[e2e] UDP length field

Vernon Schryver vjs at calcite.rhyolite.com
Sat Apr 21 06:46:33 PDT 2001

> From: Thomas "H." Ptacek <tqbf at sonicity.com>

> > The validity checks is another issue, all UDPs that we've 
> > looked at seems to adhere to Craigs rules. There is one 
> > exception which is Quake's home-grown UDP. They do some
> Quake has its own UDP? I can see (evil) reasons for building ones
> own TCP, but almost no benefit to a custom UDP. Is this an OS issue
> (ie, they reimplemented sockets for speed) or did they also build
> their own IP? Does anyone know the answer to this?

Consider their target market, recall when they hit it, and ask yourself
if you would have wanted to rely on the available UDP/IP application
*binary* interfaces available then.  Does't Quake like to use multicasting,
and wasn't that support a little thin in Winsock 1.0?  Was there any
sort of Winsock library for Novell systems when Quake first appeared?

In ancient days, I tried to convince customers of my then employer to not
use raw Ethernet packets.  Many people in the 1980's were convinced beyond
reasoning that UDP/IP and TCP/IP were unavoidably slow and impossible to
implement on many systems.  They may have had help from other vendors in
their belief that TCP/IP could not handle 20 messages per second.
Sometimes I was able to convince them to put 28 constant bytes between
their Ethernet headers and payloads and that their other, IP-stupid systems
could ignore.

Vernon Schryver    vjs at rhyolite.com

More information about the end2end-interest mailing list