General Advice

Your requirements don't directly mandate an MMU, but you'll possibly wind up with one for free anyway. Sharp has an ARM7 with on-chip MAC that might be interesting. LH79525,

You can actually do everything you want in an OSless 8-bit micro, but the annoying part is sourcing and integrating the MAC. I'd suggest that Sharp part and eCos.

Zilog eZ80f91 sounds about right. They are a bit expensive in small quantities, but otherwise seem to have everything else you want. Development system with C, OS and lots of examples $99 I think.

Paul Burke

See Rabbit Semiconductor as a starting point. They make 8-bit modules that have Ethernet and limited SSL support.

These guys may have SSL in the works:

Richard

Maybe check out Axis ETRAX MCM4+16, or the 100LX.

SNIP

SNIP

If you want to run SSL you will need a 32 bit processor (maybe you get away with 16 bits with clever coding and fast enough clock). For an 8 bit processor AES is a better choice than Blowfish. Or use a crypto-coprocessor (which are rare and expensive) or a DSP

Wim

See _Implementing SSL on 8-bit micros_ at

away

crypto-coprocessor

Interesting article! Is RC4 freely available today or covered by patents? Some time ago it was a Mirocsoft prorietary algorithm.

Wim

Close, but incorrect. RC4 is and always has been owned by RSA Data Security, Inc. Jim Bidzos, President of RSA, decided long ago to not patent RC4 but to instead rely upon trade secret protection - which means that RSA has never revealed the RC4 algorithm to the public, a requirement for obtaining a patent.

Revealing, publishing, or using a trade secret is not illegal. Trade secrets are protected by contracts instead of laws. If you have signed a contract with RSA, you are bound by that contract and could be sued if you violate it.

If, on the other hand, you don't have a contract with RSA and are using a copy that has been reverse-engineered, you are free and clear to use it.

An interesting aspect of this is that we don't *really* know whether what we call "RC4" is the same as the RC4 owned by RSA. We do know that the two can encrypt and decrypt each other's output, and that this has been tested on millions of files. For this reason, some people call the RC4 that we use "ARCFOUR" instead of "RC4."

Bottom line: go ahead and use it.

Nonetheless, RC4 is a trade secret, whether you think so or not.

In other words, something that we are pretty sure is RC4, but we are not completely sure ss RC4, puplished by someone who may or may not have correctly reverse-engineered RC4.

Comments by pepole who *don't* own RC4 have no validity. The official comment by the people who *do* own RC4 is:

"RSA Security does not hold any patents nor does it have any pending applications on the RC4 algorithm. However, RSA Security does not represent or warrant that implementations of the algorithm will not infringe the intellectual property rights of any third party. Proprietary implementations of the RC4 encryption algorithm are available under license from RSA Security Inc. For licensing information, contact: RSA Security Inc. 2955 Campus Drive, Suite 400, San Mateo, CA 94403-2507, USA, or

Are you implying that RSA does not own RC4?

Guy Macon wrote on Mon, 13 Dec 2004 04:36:09 +0000:

If RC4 isn't a secret, it isn't also a trade secret. IMO it is logical for an average engineer. Although I agree that lawyers (and marketroids) use different kind of logic. ;-)

Free RC4 implementations have been widely used for 10 years. IMO it is long enough to prove that they actually implement RC4.

I'm not. I'm only saying that R. Rivest is the inventor of RC4. Another interesting link:

Regards, /J.D.