Re: [lwip-users] Performance drop due port numbers reused too fast

[Indan Zupancic]

Hello Jochem,

On 2022-10-19 18:51, Jochen Strohbeck wrote:
> from time to time I experience a significant network performance drop.

Is this during normal use or during load testing/benchmarking?

> I'm running an lwip http 1.0 server on SAME70/FreeRTOS and do http
> requests on Windows or Linux.
>
> In the wireshark trace I see ACK+RST after a client SYN in case a port
> number is used in the request which has been already used a few
> milliseconds before.

lwIp does the right thing here, and one extra RST to force Windows to
use another port number should not cause a huge performance drop.

> I wonder why http clients running on Windows or Linux are re-using port
> numbers that fast. Shouldn't the OS wait for at least MSL before
> re-using ports?

Yes, but it's a lot easier to let the remote check for port-reuse than
to check all sockets in TIME_WAIT. What happens is that the port is
chosen randomly.

> I guess this can be omitted using persistent connections. Is this the
> only remedy to that problem?

I think the main issue is that Windows propagates the RST as an error
to the program doing a connect, instead of silently retrying with a
different port number. So the huge performance drop may be caused by
the program waiting before reconnecting.

You can reduce this problem by increasing the ephemeral port range on
the client PC.

Greetings,

Indan