[c-nsp] Layer 2 vs Layer 3 Performance Question

Jason Luke jluke at truarx.com
Mon Jun 22 10:35:33 EDT 2009


I'm not sure if it matters but if it does, I would like to know.

I have two buildings, connected via fiber (1 Gig speed only). In bldg one, I have my core switch (6509) and that core switch has the default route for everything in bldg 2. In other words, bldg 2 does not have a router, just a switch.
In bldg one, I have my SAN, storing all my virtual machines, all talking iSCSI  to each other on VLAN 5.
In bldg two, I have my backup server and my data duplication (DD) storage device on VLAN 10.

I could just as easily trunk VLAN 5 over to the switch at bldg two and put the backup server and DD box on the iSCSI VLAN 5. This would mean that ALL backup comms only happen on that same VLAN, no routing/traversal of that VLAN.
Or I could leave it as is and route the traffic. The actual physical paths are no different.

So the question is, is there a difference?  I have hundreds of GB's to backup so if there is a performance difference one way or the other, I would want to know.  The only thing I can see as a difference is if I trunk the iSCSI vlan, then the traffic never hits the 6509's routing module. Maybe that helps performance?



More information about the cisco-nsp mailing list