3 # Monkeysphere Documentation #
8 * Monkeysphere relies on [GnuTLS](http://gnutls.org/) version 2.4.0 or later.
12 * [Initial specifications at CMRG](http://cmrg.fifthhorseman.net/wiki/OpenPGPandSSH)
13 * [OpenPGP (RFC 4880)](http://tools.ietf.org/html/rfc4880)
14 * [Secure Shell Authentication Protocol (RFC 4252)](http://tools.ietf.org/html/rfc4252)
15 * [URI scheme for SSH, RFC draft](http://tools.ietf.org/wg/secsh/draft-ietf-secsh-scp-sftp-ssh-uri/)
17 ## Similar Projects ##
19 The monkeysphere isn't the only project intending to implement a PKI
20 for OpenSSH. We provide links to these other projects because they're
21 interesting, though we have concerns with their approaches.
23 All of the other projects we've found so far require a patched version
24 of OpenSSH, which makes adoption more difficult. Most people don't
25 build their own software, and simply overlaying a patched binary is
26 associated with significant maintenance (and therefore security)
27 problems. A PKI becomes more useful the more people participate in
28 it, so widespread adoption is important.
32 [openssh-gpg](http://www.red-bean.com/~nemo/openssh-gpg/) is a patch
33 against OpenSSH to support OpenPGP certificates. According to its
34 documentation, it is intended to support [`pgp-sign-rsa` and
35 `pgp-sign-dss` public key algorithms for hosts, as specified by the
36 IETF](http://tools.ietf.org/html/rfc4253#section-6.6).
38 Some concerns with `openssh-gpg`:
40 * This patch is significantly old; it doesn't appear to have been
41 maintained beyond OpenSSH 3.6p1. As of this writing, OpenSSH is on
44 * It only provides infrastructure in one direction: the user
45 authenticating the host by name. There doesn't seem to be a
46 mechanism for dealing with identifying users by name, or allowing
47 users to globally revoke or update keys.
49 * The choice of User ID (`anything goes here (and here!)
50 <ssh@foo.example.net>`) for host keys overlaps with the current use
51 of the User ID space. While it's unlikely that someone actually
52 uses this e-mail address in the web of trust, it would be a nasty
53 collision, as the holder of that key could impersonate the server
54 in question. The monkeysphere uses [User IDs of the form
55 `ssh://foo.example.net`](http://tools.ietf.org/wg/secsh/draft-ietf-secsh-scp-sftp-ssh-uri/)
56 to avoid collisions with existing use.
58 * It's not clear that `openssh-gpg` acknowledges or respects the
59 usage flags on the host keys.
61 * It requires patching OpenSSH.
64 ### Perspectives OpenSSH client ###
66 [The Perspectives project](http://www.cs.cmu.edu/~perspectives/) at
67 CMU has released an [openssh client that uses network
68 notaries](http://www.cs.cmu.edu/~perspectives/openssh.html) to bolster
69 your confidence in new keys. This offers a defense against a narrow
70 MITM attack (e.g. by someone who controls your local gateway) by
71 simply verifying that other machines from around the network see the
72 same keys for the remote host that you're seeing.
74 This is quite useful, but doesn't take the system as far as it could
75 go, and doesn't tie into the existing web of trust.
77 Some concerns with the Perspectives OpenSSH client:
79 * This client won't help if you are connecting to machines behind
80 firewalls, on NAT'ed LANs, with source IP filtering, or otherwise
81 in a restricted network state.
83 * There is still a question of why you should trust these particular
84 notaries during your verification. Who are the notaries? How
85 could they be compromised?
87 * It only provides infrastructure in one direction: the user
88 authenticating the host by name. There is no mechanism for dealing
89 with identifying users by name, or allowing users to globally
90 revoke or change keys.
92 * It requires patching OpenSSH
94 ### OpenSSH with X.509v3 certificates ###
96 Roumen Petrov [maintains a patch to OpenSSH that works with the X.509
97 PKI model](http://www.roumenpetrov.info/openssh/). This is the
98 certificate hierarchy commonly used by TLS (and SSL before that).
100 Some concerns about OpenSSH with X.509v3:
102 * the X.509 certificate specification itself [encourages corporate
103 consolidation and centralized global "trust" because of its
104 single-issuer architectural
105 limitation](http://lair.fifthhorseman.net/~dkg/tls-centralization/).
106 This results in an expensive and cumbersome system for smaller
107 players, and it also doesn't correspond to the true distributed
108 nature of human-to-human trust. Furthermore, centralized global
109 "trusted authorities" create a tempting target for attack, and a
110 single-point-of-failure if an attack is successful.
112 Depending on how you declare your trust relationships, OpenPGP is
113 capable of providing the same hierarchical structure as X.509, but
114 it is not limited to such a structure. The OpenPGP Web of Trust
115 model is more flexible and more adaptable to represent real-world
116 trust than X.509's rigid hierarchy.
118 * It requires patching OpenSSH.