A company wants to deploy VMware NSX for vSphere and ensure the least amount of bandwidth
consumption in the underlying physical architecture. Which replication mode should the logical
switches be deployed with?
A.
Multicast Replication Mode
B.
Unicast Replication Mode
C.
Hybrid Replication Mode
D.
vSphere Replication Mode
C
But others say A
Definitely A, the downside for a L3 network is that it requieres configuration of PIM.
Multicast -> one transmit
Hybrid (L3) -> one multicast transmit for L2 segment, unicast to the VTEP proxy (UTEP)
But with multi-cast mode there is no BUM traffic reduction. So a broadcast frame will be sent out of the source VTEP as a multi-cast frame which will then be propagated using PIM across to other segments. Whereas if you use Hybrid, then you still have controllers that will inform the ESXi host of which M-TEP to send the traffic to, which will then replicate locally using multi-cast, but you get all the benefits of the controllers to reduce arp traffic etc.. So I believe the answer is C.
Just food for thought, what about “B” because NSX in unicast Mode will use remote proxy UTEP to replicate in the segment. The Higher overhead is on each VTEP and UTEP and increase CPU on the hosts.
“In NSX, the default mode of traffic replication is unicast. Initially no multicast support is required on the physical network”
“This mode reduces network dependencies to only increase in MTU”
Hybrid makes more sense: http://planetvm.net/blog/?p=2894
Unicast Mode:
The major downside is that in the case of many segments, many packets may be sent from a single server rather than just one in the Multicast mode. This can eat bandwidth in high density cases.
I will say B for the least bandwidth, because Unicast or Hybrid mode also includes ARP suppression. But Hybrid mode will multicast to multicast group and unicast to UTEP, which increases bandwidth compared to unicast only.
https://telecomoccasionally.wordpress.com/2015/01/11/nsx-for-vsphere-vxlan-control-plane-modes-explained/
In Multicast mode, host has no other choice but to flood a frame for which it doesn’t have a MAC:VTEP entry, in hope that one of the copies of flooded frame will reach the destination VM, causing it to respond back, so that the above process can complete.
Flooding will also create MAC:VTEP cache entry for the source VM’s MAC and VTEP of the host it’s running on on all hosts participating in VXLAN this VM is connected to.
In Hybrid and Unicast mode, hosts will first query Controller for this mapping, and use information in Controller’s response to populate their local cache, avoiding initial flooding.