/[gentoo]/xml/htdocs/proj/en/glep/glep-0059.txt
Gentoo

Contents of /xml/htdocs/proj/en/glep/glep-0059.txt

Parent Directory Parent Directory | Revision Log Revision Log


Revision 1.2 - (hide annotations) (download)
Wed Oct 22 17:59:43 2008 UTC (5 years, 8 months ago) by robbat2
Branch: MAIN
Changes since 1.1: +6 -6 lines
File MIME type: text/plain
Fix RST validation for tree-signing GLEPS.

1 cardoe 1.1 GLEP: 59
2     Title: Manifest2 hash policies and security implications
3 robbat2 1.2 Version: $Revision: 1.1 $
4     Last-Modified: $Date: 2008/10/21 23:30:47 $
5 cardoe 1.1 Author: Robin Hugh Johnson <robbat2@gentoo.org>,
6     Status: Draft
7     Type: Standards Track
8     Content-Type: text/x-rst
9     Requires: 44
10     Created: October 2006
11     Updated: November 2007, June 2008, July 2008
12     Updates: 44
13    
14     Abstract
15     ========
16     While Manifest2 format allows multiple hashes, the question of which
17     checksums should be present, why, and the security implications of such
18     have never been resolved. This GLEP covers all of these issues, and
19     makes recommendations as to how to handle checksums both now, and in
20     future.
21    
22     Motivation
23     ==========
24     This GLEP is being written as part of the work on signing the Portage
25     tree, but is only tangentially related to the actual signing of
26     Manifests. Checksums present one possible weak point in the overall
27     security of the tree - and a comprehensive security plan is needed.
28    
29     Specification
30     =============
31     The bad news
32     ------------
33     First of all, I'd like to cover the bad news in checksum security.
34     A much discussed point, as been the simple question: What is the
35     security of multiple independent checksums on the same data?
36     The most common position (and indeed the one previously held by myself),
37     is that multiple checksums would be an increase in security, but we
38     could not provably quantify the amount of security this added.
39     The really bad news, is that this position is completely and utterly
40     wrong. Many of you will be aghast at this. There is extremely little
41     added security in multiple checksums [J04]. For any set of checksums,
42     the actual strength lies in that of the strongest checksum.
43    
44     How fast can MD5 be broken?
45     ---------------------------
46     For a general collision, not a pre-image attack, since the original
47     announcement by Wang et al [W04], the time required to break MD5 has
48     been massively reduced. Originally at 1 hour on a near-supercomputer
49     (IBM P690) and estimated at 64 hours with a Pentium-3 1.7Ghz. This has
50     gone down to less than in two years, to 17 seconds [K06a]!
51    
52     08/2004 - 1 hour, IBM pSeries 690 (32x 1.7Ghz POWER4+) = 54.4 GHz-Hours
53     03/2005 - 8 hours, Pentium-M 1.6Ghz = 12.8 Ghz-Hours
54     11/2005 - 5 hours, Pentium-4 1.7Ghz = 8.5 Ghz-Hours
55     03/2006 - 1 minute, Pentium-4 3.2Ghz = .05 Ghz-Hours
56     04/2006 - 17 seconds, Pentium-4 3.2Ghz = .01 Ghz-Hours
57    
58     If we accept a factor of 800x as a sample of how much faster a checksum
59     may be broken over the course of 2 years (MD5 using the above data is
60     >2000x), then existing checksums do not stand a significant chance of
61     survival in the future. We should thus accept that whatever checksums we
62     are using today, will be broken in the near future, and plan as best as
63     possible. (A brief review [H04] of the present SHA1 attacks indicates an
64     improvement of ~600x in the same timespan).
65    
66     And for those that claim implementation of these procedures is not yet
67     feasible, see [K06b] for an application that can produce two
68     self-extracting .exe files, with identical MD5s, and whatever payload
69     you want.
70    
71     The good news
72     -------------
73     Of the checksums presently used by Manifest2, one stands close to being
74     completely broken: SHA1. The SHA2 series has suffered some attacks, but
75     still remains reasonably solid [G07],[K08]. No attacks against RIPEMD160
76     have been published, however it is constructed in the same manner as
77     MD5, SHA1 and SHA2, so is also vulnerable to the new methods of
78     cryptanalysis [H04].
79    
80     To reduce the potential for future problems and any single checksum
81     break leading to a rapid decrease in security, we should incorporate the
82     strongest hash available from each family of checksums, and be prepared
83     to retire old checksums actively, unless there is a overriding reason to
84     keep a specific checksum.
85    
86     What should be done
87     -------------------
88     Portage should always try to verify all supported hashes that are
89     available in a Manifest2, starting with the strongest ones as maintained
90     by a preference list. Over time, the weaker checksums should be removed
91     from Manifest2 files, once all old Portage installations have had
92     sufficient time to upgrade. We should be prepared to add stronger
93     checksums wherever possible, and to remove those that have been
94     defeated.
95    
96     An unsupported hash is not considered to be a failure unless no
97     supported hashes are available.
98    
99     Checksum depreciation
100     ~~~~~~~~~~~~~~~~~~~~~
101     For the current Portage, SHA1 should be gradually removed, as presents
102     no advantages over SHA256. Beyond one specific problem (see the next
103     paragraph), we should add SHA512 (SHA2, 512 bit size), the Whirlpool
104     checksum (standardized checksum, with no known weaknesses). In future,
105     as stream-based checksums are developed (in response to the development
106     by NIST [AHS]), they should be considered and used.
107    
108     There is one temporary stumbling block at hand - the existing Portage
109     infrastructure does not support SHA384/512 or Whirlpool, thus hampering
110     their immediate acceptance. SHA512 is available in Python 2.5, while
111     SHA1 is already available in Python 2.4. After Python2.5 is established
112     in a Gentoo media release, that would be a suitable time to remove SHA1
113     from Manifest2 files.
114    
115     Backwards Compatibility
116     =======================
117     Old versions of Portage may support and expect only specific checksums.
118     This is accounted for in the checksum depreciation discussion.
119    
120     References
121     ==========
122    
123     [AHS] NIST (2007). "NIST's Plan for New Cryptographic Hash Functions",
124     (Advanced Hash Standard). http://csrc.nist.gov/pki/HashWorkshop/
125    
126     [BOBO06] Boneh, D. and Boyen, X. (2006). "On the Impossibility of
127     Efficiently Combining Collision Resistant Hash Functions"; Proceedings
128     of CRYPTO 2006, Dwork, C. (Ed.); Lecture Notes in Computer Science
129     4117, pp. 570-583. Available online from:
130     http://crypto.stanford.edu/~dabo/abstracts/hashing.html
131    
132     [H04] Hawkes, P. and Paddon, M. and Rose, G. (2004). "On Corrective
133     Patterns for the SHA-2 Family". CRYPTO 2004 Cryptology ePrint Archive,
134     Report 2004/204. Available online from:
135     http://eprint.iacr.org/2004/207.pdf
136    
137 robbat2 1.2 [J04] Joux, Antoie. (2004). "Multicollisions in Iterated Hash
138     Functions - Application to Cascaded Constructions;" Proceedings of
139     CRYPTO 2004, Franklin, M. (Ed); Lecture Notes in Computer Science
140     3152, pp. 306-316. Available online from:
141 cardoe 1.1 http://web.cecs.pdx.edu/~teshrim/spring06/papers/general-attacks/multi-joux.pdf
142    
143     [K06a] Klima, V. (2006). "Tunnels in Hash Functions: MD5 Collisions
144     Within a Minute". Cryptology ePrint Archive, Report 2006/105.
145     Available online from: http://eprint.iacr.org/2006/105.pdf
146    
147     [K06b] Klima, V. (2006). "Note and links to high-speed MD5 collision
148     proof of concept tools". Available online from:
149     http://cryptography.hyperlink.cz/2006/trick.txt
150    
151     [K08] Klima, V. (2008). "On Collisions of Hash Functions Turbo SHA-2".
152     Cryptology ePrint Archive, Report 2008/003. Available online from:
153     http://eprint.iacr.org/2008/003.pdf
154    
155     [G07] Gligoroski, D. and Knapskog, S.J. (2007). "Turbo SHA-2".
156     Cryptology ePrint Archive, Report 2007/403. Available online from:
157     http://eprint.iacr.org/2007/403.pdf
158    
159     [W04] Wang, X. et al: "Collisions for Hash Functions MD4, MD5,
160     HAVAL-128 and RIPEMD", rump session, CRYPTO 2004, Cryptology ePrint
161     Archive, Report 2004/199, first version (August 16, 2004), second
162     version (August 17, 2004). Available online from:
163     http://eprint.iacr.org/2004/199.pdf
164    
165     Thanks to
166     =========
167     I'd like to thank the following folks, in no specific order:
168     - Ciaran McCreesh (ciaranm) - for pointing out the Joux (2004) paper,
169     and also being stubborn enough in not accepting a partial solution.
170     - Marius Mauch (genone), Zac Medico (zmedico) and Brian Harring
171     (ferringb): for being knowledgeable about the Portage Manifest2
172     codebase.
173    
174     Copyright
175     =========
176     Copyright (c) 2006 by Robin Hugh Johnson. This material may be
177     distributed only subject to the terms and conditions set forth in the
178     Open Publication License, v1.0.
179    
180     vim: tw=72 ts=2 expandtab:

  ViewVC Help
Powered by ViewVC 1.1.20