-2011 11:06 AM
Matt,
Again, thank you very much for your assistance. We'll set it up in the lab and give it a try.
Yours,
David
06-02-2015 05:33 AM
Matthew/Cisco Experts,
Is this answer of 32 still valid? I've read a Technical Presentation recently which stated that 16 Nodes were advised, but that 32 had been tested.
Is there a revised "Advisory Level" (referring to the 16 Nodes), given that time has elapsed and REP has been developed since your original answer was posted?
Are you effectively saying there is no hard limit, other than the REP Segment restoration time being acceptable to us - so assuming higher REP Segment size can mean a higher REP Heal/Restoration time than the 50 - 250ms advisory figures?
Thanks in advance.
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