RB3011 throughput?

How do I read the information about throughput? Is it only telling the lan-routing?
I´m interested in throughput wan - lan.

If I remember correctly, Mikrotik does this test using all ports, with traffic flowing in a single direction in each port, crossing the router.

Look at the table, and compare it with the block diagram. The RB3011 has four gigabit links - two from each switch to the CPU. Hence, the maximum speed of almost 4 Gb.

And I need to add that it does not matter how you label the ports. The wan/lan is only some reduced way of thinking about the routing.

This is no small task to test actual throughput because there is multiple ways to configure a MikroTik router. Taking into consideration features like FastTrack, FastPath, using Filter rules and SPI Firewall, internal Switch being used without Bridge, sizes of packets and different protocols passing through, etc. Many things will hog CPU or go easy on it.

I’m currently testing 2 MikroTik routers (RB750Gr3 (hEX v3) and hAP mini) to see if they can perform as advertised.
For now I can tell that “hAp mini” can’t provide more than 97Mbit UDP and 57Mbit TCP in both directions via 2 Bonded ports without maxing out its CPU. This is without using Bridge and with disabled Firewall.
It’s a toy, so don’t expect 150Mbit via Wireless interface from it or anything special in terms of performance.
So by looking at table for hAP mini, I can safely assume numbers from it could be easily divided by half or more:

mikrotik web site tests are useful to compare devices performance, but nobody can predict the performance of a device on a specific configuration or scenario because the possible combinations are infinite.

for example a simple nat + fastrack configuration can get several hundred of mbps even on small devices, but put some firewall and QoS and the performance drops to 1/10 of that

for example a heavy QoS configuration can kneel down a rb1100ahx2 to 100mbps of internet throughput from the maximum theoretical of 5000 mbps

for example a heavy QoS configuration can kneel down a ccr1036 to 1000mbps of internet throughout from the maximum theoretical of 28000 mbps

Its very frequent that a isp want to improve their configuration on a well performing device but the new more complex configuration leaves the device with cpu usage on 100% hurting performance

and thats ok its because the versatility of routeros, performance allways depend of the scenario and configuration

there is no shortcuts in this you need to test your configuration on a device in real conditions to know the expected performance

Hopefully one day when Router OS 7 releases we can have filter rules that spread the load over all cores

@sizeofbool if you’re just using the internal btest function to test your mini, you wont be able to reach the maximum available throughput that way regardless due to the intensity of btest on the cpu.