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

Parent Directory Parent Directory | Revision Log Revision Log

Revision 1.10 - (show annotations) (download) (as text)
Sun Feb 7 10:39:52 2010 UTC (4 years, 5 months ago) by robbat2
Branch: MAIN
Changes since 1.9: +52 -25 lines
File MIME type: text/html
Fix items per mail <robbat2-20100204T024714-596935539Z@orbis-terrarum.net>.

1 <?xml version="1.0" encoding="utf-8" ?>
2 <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
3 <html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en" lang="en">
5 <head>
6 <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
7 <meta name="generator" content="Docutils 0.6: http://docutils.sourceforge.net/" />
8 <title>GLEP 59 -- Manifest2 hash policies and security implications</title>
9 <link rel="stylesheet" href="tools/glep.css" type="text/css" /></head>
10 <body bgcolor="white">
11 <table class="navigation" cellpadding="0" cellspacing="0"
12 width="100%" border="0">
13 <tr><td class="navicon" width="150" height="35">
14 <a href="http://www.gentoo.org/" title="Gentoo Linux Home Page">
15 <img src="http://www.gentoo.org/images/gentoo-new.gif" alt="[Gentoo]"
16 border="0" width="150" height="35" /></a></td>
17 <td class="textlinks" align="left">
18 [<b><a href="http://www.gentoo.org/">Gentoo Linux Home</a></b>]
19 [<b><a href="http://www.gentoo.org/proj/en/glep">GLEP Index</a></b>]
20 [<b><a href="http://www.gentoo.org/proj/en/glep/glep-0059.txt">GLEP Source</a></b>]
21 </td></tr></table>
22 <table class="rfc2822 docutils field-list" frame="void" rules="none">
23 <col class="field-name" />
24 <col class="field-body" />
25 <tbody valign="top">
26 <tr class="field"><th class="field-name">GLEP:</th><td class="field-body">59</td>
27 </tr>
28 <tr class="field"><th class="field-name">Title:</th><td class="field-body">Manifest2 hash policies and security implications</td>
29 </tr>
30 <tr class="field"><th class="field-name">Version:</th><td class="field-body">1.7</td>
31 </tr>
32 <tr class="field"><th class="field-name">Last-Modified:</th><td class="field-body"><a class="reference external" href="http://www.gentoo.org/cgi-bin/viewcvs.cgi/xml/htdocs/proj/en/glep/glep-0059.txt?cvsroot=gentoo">2010/02/02 05:49:27</a></td>
33 </tr>
34 <tr class="field"><th class="field-name">Author:</th><td class="field-body">Robin Hugh Johnson &lt;robbat2&#32;&#97;t&#32;gentoo.org&gt;,</td>
35 </tr>
36 <tr class="field"><th class="field-name">Status:</th><td class="field-body">Draft</td>
37 </tr>
38 <tr class="field"><th class="field-name">Type:</th><td class="field-body">Standards Track</td>
39 </tr>
40 <tr class="field"><th class="field-name">Content-Type:</th><td class="field-body"><a class="reference external" href="glep-0002.html">text/x-rst</a></td>
41 </tr>
42 <tr class="field"><th class="field-name">Requires:</th><td class="field-body"><a class="reference external" href="http://www.gentoo.org/proj/en/glepglep-0044.html">44</a></td>
43 </tr>
44 <tr class="field"><th class="field-name">Created:</th><td class="field-body">October 2006</td>
45 </tr>
46 <tr class="field"><th class="field-name">Updated:</th><td class="field-body">November 2007, June 2008, July 2008, October 2008, January 2010</td>
47 </tr>
48 <tr class="field"><th class="field-name">Updates:</th><td class="field-body">44</td>
49 </tr>
50 <tr class="field"><th class="field-name">Post-History:</th><td class="field-body">December 2009, January 2010</td>
51 </tr>
52 </tbody>
53 </table>
54 <hr />
55 <div class="contents topic" id="contents">
56 <p class="topic-title first">Contents</p>
57 <ul class="simple">
58 <li><a class="reference internal" href="#abstract" id="id1">Abstract</a></li>
59 <li><a class="reference internal" href="#motivation" id="id2">Motivation</a></li>
60 <li><a class="reference internal" href="#specification" id="id3">Specification</a><ul>
61 <li><a class="reference internal" href="#the-bad-news" id="id4">The bad news</a></li>
62 <li><a class="reference internal" href="#how-fast-can-md5-be-broken" id="id5">How fast can MD5 be broken?</a></li>
63 <li><a class="reference internal" href="#the-good-news" id="id6">The good news</a></li>
64 <li><a class="reference internal" href="#what-should-be-done" id="id7">What should be done</a></li>
65 <li><a class="reference internal" href="#checksum-depreciation-timing" id="id8">Checksum depreciation timing</a><ul>
66 <li><a class="reference internal" href="#general-principles" id="id9">General principles:</a></li>
67 <li><a class="reference internal" href="#immediate-plans" id="id10">Immediate plans:</a></li>
68 </ul>
69 </li>
70 </ul>
71 </li>
72 <li><a class="reference internal" href="#backwards-compatibility" id="id11">Backwards Compatibility</a></li>
73 <li><a class="reference internal" href="#references" id="id12">References</a></li>
74 <li><a class="reference internal" href="#thanks-to" id="id13">Thanks to</a></li>
75 <li><a class="reference internal" href="#copyright" id="id14">Copyright</a></li>
76 </ul>
77 </div>
78 <div class="section" id="abstract">
79 <h1><a class="toc-backref" href="#id1">Abstract</a></h1>
80 <p>While Manifest2 format allows multiple hashes, the question of which
81 checksums should be present, why, and the security implications of such
82 have never been resolved. This GLEP covers all of these issues, and
83 makes recommendations as to how to handle checksums both now, and in
84 future.</p>
85 </div>
86 <div class="section" id="motivation">
87 <h1><a class="toc-backref" href="#id2">Motivation</a></h1>
88 <p>This GLEP is being written as part of the work on signing the Portage
89 tree, but is only tangentially related to the actual signing of
90 Manifests. Checksums present one possible weak point in the overall
91 security of the tree - and a comprehensive security plan is needed.</p>
92 <p>This GLEP is not mandatory for the tree-signing specification, but
93 instead aims to improve the security of the hashes used in Manifest2.
94 As such, it is also able to stand on it's own.</p>
95 </div>
96 <div class="section" id="specification">
97 <h1><a class="toc-backref" href="#id3">Specification</a></h1>
98 <div class="section" id="the-bad-news">
99 <h2><a class="toc-backref" href="#id4">The bad news</a></h2>
100 <p>First of all, I'd like to cover the bad news in checksum security.
101 A much discussed point, as been the simple question: What is the
102 security of multiple independent checksums on the same data?
103 The most common position (and indeed the one previously held by myself),
104 is that multiple checksums would be an increase in security, but we
105 could not provably quantify the amount of security this added.
106 The really bad news, is that this position is completely and utterly
107 wrong. Many of you will be aghast at this. There is extremely little
108 added security in multiple checksums as noted by Joux [J04]. For any set
109 of checksums, the actual strength lies in that of the strongest
110 checksum.</p>
111 <p>Wang et al [W04] extended Joux's [J04] work on SHA-0 to cover MD4, MD5,
112 HAVAL-128 and RIPEMD families of hashes.</p>
113 </div>
114 <div class="section" id="how-fast-can-md5-be-broken">
115 <h2><a class="toc-backref" href="#id5">How fast can MD5 be broken?</a></h2>
116 <p>For a general collision, not a pre-image attack, since the announcement
117 by Wang et al [W04], the time required to break MD5 has been massively
118 reduced. Originally at 1 hour on a near-supercomputer (IBM P690) and
119 estimated at 64 hours with a Pentium-3 1.7Ghz. This has gone down to
120 less than in two years, to 17 seconds [K06a].</p>
121 <ul class="simple">
122 <li>08/2004 - 1 hour, IBM pSeries 690 (32x 1.7Ghz POWER4+) = 54.4 GHz-Hours</li>
123 <li>03/2005 - 8 hours, Pentium-M 1.6Ghz = 12.8 Ghz-Hours</li>
124 <li>11/2005 - 5 hours, Pentium-4 1.7Ghz = 8.5 Ghz-Hours</li>
125 <li>03/2006 - 1 minute, Pentium-4 3.2Ghz = .05 Ghz-Hours</li>
126 <li>04/2006 - 17 seconds, Pentium-4 3.2Ghz = .01 Ghz-Hours</li>
127 </ul>
128 <p>If we accept a factor of 800x as a sample of how much faster a checksum
129 may be broken over the course of 2 years (MD5 using the above data is
130 &gt;2000x), then existing checksums do not stand a significant chance of
131 survival in the future. We should thus accept that whatever checksums we
132 are using today, will be broken in the near future, and plan as best as
133 possible. (A brief review [H04] of the SHA1 attacks indicates an
134 improvement of ~600x in the same timespan).</p>
135 <p>And for those that claim implementation of these procedures is not yet
136 feasible, see [K06b] for an application that can produce two
137 self-extracting EXE files, with identical MD5s, and whatever payload you
138 want.</p>
139 </div>
140 <div class="section" id="the-good-news">
141 <h2><a class="toc-backref" href="#id6">The good news</a></h2>
142 <p>Of the checksums presently used by Manifest2 (SHA1, SHA256, RIPEMD160),
143 one stands close to being completely broken: SHA1; and another is
144 significantly weakened: RIPEMD160. The SHA2 series has suffered some
145 attacks, but still remains reasonably solid [G07],[K08].</p>
146 <p>To reduce the potential for future problems and any single checksum
147 break leading to a rapid decrease in security, we should incorporate the
148 strongest hash available from each family of checksums, and be prepared
149 to retire old checksums actively, unless there is a overriding reason to
150 keep a specific checksum, such as part of a migration plan.</p>
151 </div>
152 <div class="section" id="what-should-be-done">
153 <h2><a class="toc-backref" href="#id7">What should be done</a></h2>
154 <p>Portage should always try to verify all supported hashes that are
155 available in a Manifest2, starting with the strongest ones as maintained
156 by a preference list. Over time, the weaker checksums should be removed
157 from Manifest2 files, once all old Portage installations have had
158 sufficient time to upgrade. Stronger checksums shall be added as soon as
159 an implementation is available in Portage. Weak checksums may be removed
160 as long as the depreciation process is followed (see below).</p>
161 <p>As soon as feasible, we should add the SHA512 and WHIRLPOOL algorithms.
162 In future, as stream-based checksums are developed (in response to the
163 development by NIST [AHS]), they should be considered and used.</p>
164 <p>The SHA512 algorithm is available in Python 2.5, which has been a
165 dependency of Portage since approximately Portage</p>
166 <p>The WHIRLPOOL checksum is not available within the PyCrypto library or
167 hashlib that is part of Python 2.5, but there are multiple alternative
168 Python implementations available, ranging from pure Python to C-based
169 (python-mhash).</p>
170 <p>The existence unsupported hash is not considered to be a failure unless
171 no supported hashes are available for a given Manifest entry.</p>
172 </div>
173 <div class="section" id="checksum-depreciation-timing">
174 <h2><a class="toc-backref" href="#id8">Checksum depreciation timing</a></h2>
175 <div class="section" id="general-principles">
176 <h3><a class="toc-backref" href="#id9">General principles:</a></h3>
177 <p>A minimum set of depreciated checksums shall be maintained only to
178 support old package manager versions where needed by historically used
179 trees:</p>
180 <ul class="simple">
181 <li>New package manager versions should NOT use depreciated checksums in</li>
182 <li>New trees with that have never used the depreciated checksums may omit
183 them for reasons of size, but are still strongly suggested to include
184 them.</li>
185 <li>Removal of depreciated checksums shall happen after no less than 18
186 months or one major Portage version cycle, whichever is greater.</li>
187 </ul>
188 </div>
189 <div class="section" id="immediate-plans">
190 <h3><a class="toc-backref" href="#id10">Immediate plans:</a></h3>
191 <p>For the current Portage, both SHA1 and RIPEMD160 should be immediately
192 removed, as they present no advantages over the already present SHA256.
193 SHA256 cannot be replaced immediately with SHA512, as existing Portage
194 versions need at least one supported algorithm present (SHA256 support
195 was added in June 2006), so it must be retained for some while.</p>
196 <p>Immediately:</p>
197 <ul class="simple">
198 <li>Add WHIRLPOOL and SHA512.</li>
199 <li>Remove SHA1 and RIPEMD160.</li>
200 </ul>
201 <p>After the majority of Portage installations include SHA512 support:</p>
202 <ul class="simple">
203 <li>Remove SHA256.</li>
204 </ul>
205 </div>
206 </div>
207 </div>
208 <div class="section" id="backwards-compatibility">
209 <h1><a class="toc-backref" href="#id11">Backwards Compatibility</a></h1>
210 <p>Old versions of Portage may support and expect only specific checksums.
211 This is accounted for in the checksum depreciation discussion.</p>
212 <p>For maximum compatiability, we should only have to include each of the
213 old algorithms that we are officially still supporting, as well as the
214 new ones that we prefer.</p>
215 </div>
216 <div class="section" id="references">
217 <h1><a class="toc-backref" href="#id12">References</a></h1>
218 <dl class="docutils">
219 <dt>[AHS] NIST (2007). &quot;NIST's Plan for New Cryptographic Hash Functions&quot;,</dt>
220 <dd>(Advanced Hash Standard). <a class="reference external" href="http://csrc.nist.gov/pki/HashWorkshop/">http://csrc.nist.gov/pki/HashWorkshop/</a></dd>
221 <dt>[BOBO06] Boneh, D. and Boyen, X. (2006). &quot;On the Impossibility of</dt>
222 <dd>Efficiently Combining Collision Resistant Hash Functions&quot;; Proceedings
223 of CRYPTO 2006, Dwork, C. (Ed.); Lecture Notes in Computer Science
224 4117, pp. 570-583. Available online from:
225 <a class="reference external" href="http://crypto.stanford.edu/~dabo/abstracts/hashing.html">http://crypto.stanford.edu/~dabo/abstracts/hashing.html</a></dd>
226 <dt>[H04] Hawkes, P. and Paddon, M. and Rose, G. (2004). &quot;On Corrective</dt>
227 <dd>Patterns for the SHA-2 Family&quot;. CRYPTO 2004 Cryptology ePrint Archive,
228 Report 2004/204. Available online from:
229 <a class="reference external" href="http://eprint.iacr.org/2004/207.pdf">http://eprint.iacr.org/2004/207.pdf</a></dd>
230 <dt>[J04] Joux, Antoie. (2004). &quot;Multicollisions in Iterated Hash</dt>
231 <dd>Functions - Application to Cascaded Constructions;&quot; Proceedings of
232 CRYPTO 2004, Franklin, M. (Ed); Lecture Notes in Computer Science
233 3152, pp. 306-316. Available online from:
234 <a class="reference external" href="http://web.cecs.pdx.edu/~teshrim/spring06/papers/general-attacks/multi-joux.pdf">http://web.cecs.pdx.edu/~teshrim/spring06/papers/general-attacks/multi-joux.pdf</a></dd>
235 <dt>[K06a] Klima, V. (2006). &quot;Tunnels in Hash Functions: MD5 Collisions</dt>
236 <dd>Within a Minute&quot;. Cryptology ePrint Archive, Report 2006/105.
237 Available online from: <a class="reference external" href="http://eprint.iacr.org/2006/105.pdf">http://eprint.iacr.org/2006/105.pdf</a></dd>
238 <dt>[K06b] Klima, V. (2006). &quot;Note and links to high-speed MD5 collision</dt>
239 <dd>proof of concept tools&quot;. Available online from:
240 <a class="reference external" href="http://cryptography.hyperlink.cz/2006/trick.txt">http://cryptography.hyperlink.cz/2006/trick.txt</a></dd>
241 <dt>[K08] Klima, V. (2008). &quot;On Collisions of Hash Functions Turbo SHA-2&quot;.</dt>
242 <dd>Cryptology ePrint Archive, Report 2008/003. Available online from:
243 <a class="reference external" href="http://eprint.iacr.org/2008/003.pdf">http://eprint.iacr.org/2008/003.pdf</a></dd>
244 <dt>[G07] Gligoroski, D. and Knapskog, S.J. (2007). &quot;Turbo SHA-2&quot;.</dt>
245 <dd>Cryptology ePrint Archive, Report 2007/403. Available online from:
246 <a class="reference external" href="http://eprint.iacr.org/2007/403.pdf">http://eprint.iacr.org/2007/403.pdf</a></dd>
247 <dt>[W04] Wang, X. et al: &quot;Collisions for Hash Functions MD4, MD5,</dt>
248 <dd>HAVAL-128 and RIPEMD&quot;, rump session, CRYPTO 2004, Cryptology ePrint
249 Archive, Report 2004/199, first version (August 16, 2004), second
250 version (August 17, 2004). Available online from:
251 <a class="reference external" href="http://eprint.iacr.org/2004/199.pdf">http://eprint.iacr.org/2004/199.pdf</a></dd>
252 </dl>
253 </div>
254 <div class="section" id="thanks-to">
255 <h1><a class="toc-backref" href="#id13">Thanks to</a></h1>
256 <dl class="docutils">
257 <dt>I'd like to thank the following folks, in no specific order:</dt>
258 <dd><ul class="first last simple">
259 <li>Ciaran McCreesh (ciaranm) - for pointing out the Joux (2004) paper,
260 and also being stubborn enough in not accepting a partial solution.</li>
261 <li>Marius Mauch (genone), Zac Medico (zmedico) and Brian Harring
262 (ferringb): for being knowledgeable about the Portage Manifest2
263 codebase.</li>
264 </ul>
265 </dd>
266 </dl>
267 </div>
268 <div class="section" id="copyright">
269 <h1><a class="toc-backref" href="#id14">Copyright</a></h1>
270 <p>Copyright (c) 2006-2010 by Robin Hugh Johnson. This material may be
271 distributed only subject to the terms and conditions set forth in the
272 Open Publication License, v1.0.</p>
273 <p>vim: tw=72 ts=2 expandtab:</p>
274 </div>
276 </div>
277 <div class="footer">
278 <hr class="footer" />
279 <a class="reference external" href="glep-0059.txt">View document source</a>.
280 Generated on: 2010-02-07 10:37 UTC.
281 Generated by <a class="reference external" href="http://docutils.sourceforge.net/">Docutils</a> from <a class="reference external" href="http://docutils.sourceforge.net/rst.html">reStructuredText</a> source.
283 </div>
284 </body>
285 </html>

  ViewVC Help
Powered by ViewVC 1.1.20