[A86] Re: Async Link Routine

To: <assembly-86@lists.ticalc.org>
Subject: [A86] Re: Async Link Routine
From: "Henk Poley" <HPoley@DDS.nl>
Date: Wed, 13 Feb 2002 13:58:47 +0100
Reply-To: assembly-86@lists.ticalc.org


Ah, so that's it. I always found it strange that when I tried to tenet my
mailserver to check e-mail it was that slow, while that way of
communicating was standard on BBSes. It's the size overhead of the TCP/IP
packets...

> The best way to handle something like this is a shell account that you
can
> dial into directly.  You would have much, much less lower latency, and it
> would actually be usable.  Ever wonder why you could easily use a BBS at
> 2400 or 9600 baud, but an SSH connection on a modem is pratically
unusable?
> That's because every keystroke requires 90-100 bytes of data.  That
doesn't
> seem like a lot for a "fast modem", but it still means you get 200-250ms
or
> higher pings.  I sure type a lot faster than that.  It's much better when
> your keystroke and the response arrive immediately.
> 
> > Is there any sort of software flow control or error correction in the
> > protocol that you could take advantage of?
> > If you're putting a higher level network stack on top of it, could you
> > just ignore errors at this level and let the checksum or whatever take
> > care of it?
> 
> 
>

Follow-Ups:

[A86] Re: Async Link Routine

From: "David Phillips" <david@acz.org>