On 08/02/2018 12:34 AM, xlar54 wrote:
> at the moment, webservers dump the pages so quickly that the wifi modem
> and/or code cant keep up.

This seems like a modem / gateway / web server failing to me.

To me, there are two communication streams, one on either side of the
gateway (like the Wifi232).  One is serial to the old computer.  The
other is TCP/IP over some sort of network (wired or wireless).

I would expect that the serial side can be paused by (I think) dropping
DTR.  (Maybe it's a different signal.)  The point being for the (retro)
computer to indicate that it is not ready to receive any more data from
the other end.

I would also expect that the TCP/IP side could similarly cause the web
server to pause sending data.  I'm not completely familiar with the
mechanics of how.  But I am confident that the TCP/IP client can tell
the sending web server how much buffer it has available.  The idea being
that the web server's sending side should not exceed said buffer.

So … I feel like this problem (as I understand it) is solved.  Or at
least it should be.  Hence why I feel like there is a failing somewhere,
be it a bug or lack of leveraging the existing mechanisms.

The Slowloris attack comes to mind as supporting evidence that a TCP/IP
client can cause a web server to slow down.  }:-)



--
Grant. . . .
unix || die

--- SoupGate-Win32 v1.05