[MUD-Dev] Re: Encryption of protocols, compression and lag...

["Adam J. Thornton" <adam#phoenix,Princeton.EDU>]

On Tue, Sep 01, 1998 at 07:22:13PM +0200, Ola Fosheim Grøstad wrote:
> Well, Caliban, if you or somebody else on this list really have this
> problem, then I simply suggest that you send me a mail with pointers to your
> software and the crypto DLL.  I will then simply add the DLL to your zipfile
> and put it on a scandinavian ftp-site.  Import of encryption code isn't
> illegal is it? (On the hysterical side, any programming toolkit with a
> decent random generator is clearly an encryption library. And what about
> sound editors? Would you US guys really worry about adding noise to a
> signal??? Anyway, I believe PGP was exported as printouts and has been
> OCRed.)

Yeah.  Well, the nice thing about ITAR is that it's wonderfully selectively
enforceable; yes, a program linked against a C library with a decent PRNG
*could* buy you a federal prison term, if the feds decided to prosecute


But if it's on paper, it's OK.  Yet another quirk.  Hence the big huge PGP
Internals book.

Anyway, this is getting less and less relevant to Mud-Dev.  The point being
the crypto DLL needs to be developed outside the US, in someplace with
friendly encryption laws.

> Is, however, encrypting the protocols A Good Idea (tm)?  I would think not,
> unless you compress the stream first.

That is, in fact, the usual technique.  What you probably want to do is to
compress the data stream first, then pad to blocksize (since this is
usually 8 or 16 characters, it's not a huge waste), then encrypt.

> Wouldn't encryption just increase
> bandwidth?

Yes, because the outgoing data stream will have near-total entropy and thus
be in effect uncompressible.

> would like to exploit that piece of hardware, or wouldn't you?

Can't.  Your data stream is going to be entropic, so you do software
compression first and let the modems cope.

Adam
-- 
adam#princeton,edu 
"There's a border to somewhere waiting, and a tank full of time." - J. Steinman