Disclaimer
The Author of this posting offers the information contained within this posting without consideration and with the reader's understanding that there's no implied or expressed suitability or fitness for any purpose. Information provided is for informational purposes only and should not be construed as rendering professional advice of any kind. Usage of this posting's information is solely at reader's own risk.
Liability Disclaimer
In no event shall Author be liable for any damages whatsoever (including, without limitation, damages for loss of use, data or profit) arising out of the use or inability to use the posting's information even if Author has been advised of the possibility of such damage.
Posting
The original Stackwise has, as I've noted, two 8 Gbp (duplex) stack ports per switch. That's where the 32 Gbps, i.e. 8 * 2 (for duplex) = 16 Gbps * 2 (ports) 32 Gbps.
With StackwisePlus, Stackwise is described as all a switch's internal traffic is placed onto the stackring. Additionally, the switch placing traffic onto the stackring is also the switch that removes the traffic (after it's transversed all the other switchs in the stack). Effectively, the stackring performs much like the 6500's classic bus (also 32 Gbps [16 Gbps, duplex], BTW). I.e. All the stack members share this bandwidth.
PS:
BTW, StackWisePlus besides doubling stack port bandwidth, uses its stackring bandwidth much, much better. First, it only places unicast traffic on the ring that is non-local to the stack member. Second, the destination stack member removes the traffic from the stackring.
The 3850's new Stack480 hasn't, as far as I know, yet documented how it uses its ring.
