Cisco switch replacement sizing based on CPU, PPS, memory utilization, etc.

Customer has a couple Catalyst 6509 core switches in a VSS core deployment, and they're looking at potential replacement options. So, while the natural replacement might be something in the Catalyst 6800 family (the 6880 would do the trick from a port density standpoint, also handling VSS), I’m wondering if something like a 4000-X VSS core could handle the load. They’ve indicated that they’ll need a maximum of (12) 10-Gigabit ports on each, but my concerns are really based more on the processing/memory load.

I have attached some trimmed down output from these units, and you can see that they’re very lightly-taxed from a memory and CPU standpoint.  The “show cpu history” has a spike within the last minute, but that was likely from running the “show tech” command to yield the output. I do see a few spikes (87, 82 and 91%) when looking at the 72-hour graph. These have a large number of 10-Gig interfaces, and they don’t appear to be processing at any higher than 200-Mbps peak input/output.  

Some additional detail on what these existing core switches are currently handling:

Routing:

• Full EIGRP routing (20 networks defined)
• 80 static routes
• 30 L3 SVIs

Handling 40 or so DHCP pools
10 ACLs (fairly small)
Voice QoS policy on all of the distribution links (6 closets with 10-Gig connectivity)

Beyond the port requirements, some other comparison items between the Sup720-10G and the 4500-X switches:

Sup720-10G (with PFC3):

• 600-MHz processor
• 450-Mpps
• 256K routing (IPv4)
• 1-GB DRAM for routing, 1-GB for switching

Catalyst 4500-X:

• Dual core, 1.5-GHz CPU
• 250-Mpps
• 256K routing
• 4-GB DRAM

What else should I be looking at here to determine if the 4500-X would NOT be a good replacement fit for this customer?

Thank you – reference docs are always appreciated.