FreeTDS Development Group

[Brian Bruns <camber AT umcc.ais.org>]

On Tue, 26 Jan 1999 tetherow AT nol.org wrote:

> I can't image how you could bring down a database with the wrong
> endian-ness (not saying it is the case, but that it just baffles me
> that someone could write a DB server that allows that kind of
> misbehaviour :)  It still bothers me that sybase/MS didn't use network
> byte order for sending stuff across the network anyway.
>  

Well, this isn't always the case, IIOP for instance will keep the
clients (I think) byte order. The reason being that byte swapping slows
things down if client and server are the same but opposite from network
byte order.  It also presumes a powerful server that can take care of the
byte swapping instead of making the client handle it. And remember that
Sybase runs across a variety of transports including ipx (ugh) and decnet,
so network order may not apply. So, its a trade off. Most internet
protocols like NNTP or SMTP assume peer machines of roughly similar
computing power, TDS is definately rooted in more of a client/server
world.

That said, I've seen some indications that certain things depend on the
servers byte order, but didn't have enough time to verify it.

I have yet to stack trace our big endian boxes (AIX), but I wouldn't say
its impossible.  Certainly leaves Sybase open to denial of service
attacks.

Brian