So I'm trying to set up this switch for a lab to troubleshoot a performance issue with some POE devices and I think I'm running into some confusion in figuring out what is the most optimal way of configuring the switch. I've gone through pcunite's post and I'm still trying to wrap my head around it. As far as what I am doing in the lab:
We have some embedded devices (Em-1 thru Em-n) that require POE and they are really dumb in that they are using a hard-coded address space and dumping much traffic to a fixed address of a multi-network collector (C1, C2). The expectation is that the embedded device and the receiver interface are the only entities on the network and the address is hard coded. So what we would like to do is the following:
1st group is for the user accessible side of the server (10.1.1.x) and clients to access the data on the servers.
2nd & 3rd group need same address space (10.9.9.x), so has to be segregated w/ no leakage.
4th is a misc segment (192.168.1.x) with a management address tied to the 10/100 port only
In order for us to troubleshoot, we would like to "partition" the switch into 4 isolated groups of ports that are in different broadcast domains.
I think plani's post is pretty close to what I'm looking for, in fact I'm going to shamelessly copy his visuals to explain what the setup is looking to accomplish:
C1 will be connected to the GREEN and RED
C2 will be connected to GREEN and BLUE
Em1-Em# will be connected to RED
Em1'-Em#' will be connected to BLUE
My initial thought was like plani's: setup 4 bridges with the 4 different networks, all traffic is going to be self contained within the 13 ports (12x 1Gb & 1x sfp+) of each group so I don't think I need to worry about tagging/untagging (I think). However, as I'm going through the posts, it seems that this configuration is not the best option. Can someone suggest a better way to do this while maintaining wire speed ?