On Mon, 2004-10-11 at 14:19, Jon Bendtsen wrote: > Den 11. okt 2004, kl. 13:54, skrev Jörn Hartmann: > > Maybe I'm wrong, but I think think it's hardly possible to strip > > openssl down to less than 32K and keep full SSL functionality. You > > might have more luck in writing your own SSL implementation. That will > > be tough enough assuming you got only a fraction of the 32K left for > > the encryption part. > > i know, and i might consider making a broken SSL implementation, such > that only some clients can connect, and stuff like that. Further more > i do not need to generate a certificate inside the device. I was > thinking of keeping: > one of AES/blowfish/3DES (depending on what browsers support). > RSA > certificate checking > CRL checking
Even with some heavy chainsaw work on OpenSSL, I don't think you'll manage getting down to 32KB. I saw some numbers on the list in the past of people targetting embedded platforms, but they were of several 100KB. I suspect a lot of people would be interested in a version of OpenSSL supporting the subset and sizes you are talking about. Unfortunately this means a lot more work than not compiling some unnecessary fetures. If you don't mind working with GPL code, you should look at http://www.matrixssl.org/ because they are down to the order of magitude you are talking about (http://www.matrixssl.org/docs.html). If GPL is an issue for your work, you can buy a licence from them. Or from other proprietary editors: Mocana (http://www.mocana.com/ssl.html), Spyrus (http://www.spyrus.com/content/products/DeviceSSL_N7.asp), etc... Sincerely, -- Mathias Brossard <[EMAIL PROTECTED]> ______________________________________________________________________ OpenSSL Project http://www.openssl.org User Support Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]