Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!swrinde!elroy.jpl.nasa.gov!jato!jdickson From: jdickson@jato.jpl.nasa.gov (Jeff Dickson) Newsgroups: comp.sys.amiga.tech Subject: Re: How to LOCK your Amiga? Message-ID: <1991Jan8.163007.3786@jato.jpl.nasa.gov> Date: 8 Jan 91 16:30:07 GMT References: <1991Jan6.233133.1@csc.fi> <1025@macuni.mqcc.mq.oz> Reply-To: jdickson@jato.Jpl.Nasa.Gov (Jeff Dickson) Organization: Jet Propulsion Laboratory, Pasadena, CA Lines: 105 In article <1025@macuni.mqcc.mq.oz> ifarqhar@sunc.mqcc.mq.oz.au (Ian Farquhar) writes: >Actually, there may be a solution to this that would be reasonably >secure, though it isn't pretty. Write a handler that sits between the >l:fastfilesystem handler and the dos.library (ie. handler entry in the >devs:mountlist points to your handler and your handler loads the >l:fastfilesystem handler), which encodes all data going to and coming >from the disk. > >The hardest is encoding the data. You have to decide two things: > >1) What to encode. You could only encode the data going to and coming > from the read and write packets, so that the directory structure > looked fine, but the data in these files was encrypted. Or you > could encode everything, which is less safe (as recovering corrupted > data would become a real pain) but more secure. > >2) How to encode. The ideal method would be a one-time pad, but this > is a little difficult to arrange :-) DES is an option, but it is > notoriously slow to implement in software. I would tend to go for > a multi-barrel 256-slot enigma machine, which is fast to implement > in assembler (it's all tables and pointers), and reasonably secure. > If you had an 8 character password, this would yield 56 usable bits > (in 7 bit ASCII), which would mean that it could set up to seven > 256 slot barrels. That is not going to stop the NSA, but it will > put Joe Average right off. > >The password, of course, is established when the handler is fired up. >This could be done by a floppy (not advised, as it would contain the >barrel tables which would not necessarily compromise the encryption, but >make it somewhat easier to break), or much better to have it on a hard >disk partition in the old filing system, from which the secure one >boots. > >Comments, anyone? A small point: the file comment field makes the ideal >place to store security information if you wanted to extend this >further. > >-- >Ian Farquhar Phone : 61 2 805-9400 >Office of Computing Services Fax : 61 2 805-7433 >Macquarie University NSW 2109 Also : 61 2 805-7420 >Australia EMail : ifarqhar@suna.mqcc.mq.oz.au Newsgroups: comp.sys.amiga.tech Subject: Re: How to LOCK your Amiga? Summary: Expires: References: <1991Jan6.233133.1@csc.fi> <1025@macuni.mqcc.mq.oz> Sender: Reply-To: jdickson@jato.Jpl.Nasa.Gov (Jeff Dickson) Followup-To: Distribution: Organization: Jet Propulsion Laboratory, Pasadena, CA Keywords: In article <1025@macuni.mqcc.mq.oz> ifarqhar@sunc.mqcc.mq.oz.au (Ian Farquhar) writes: >Actually, there may be a solution to this that would be reasonably >secure, though it isn't pretty. Write a handler that sits between the >l:fastfilesystem handler and the dos.library (ie. handler entry in the >devs:mountlist points to your handler and your handler loads the >l:fastfilesystem handler), which encodes all data going to and coming >from the disk. > >The hardest is encoding the data. You have to decide two things: > >1) What to encode. You could only encode the data going to and coming > from the read and write packets, so that the directory structure > looked fine, but the data in these files was encrypted. Or you > could encode everything, which is less safe (as recovering corrupted > data would become a real pain) but more secure. > >2) How to encode. The ideal method would be a one-time pad, but this > is a little difficult to arrange :-) DES is an option, but it is > notoriously slow to implement in software. I would tend to go for > a multi-barrel 256-slot enigma machine, which is fast to implement > in assembler (it's all tables and pointers), and reasonably secure. > If you had an 8 character password, this would yield 56 usable bits > (in 7 bit ASCII), which would mean that it could set up to seven > 256 slot barrels. That is not going to stop the NSA, but it will > put Joe Average right off. > >The password, of course, is established when the handler is fired up. >This could be done by a floppy (not advised, as it would contain the >barrel tables which would not necessarily compromise the encryption, but >make it somewhat easier to break), or much better to have it on a hard >disk partition in the old filing system, from which the secure one >boots. > >Comments, anyone? A small point: the file comment field makes the ideal >place to store security information if you wanted to extend this >further. > >-- >Ian Farquhar Phone : 61 2 805-9400 >Office of Computing Services Fax : 61 2 805-7433 >Macquarie University NSW 2109 Also : 61 2 805-7420 >Australia EMail : ifarqhar@suna.mqcc.mq.oz.au The problem with having an additional channel for disk data is that it greatly slows the system. Why not request an exclusive write lock on the root of the filing system. Others users could use your stuff, but they couldn't overwrite or delete it and that's what you're after - right? Jeff