Glad you got the tunnel up, droogiee. If you are now maxing around
80 Mbps, that is about:
80 Mbps ÷ 8 = 10 MB/s
So yes, that is still far below what a clean
500/500 site-to-site link should be capable of.
Most likely causes
The big suspects are:
- Synology router CPU/encryption bottleneck
- IPsec settings causing slower encryption
- MTU/MSS fragmentation
- SMB/file copy inefficiency over routed VPN
- Traffic inspection/QoS/firewall processing still active
- ISP routing/latency/packet loss between the two sites
- One side not actually getting full 500 Mbps upload at the time of testing
- Testing over Wi-Fi instead of wired Ethernet
Since you tested multiple PCs both ways and disabling
Safe Access /
Threat Prevention helped slightly, that strongly suggests the routers are doing extra packet processing and/or hitting a processing limit.
First: separate VPN speed from file-copy speed
Do not use SMB copy speed as the first benchmark. Test the tunnel with
iperf3.
1. On a wired PC at work
iperf3 -s
2. On a wired PC at home
iperf3 -c WORK_PC_IP
3. Test with multiple streams
iperf3 -c WORK_PC_IP -P 8
4. Test reverse direction
iperf3 -c WORK_PC_IP -P 8 -R
What the results mean:
Code:
iperf3 around 80 Mbps = VPN/router/ISP/MTU bottleneck
iperf3 much faster but SMB slow = Windows file sharing/SMB issue
one direction fast, other slow = upload/routing/ISP/asymmetric issue
single stream slow, multi-stream faster = latency/TCP window issue
Check router CPU during the test
While running
iperf3, watch both Synology routers.
1. Open SRM on both routers
Go to:
SRM > Resource Monitor
2. Run the VPN speed test
Use:
iperf3 -c REMOTE_PC_IP -P 8
3. Watch CPU usage
If either router hits high CPU during the VPN test, that router is the bottleneck.
This is very possible. VPN throughput is not just “internet speed.” The router has to encrypt/decrypt every packet.
Check your Site-to-Site encryption settings
For testing only, use a lighter-but-still-reasonable config.
Try:
Code:
IKE version: IKEv2
Encryption: AES-128 or AES-256
Authentication: SHA-256
PFS: disabled temporarily for testing
Security note: Disabling PFS is only for speed testing. If it makes a big difference, you found part of the bottleneck. You can then decide whether the speed gain is worth the security tradeoff.
MTU/MSS may be killing throughput
Bad MTU can absolutely cap VPN speeds.
1. Test ping with no fragmentation
From one side to the other:
ping REMOTE_PC_IP -f -l 1472
If that fails, lower the number:
Code:
ping REMOTE_PC_IP -f -l 1400
ping REMOTE_PC_IP -f -l 1360
ping REMOTE_PC_IP -f -l 1320
The largest value that works helps estimate the usable MTU.
2. Try lowering VPN MTU/MSS if Synology allows it
Common working values:
Code:
MTU: 1400
MTU: 1380
MTU: 1360
If there is an
MSS clamping option, try around:
MSS: 1360
If MTU is the problem, you may see a large jump immediately.
Make sure the test PCs are wired
For accurate testing:
- Use Ethernet on both ends.
- Confirm link speed is
1.0 Gbps.
- Do not test over Wi-Fi.
- Avoid testing to/from a sleeping NAS, old laptop, USB Ethernet adapter, or powerline adapter.
On Windows, check link speed:
Get-NetAdapter | Select Name, Status, LinkSpeed
Check for remaining Synology features that can slow routing
Even with Safe Access and Threat Prevention disabled, check these too:
Code:
Traffic Control
QoS / bandwidth control
Firewall logging
Smart WAN policy routing
Application Layer Gateway features
DoH/parental filtering
IDS/IPS remnants
Device-specific bandwidth limits
VPN logging/debug mode
For testing, temporarily simplify both routers as much as possible.
Undo: Re-enable the protections after testing once you know which feature is affecting throughput.
SMB may still be slower than the VPN
If
iperf3 gives you decent numbers but file copies are still poor, use
robocopy with multithreading instead of Explorer drag-and-drop.
Example:
robocopy "\\REMOTE_PC_IP\Share" "D:\TestCopy" /E /MT:16 /R:2 /W:2
Or:
robocopy "D:\TestFolder" "\\REMOTE_PC_IP\Share" /E /MT:16 /R:2 /W:2
Also test with one large file, not thousands of tiny files. Tiny files are much slower over VPN.
Test the raw internet path
Run speed tests at both sites using wired PCs, but also test latency and packet loss between the two public IPs if possible.
From one side:
ping OTHER_SITE_PUBLIC_IP -n 50
Also:
pathping OTHER_SITE_PUBLIC_IP
High latency, jitter, or packet loss can destroy VPN throughput even when both sites show “500/500” to a nearby speed test server.
What speed should be realistic?
With 500/500 at both ends:
Code:
Theoretical max: 500 Mbps / 62.5 MB/s
Good site-to-site VPN: 250–450 Mbps
Excellent setup: 400+ Mbps
Current result: 80 Mbps / 10 MB/s
So your result is not normal if both routers are healthy, wired, and configured well.
My bet
Given that disabling
Safe Access and
Threat Protection helped, I suspect one of these:
Code:
Router CPU/packet-processing limit
MTU/MSS issue
Remaining SRM traffic inspection/QoS/firewall processing
IPsec encryption setting causing poor acceleration
What to send Synology support
Send them these exact test results:
Code:
Router model at home:
Router model at work:
SRM version on both:
VPN Plus Server version on both:
Site-to-site encryption settings:
Home LAN subnet:
Work LAN subnet:
iperf3 single stream result:
iperf3 -P 8 result:
iperf3 -P 8 -R result:
CPU usage on both routers during test:
MTU test results:
Whether Safe Access/Threat Prevention are disabled:
Whether clients are wired or Wi-Fi:
If you post your
iperf3 results here, especially single-stream,
-P 8, and reverse
-R, I can help narrow down whether this is
router CPU,
MTU,
SMB, or
ISP path.