Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 28 May 1998 21:14:44 EDT
From:      "Angelos D. Keromytis" <angelos@dsl.cis.upenn.edu>
To:        simestd@alaska.net
Cc:        freebsd-security@FreeBSD.ORG
Subject:   Re: SKIP problems
Message-ID:  <199805290119.VAA03275@adk.gr>

next in thread | raw e-mail | index | archive | help
-----BEGIN PGP SIGNED MESSAGE-----

To: simestd@alaska.net
Subject: Re: SKIP problems
Cc: freebsd-security@FreeBSD.ORG
Date: 05/28/98, 21:14:42


> SKIP is not an IETF standard and was rejected by the IPSEC wg several
> years ago (along with photuris))...

Touche'
Although the circumstances were different in the two situations.

> 1. freebsd/NRL/psu/me as found at http://www.cs.pdx.edu/research/SMN
> in case you have been asleep...    VPNs via route(8), route(4),
> and keyadmin(1).   I could try to briefly clarify on-line if

Sounds similar to the OpenBSD code, although I haven't seen the NRL
code (being a foreigner and all that).

> there was interest.  I suspect there are at least two or more
> IPSEC implementor (camps) that read this list.  Maybe we could
> all do that (or I could just go on vacation).

FYI, the linux-ipsec@clinet.fi mailing list (which was originally
intended to be the FreeSWAN list) has occasionally interesting
discussions, and at least 4 or 5 implementors are on it (that I know
of). 

> 2. the openBSD used to be netBSD implementation.  

...used to be BSD/OS :-)

> what about ISAKMP?

Not using it. No good free implementations available (yet). FreeSWAN
is working on it, maybe when they have something stable.

> what are the kernel interfaces?

PF_ENCAP (looks like a simplified PF_KEYv2)

> how do the kernel parts work?

http://www.cis.upenn.edu/~angelos/ipsec.ps.gz

>how do you add a new security transform?

ipsecadm (1) or photurisd (8) and possibly isakmp in the future

>how tested is the code? (how buggy?)

The first version of the code was written back in 1995. I can't claim
that there aren't any bugs left, but the code has been tested (and is
being tested) and used

>        is the code well written?

I'm probably biased, but I've heard from 3 people who have no
connection to the project that it's well written and tight.
It's at least reasonably good.

> what is the user (or sysadmin) api?

PF_ENCAP..expect a draft soon

>how does key management work?

Which part ?

>        is ASN involved :->

No!

> does it support user-level or only network level?

Supports both user-level and network level. It'll at some point be
able to also act as bump-in-the-wire

>policy for packets in/out in the o.s.; i.e., when to IPSEC
>        and when not?

Outgoing packets based on source/destination addresses (possibly
subnetted), transport protocol, UDP/TCP source/destination ports. 
Recently added per-socket policies with the automated keying.

> tunnel security attributes?

Yes.

> could joe average routing daemon use it?

Routing daemon ?! You could have your routing infrastructure point
everything at your IPsec firewall, if that's what you mean.

> multicast semantics?

Not completed (yet). We haven't focused on that, since there hasn't
been much demand on it (yet).

> how many tons of docs, if any?

Not many. Some man pages, a paper (URL above). A short article in the
OpenBSD Journal.

>you claim "interoperation", exactly what did that mean?
>        end to end apps
>        end to router tunnel
>        AH with transform Y
>        which AH acc. to which RFC/draft

End-to-end, firewall-to-firewall tunnel telnet, ping and ftp.
Have tested:

  old ESP DES and 3DES
  old AH MD5 and SHA1
  new ESP DES and 3DES, with MD5 and SHA1
  new AH MD5 and SHA1

I believe Rodney Thayer (rodney@sabletech.com) maintains a
sort-of-recent interoperability matrix, you can find OpenBSD there.
The most recent tests were last September at the ANX Interop Workshop
in Ottawa (interoped with 2-3 implementors, I remember mentat.com and
I think IBM), and at SNDSS at the end of March with Dan McDonald (Sun
Microsystems -- Solaris implementation).

We also support RIPEMD-160 authentication and CAST128 and Blowfish
encryption.

>and of course, our favorite, export control aspects.

No export control, as the code was written, lives and is being
maintained outside the US.

Hope this is informational enough.
Cheers,
- -Angelos

-----BEGIN PGP SIGNATURE-----
Version: 2.6.3i
Charset: noconv
Comment: Processed by Mailcrypt 3.4, an Emacs/PGP interface

iQCVAwUBNW4MAr0pBjh2h1kFAQFBpwQAhAPJqPIW39D8efkBwsmd7RxqT5oarcVH
l1FJayHW2z7T9qqvOriklnMXOZ4E/m0bZzOgfBjj960sdZiJNKK29fBhxVeubuQJ
2ol26SAcGYtfAFuGOZMY6FMxCJhK9fxgM8NFOUkVcV2lvAx/jbUqgJz2SD6pHbZ3
tXmVyNAiNLM=
=HxLS
-----END PGP SIGNATURE-----

To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe security" in the body of the message



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199805290119.VAA03275>