I just wanna now how many real tcp/ip throughput RB1000U can handle only on routing.
I just tried using mikrotik bw-test its can handle only about 17mbps up and 97mbps down with 100% cpu resources. We had a STM1 connection, should we using a PC or thereis any other way?
Thanks.

Did you push those packets through the router, or did you have the router generate or directly receive the packets used for the bandwidth test? There's a huge performance difference between the two.

Did you push those packets through the router, or did you have the router generate or directly receive the packets used for the bandwidth test? There's a huge performance difference between the two.

rb1000u act as bw-test server and the other as bw-test client. how big the different is?

It's a huge difference. Generating packets is a lot more work than routing packets, and in normal operations only a tiny fraction of all traffic is generated by or terminated on the router so your test doesn't at all simulate real world traffic conditions.

Use two PCs with the router in the middle.

It's a huge difference. Generating packets is a lot more work than routing packets, and in normal operations only a tiny fraction of all traffic is generated by or terminated on the router so your test doesn't at all simulate real world traffic conditions.

Use two PCs with the router in the middle.

ok, thanks fewi.
Usually, how much tcp/ip real traffic rb1000u can handle?just as reference.

Depends on your configuration. For just routing without any QoS or firewall rules the data sheet contains the official test data, it's available on routerboard.com. Any QoS rule or firewall rule means that the router spends time not routing, affecting throughput performance.

Depends on your configuration. For just routing without any QoS or firewall rules the data sheet contains the official test data, it's available on routerboard.com. Any QoS rule or firewall rule means that the router spends time not routing, affecting throughput performance.

Thanks again fewi