I am exploring the VSS solution for our data centers. However, since we are
already esatblished the only driver for us is the ability to scale and
increase the amount of hosts we can conncet to a VSS cluster.
Our design takes:
Two data centers.
Two aggregation switches in each. The two pairs are connected with dark
fiber. So across the two data centers we have the same Vlans configured.
(Stretched VLANS) and MSTP instances.
In each Data center we have 1 Load balancer and 1 FW switch blades.
For the whole MST region we have Layer 2 connections between the
All our access switches (sup32) are thus redundantly connected to the local
Traffic outside of the MST VLANS is routed to the core - MPLS and MBGP.
This works very well. With a heavy reliance on STP - blocked paths
However as you see - we can sale to a maximum of 5040 ports/ hosts
7 blades (UTP/MM/SM 10/100/1000) * 48 ports * sup32 15 switches = 5040
7*48*15 = 5040
Every HoSt/Server in the DC LAN is redundantly connected upstream.
So the whole MST region - 2 datacenters- Layer 2 MST cloud has 10080 host
ports and half of them are in blocked state due to STP.
Therefore i cant scale beyond perhaps 15000 active/active ports in the
To be scalable and cost effective VSS seems to help me in this. I could
scale to 40 - 50 K hosts ports that are also active.
Can someone help me understand whether my thinking is correct.
Thanks in advance