Hello,
If i would like to measure the impact of the VSAN network as part of the latency, how would one best go about it?
Example: A write takes place. It is of course shipped over the VSAN Network to another host as a component of the object lies there too. Transporting the data over the network and handing it over to the physical disk layer take time.
Using vSAN Observer, it is easy to see the latency of the physical disks and how much they contribute to latency. Looking at latency from the "client" level however shows a different picture which is logical as it's higher up in the stack.
As i cannot afford 10Gbit Ethernet, my Lab network is a 1Gbps network where each node has two NIC's in a LACP LAG and i clearly see a node talk to one other node over 1 NIC and talk to another node over the other NIC as LACP nicely splits the traffic over the two cards.
During rebuilds I notice a nice full utilization of both links on servers so from a bandwith-perspective, using LACP like this seems to be worth it. But would it increase latency to an undesirable extent? I also noticed that during normal operations, the 1Gig links never even come close to getting saturated, kind of deeming 10Gig Ethernet unjustified for my small 4 node environment (i run about 80VM's on it).
What i would like to be able to do is determine how much latency is added by the networking-layer.
