home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!haven.umd.edu!darwin.sura.net!mips!swrinde!zaphod.mps.ohio-state.edu!usc!isi.edu!allard
- From: allard@isi.edu (Dennis Allard)
- Newsgroups: comp.protocols.ppp
- Subject: Re: PAP & CHAP
- Message-ID: <22257@venera.isi.edu>
- Date: 21 Aug 92 18:22:03 GMT
- References: <1992Aug20.195258.13900@lambda.msfc.nasa.gov> <1992Aug20.174748.23770@ornl.gov>
- Sender: news@isi.edu
- Reply-To: allard@isi.edu (Dennis Allard)
- Organization: USC Information Sciences Institute
- Lines: 38
-
- mhump@sparky.msfc.nasa.gov (Mark Humphries) writes:
- someone else had written:
- > > Could someone explain to me the benefit of PAP/CHAP authentication ...
- >
- > Password Athentication Protocol (PAP) uses a username and password pair
- > for user validation. ...
- >
- > Challenge-Handshake Authentication Protocol (CHAP) uses a "secret" that
- > is known by the peer (the host attempting to dial in) and the authenticator
- > (PPP server/host being dialed into). ...
-
- An important issue here has to do with what level connections are being
- made at (application vs. session or transport) and whether or not human
- intervention is appropriate or convenient each time a connectio is made.
-
- I am new to protocols, new to this newsgroup, and know little about PPP,
- but here is my two cents on this matter.
-
- I got interested in protocols because I recently built my a session level
- message protocol over Unix sockets (TCP/IP) and an RPC protocol
- over that. This was in the context of an Emacs/Lisp interface. In this
- application, new connections are being made at the session level, and
- are often made by processes without the intervention of a human. The
- nature of the connection permits complete access to a Lisp address space
- by a different Lisp or Emacs running anywhere on the internet. Although
- we have not exported this facility yet, we initially put in a very simple
- 'password' scheme, consisting of an integer, which the server end of a
- new connection knows and which the client end must send to connect. (Once a
- connection is made, the server and client have completely symmetrice roles).
- Typically, we just use the integer 0, inhouse, since we trust each other.
-
- In considering how to provide a more secure connection mechanism, I was
- intriqued to see this discussion in this newsgroup, which, I repeat, I
- just joined. I will follow this with interest and try to learn more
- about PPP. Maybe that's what I should be using.
-
- Dennis
- allard@isi.edu
-