|
|||||||||||
|
[Design] Raw doc in the policy group files
From: Hugh Daniel <hugh(at)road.toad.com>
Date: Sun Feb 16 2003 - 10:03:13 EST
Is there any reason to put a default "private-or-clear"? The way it is now if someone messes with it 'bad' things can happen. I would suggest putting it in a comment with some supporting text. Obviously I am assuming that if this file is not there (and a comments only is supposed to be the same as not there at all) we get the default of 0.0.0.0/0. The per-file header doc shoud mention which one of the policies is "OE", "iOE" and "pOE" and (quickly) explain the terms. On to the policygroups.html documentation file.
The first paragraph in there is this:
Policy groups are a new feature that we're getting ready for 2.0.
Testing is appreciated, as is feedback via design@lists.freeswan.org.
Say what?:
Our Base Policy Groups rely on Opportunistic Encryption (OE) to do
this.
This page is reading like a group of scattered notes, just sentences with out plan, scatter shot onto paper. It need to be sat down with and the whole thing explained in the first four or five paragraphs. Lets continue with minutia for a bit:
No, this should be all more along the lines of VPN or OE or nothing:
private (OE-based VPN)
Attempt to negotiate opportunistically. On failure, block.
The link in this paragraph:
All Policy Groups are bidirectional. This chart shows some technical
details. FreeS/WAN does not support one-way encryption, since it can
give users a false sense of security.
As above, I hope that this is not quite true:
Base Policy Groups rely on OE. To use the following examples, you must
first become OE-capable, as described here.
Rather I would say it's as easy making a shopping list or two. Then
going into a bit more detail:
This is as easy as putting names, IPs or IP ranges in
/etc/ipsec.d/policies/[groupname].
Right, here is the part that says to comment out the 0.0.0.0/0 to turn OE off. In the case of random file corruption (say sysadmins had one too many bheers...) lets have it fail to secure (OE) rather then fail to clear. We want the built in default to be OE and you have to EXPRESSLY turn OE off, not just corrupt the private-or-clear file. Thats enough for now, I hope you get the idea of what I am looking for, both in documentation and in secure robustness. ||ugh Daniel Testing Fool hugh@freeswan.org -----BEGIN PGP SIGNATURE----- Version: 2.6.3ia Charset: noconv Comment: For the matching public key, finger the Reply-To: address.
iQCVAwUBPk+oLVZpdJR7FBQRAQGyBgP/a+/zGYRXfop14NgOe7QQvJ81JbC5MMvJ
1BMLIJKaEvr2q/q88ZE7P7jH7bCLcbTtngs3S2PzDR3ZFcFnW0fmqy2FDxuX4WNs
3oynGWc/oKw2gYW+cpE17KVpKUhthZEJh41sTkx2NsrdQs8syjNr8PFysN3zZLZd
vzpbnExheLA=
Design mailing list Design@lists.freeswan.org http://lists.freeswan.org/mailman/listinfo/design Received on Sun Feb 16 10:42:51 2003 This archive was generated by hypermail 2.1.8 : Wed Aug 23 2006 - 12:59:32 EDT |
||||||||||
|
|||||||||||