I have 1G/35M at home, but have a relative with symmetrical 2G. I duplicated my QNAP box and hooked it into a hEX at their place, then set up two WG tunnels between it and my 5009; traffic to .124/24 on my end ends up on their end, traffic to .126/24 on their end ends up on mine.
But I also expected to do QNAP RTRR backups and NFS mounts from them to me and only getting ~175Mbits/sec thruput, and was surprised at such a slowdown, also verified via iPerf3. One thing I noticed in /ip/firewall/connections is there's no Fastrack on either side of this tunnel. If I dst-nat a hole from my side's external IP to the iPerf server on their side (i.e., bypassing the local-network tunnel) speeds are in the ~750Mbit range, as expected, so it's not a thruput issue per se.
Re: Are Fastrack and Wireguard mutually exclusive?
Yes, fasttrack does not play any role in the wireguard process.
But ... major thing to consider:
the complete wireguard process is 100% handled by CPU. No HW offloading, no shortcuts.
So I'm thinking that Hex is the bottleneck here as far as resources are concerned.
When you do that test, observe CPU on Hex. I think you will see 100% (or close) on it.
If you happen to have a beefier Mikrotik avaible, redo the test and see what it gives then.
But ... major thing to consider:
the complete wireguard process is 100% handled by CPU. No HW offloading, no shortcuts.
So I'm thinking that Hex is the bottleneck here as far as resources are concerned.
When you do that test, observe CPU on Hex. I think you will see 100% (or close) on it.
If you happen to have a beefier Mikrotik avaible, redo the test and see what it gives then.
Who is online
Users browsing this forum: seriosha and 60 guests