Hi all,
This is a new setup, 2x HPE (NVIDIA) SN3420M 25GB/100GB switches in MLAG with 4x servers running Windows 2025 connected redundantly to each switch using Broadcom NICs. We have managed to fine tune the iperf/ntttcp transfer rates but still seeing fluctuations and considerably high ICMP response times between hosts directly connected to the switch (es). Reading on other reddit posts relating to ICMP (Ping), it has been “downvoted” to ‘measure’ networking performance using ICMP. Is this right and is the below ‘ok’ for a high speed DC switch?
Source address is XX.XXX.XX.211; using ICMP echo-request, ID=147c
Pinging XX.XXX.XX.212 [XX.XXX.XX.212]
with 32 bytes data (60 bytes IP):
From XX.XXX.XX.212: bytes=60 seq=0001 TTL=128 ID=e40a time=1.182ms
From XX.XXX.XX.212: bytes=60 seq=0002 TTL=128 ID=e421 time=1.293ms
From XX.XXX.XX.212: bytes=60 seq=0003 TTL=128 ID=e43b time=1.111ms
From XX.XXX.XX.212: bytes=60 seq=0004 TTL=128 ID=e4aa time=1.321ms
From XX.XXX.XX.212: bytes=60 seq=0005 TTL=128 ID=e4c3 time=0.979ms
From XX.XXX.XX.212: bytes=60 seq=0006 TTL=128 ID=e4da time=0.766ms
From XX.XXX.XX.212: bytes=60 seq=0007 TTL=128 ID=e4f3 time=0.985ms
From XX.XXX.XX.212: bytes=60 seq=0008 TTL=128 ID=e50a time=1.226ms
From XX.XXX.XX.212: bytes=60 seq=0009 TTL=128 ID=e523 time=1.156ms
From XX.XXX.XX.212: bytes=60 seq=000a TTL=128 ID=e5b9 time=1.158ms
From XX.XXX.XX.212: bytes=60 seq=000b TTL=128 ID=e5d2 time=0.948ms
From XX.XXX.XX.212: bytes=60 seq=000c TTL=128 ID=e5e9 time=1.078ms
From XX.XXX.XX.212: bytes=60 seq=000d TTL=128 ID=e60e time=1.028ms
From XX.XXX.XX.212: bytes=60 seq=000e TTL=128 ID=e625 time=1.138ms
From XX.XXX.XX.212: bytes=60 seq=000f TTL=128 ID=e64c time=1.143ms
From XX.XXX.XX.212: bytes=60 seq=0010 TTL=128 ID=e663 time=1.166ms
From XX.XXX.XX.212: bytes=60 seq=0011 TTL=128 ID=e67c time=1.217ms
From XX.XXX.XX.212: bytes=60 seq=0012 TTL=128 ID=e6a9 time=1.149ms
From XX.XXX.XX.212: bytes=60 seq=0013 TTL=128 ID=e6c2 time=1.176ms
From XX.XXX.XX.212: bytes=60 seq=0014 TTL=128 ID=e6e3 time=1.209ms
From XX.XXX.XX.212: bytes=60 seq=0015 TTL=128 ID=e6fc time=1.364ms
From XX.XXX.XX.212: bytes=60 seq=0016 TTL=128 ID=e713 time=1.214ms
From XX.XXX.XX.212: bytes=60 seq=0017 TTL=128 ID=e72c time=1.261ms
From XX.XXX.XX.212: bytes=60 seq=0018 TTL=128 ID=e743 time=0.930ms
Cheers!