tor-webwml.git

1) ## translation metadata

2) # Revision: $Revision$

3) # Translation-Priority: 4-optional
4) 
5) #include "head.wmi" TITLE="Tor: Volunteer" CHARSET="UTF-8"
6) <div id="content" class="clearfix">
7)   <div id="breadcrumbs">

8)     <a href="<page index>">Home &raquo; </a>

9)     <a href="<page getinvolved/volunteer>">Volunteer</a>
10)   </div>
11)   <div id="maincol"> 
12)     <!-- PUT CONTENT AFTER THIS TAG -->
13)     <h1>A few things everyone can do now:</h1>
14)     <ol>
15)     <li>Please consider <a href="<page docs/tor-doc-relay>">running
16)     a relay</a> to help the Tor network grow.</li>
17)     <li>Tell your friends! Get them to run relays. Get them to run hidden
18)     services. Get them to tell their friends.</li>
19)     <li>If you like Tor's goals, please <a href="<page donate/donate>">take a moment
20)     to donate to support further Tor development</a>. We're also looking
21)     for more sponsors &mdash; if you know any companies, NGOs, agencies,
22)     or other organizations that want anonymity / privacy / communications
23)     security, let them know about us.</li>
24)     <li>We're looking for more <a href="<page about/torusers>">good examples of Tor
25)     users and Tor use cases</a>. If you use Tor for a scenario or purpose not
26)     yet described on that page, and you're comfortable sharing it with us,
27)     we'd love to hear from you.</li>
28)     </ol>
29)     
30)     <p>Tor has <a href="<page getinvolved/open-positions>">two open positions</a>.
31)     Please <a href="<page about/contact>">contact us</a> if you are qualified!</p>
32)     
33)     <a id="Documentation"></a>
34)     <h2><a class="anchor" href="#Documentation">Documentation</a></h2>
35)     <ol>
36)     <li>Help translate the web page and documentation into other
37)     languages. See the <a href="<page getinvolved/translation>">translation
38)     guidelines</a> if you want to help out. We especially need Arabic or
39)     Farsi translations, for the many Tor users in censored areas.</li>
40)     <li>Evaluate and document
41)     <a href="https://trac.torproject.org/projects/tor/wiki/TheOnionRouter/TorifyHOWTO">our
42)     list of programs</a> that can be configured to use Tor.</li>
43)     <li>We have a huge list of <a
44)     href="https://trac.torproject.org/projects/tor/wiki/TheOnionRouter/SupportPrograms">potentially useful
45)     programs that interface to Tor</a>. Which ones are useful in which
46)     situations? Please help us test them out and document your results.</li>
47)     </ol>
48)     
49)     <a id="Advocacy"></a>
50)     <h2><a class="anchor" href="#Advocacy">Advocacy</a></h2>
51)     <ol>
52)     <li>Create a <a href="https://trac.torproject.org/projects/tor/wiki/CommunityLogos">community logo</a> under a Creative Commons license that all can use and modify</li>
53)     <li>Create a presentation that can be used for various user group meetings around the world</li>

54)     <li>Create a video about the positive uses of Tor, what Tor is,
55)     or how to use it.  Some have already started on <a
56)     href="http://media.torproject.org/video/">Tor's Media server</a>,
57)     <a

58)     href="http://www.howcast.com/videos/90601-How-To-Circumvent-an-Internet-Proxy">Howcast</a>,
59)     and <a href="http://www.youtube.com/thetorproject">Youtube</a>.</li> 
60)     <li>Create a poster, or a set of posters, around a theme,
61)     such as "Tor for Freedom!"</li>
62)     </ol>
63)     
64)     <a id="Coding"></a>
65)     <a id="Summer"></a>
66)     <a id="Projects"></a>
67)     <h2><a class="anchor" href="#Projects">Good Coding Projects</a></h2>
68)     
69)     <p>
70)     You may find some of these projects to be good <a href="<page
71)     about/gsoc>">Google Summer of Code 2010</a> ideas. We have labelled each idea
72)     with how useful it would be to the overall Tor project (priority), how
73)     much work we expect it would be (effort level), how much clue you should
74)     start with (skill level), and which of our <a href="<page

75)     about/corepeople>">core developers</a> would be good mentors.

76)     If one or more of these ideas looks promising to you, please <a
77)     href="<page about/contact>">contact us</a> to discuss your plans rather than
78)     sending blind applications. You may also want to propose your own project
79)     idea &mdash; which often results in the best applications.
80)     </p>
81)     
82)     <ol>
83)     
84)     <li>

85)     <b>Audit Tor Browser Bundles for data leaks</b>

86)     <br>

87)     Priority: <i>High</i>

88)     <br>

89)     Effort Level: <i>High</i>

90)     <br>

91)     Skill Level: <i>Medium</i>

92)     <br>

93)     Likely Mentors: <i>Steven, Erinn, Jacob, Andrew</i>

94)     <p>The Tor Browser Bundle incorporates Tor, Firefox, Polipo, and the Vidalia

95)     user interface (and optionally the <a href="http://pidgin.im/">Pidgin</a>
96)     Instant Messaging client). Components are pre-configured to operate in a
97)     secure way, and it has very few dependencies on the installed operating
98)     system. It has therefore become one of the most easy to use, and popular,

99)     ways to use Tor on Windows.</p>
100)     <p>This project is to identify all of the traces left behind by
101)     using a Tor Browser Bundle on Windows, Mac OS X, or Linux.  Developing
102)     ways to stop, counter, or remove these traces is a final step.</p>
103)     <p>Students should be familiar with operating system analysis,
104)     application development on one or preferably Windows, Linux,
105)     and Mac OS X, and be comfortable with C/C++ and shell scripting.</p>
106)     <p>If you would like to help extend or do security auditing for
107)     TBB, please contact Erinn.</p>

108)     </li>
109)     
110)     <li>
111)     <b>Help track the overall Tor Network status</b>

112)     <br>

113)     Priority: <i>Medium to High</i>

114)     <br>

115)     Effort Level: <i>Medium</i>

116)     <br>

117)     Skill Level: <i>Medium</i>

118)     <br>

119)     Likely Mentors: <i>Karsten, Roger</i>

120)     <p>It would be great to set up an automated system for tracking network

121)     health over time, graphing it, etc. Part of this project would involve
122)     inventing better metrics for assessing network health and growth. Is the
123)     average uptime of the network increasing? How many relays are qualifying
124)     for Guard status this month compared to last month? What's the turnover
125)     in terms of new relays showing up and relays shutting off? Periodically
126)     people collect brief snapshots, but where it gets really interesting is

127)     when we start tracking data points over time.</p>
128)     <p>Data could be collected from the Tor Network Scanners in <a

129)     href="https://svn.torproject.org/svn/torflow/trunk/README">TorFlow</a>, from
130)     the server descriptors that each relay publishes, and from other
131)     sources. Results over time could be integrated into one of the <a
132)     href="https://torstatus.blutmagie.de/">Tor Status</a> web pages, or be
133)     kept separate. Speaking of the Tor Status pages, take a look at Roger's
134)     <a href="http://archives.seul.org/or/talk/Jan-2008/msg00300.html">Tor

135)     Status wish list</a>.</p>

136)     </li>
137)     
138)     <li>
139)     <b>Improving Tor's ability to resist censorship</b>

140)     <br>

141)     Priority: <i>Medium to High</i>

142)     <br>

143)     Effort Level: <i>Medium to High</i>

144)     <br>

145)     Skill Level: <i>High</i>

146)     <br>

147)     Likely Mentors: <i>Roger, Nick, Steven</i>

148)     <p>The Tor 0.2.1.x series makes <a

149)     href="<svnprojects>design-paper/blocking.html">significant
150)     improvements</a> in resisting national and organizational censorship.
151)     But Tor still needs better mechanisms for some parts of its

152)     anti-censorship design.</p>
153)     <p>One huge category of work is adding features to our <a

154)     href="http://gitweb.torproject.org//bridgedb.git?a=tree">bridgedb</a>
155)     service (Python). Tor aims to give out <a href="<page docs/bridges>">bridge
156)     relay addresses</a> to users that can't reach the Tor network
157)     directly, but there's an arms race between algorithms for distributing
158)     addresses and algorithms for gathering and blocking them. See <a

159)     href="<blog>bridge-distribution-strategies">our

160)     blog post on the topic</a> as an overview, and then look at <a
161)     href="http://archives.seul.org/or/dev/Dec-2009/msg00000.html">Roger's
162)     or-dev post</a> from December for more recent thoughts &mdash; lots of

163)     design work remains.</p>
164)     <p>If you want to get more into the guts of Tor itself (C), a more minor problem

165)     we should address is that current Tors can only listen on a single
166)     address/port combination at a time. There's
167)     <a href="<gitblob>doc/spec/proposals/118-multiple-orports.txt">a
168)     proposal to address this limitation</a> and allow clients to connect
169)     to any given Tor on multiple addresses and ports, but it needs more

170)     work.</p>
171)     <p>This project could involve a lot of research and design. One of the big

172)     challenges will be identifying and crafting approaches that can still
173)     resist an adversary even after the adversary knows the design, and

174)     then trading off censorship resistance with usability and
175)     robustness.</p>

176)     </li>
177)     
178)     <li>
179)     <b>Tor Controller Status Event Interface for Vidalia</b>

180)     <br>

181)     Priority: <i>Medium</i>

182)     <br>

183)     Effort Level: <i>Medium</i>

184)     <br>

185)     Skill Level: <i>Low to Medium</i>

186)     <br>

187)     Likely Mentors: <i>Matt</i>

188)     <p>There are a number of status changes inside Tor of which the user may need

189)     to be informed. For example, if the user is trying to set up his Tor as a
190)     relay and Tor decides that its ports are not reachable from outside
191)     the user's network, we should alert the user. Currently, all the user
192)     gets is a couple log messages in Vidalia's 'message log' window, which they
193)     likely never see since they don't receive a notification that something
194)     has gone wrong. Even if the user does actually look at the message log,

195)     most of the messages make little sense to the novice user.</p>
196)     <p>Tor has the ability to inform Vidalia of many such status changes, and

197)     we recently implemented support for a couple of these events. Still,
198)     there are many more status events the user should be informed of and we

199)     need a better UI for actually displaying them to the user.</p>
200)     <p>The goal of this project then is to design and implement a UI for

201)     displaying Tor status events to the user. For example, we might put a
202)     little badge on Vidalia's tray icon that alerts the user to new status
203)     events they should look at. Double-clicking the icon could bring up a
204)     dialog that summarizes recent status events in simple terms and maybe
205)     suggests a remedy for any negative events if they can be corrected by
206)     the user. Of course, this is just an example and one is free to

207)     suggest another approach.</p>
208)     <p>A person undertaking this project should have good UI design and layout

209)     and some C++ development experience. Previous experience with Qt and
210)     Qt's Designer will be very helpful, but are not required. Some
211)     English writing ability will also be useful, since this project will
212)     likely involve writing small amounts of help documentation that should
213)     be understandable by non-technical users. Bonus points for some graphic

214)     design/Photoshop fu, since we might want/need some shiny new icons too.</p>

215)     </li>
216)     
217)     <li>
218)     <b>Improve our unit testing process</b>

219)     <br>

220)     Priority: <i>Medium</i>

221)     <br>

222)     Effort Level: <i>Medium</i>

223)     <br>

224)     Skill Level: <i>Medium</i>

225)     <br>

226)     Likely Mentors: <i>Nick, Erinn</i>

227)     <p>Tor needs to be far more tested. This is a multi-part effort. To start

228)     with, our unit test coverage should rise substantially, especially in
229)     the areas outside the utility functions. This will require significant
230)     refactoring of some parts of Tor, in order to dissociate as much logic

231)     as possible from globals.</p>
232)     <p>Additionally, we need to automate our performance testing. We've got

233)     buildbot to automate our regular integration and compile testing already
234)     (though we need somebody to set it up on Windows),
235)     but we need to get our network simulation tests (as built in <a
236)     href="https://svn.torproject.org/svn/torflow/trunk/README">TorFlow</a>)
237)     updated for more recent versions of Tor, and designed to launch a test
238)     network either on a single machine, or across several, so we can test

239)     changes in performance on machines in different roles automatically.</p>

240)     </li>
241)     
242)     <li>
243)     <b>Help with independent Tor client implementations</b>

244)     <br>

245)     Priority: <i>Medium</i>

246)     <br>

247)     Effort Level: <i>High</i>

248)     <br>

249)     Skill Level: <i>Medium to High</i>

250)     <br>

251)     Likely Mentors: <i>Bruce, Nathan</i>

252)     <p>Others are currently working on Tor clients for Java, Android, and Maemo

253)     environments.  The first step is to get a handle on the current state of
254)     the project in which you are interested in helping; <a
255)     href="http://github.com/brl/JTor">Tor for Java</a>,
256)     <a href="https://svn.torproject.org/svn/projects/android/trunk/">Android/Orbot</a>
257)     , or <a href="<page docs/N900>">Tor for Maemo</a>. Check out the
258)     repository and familiarize yourself
259)     with the source code.  Further, support for requesting or even providing

260)     Tor hidden services would be neat, but not required.</p>
261)     <p>A prospective developer should be able to understand and write new Java

262)     code, including a Java cryptography API. Being able to read C code would be helpful,
263)     too. One should be willing to read the existing documentation,
264)     implement code based on it, and refine the documentation
265)     when things are underdocumented. This project is mostly about coding and

266)     to a small degree about design.</p>

267)     </li>

268) 

269)     <li>
270)     <b>More on Orbot &amp; Android OS-specific development</b>
271)     <br/>

272)     <br>

273)     Priority: <i>Medium</i>

274)     <br>

275)     Effort Level: <i>High</i>

276)     <br>

277)     Skill Level: <i>Medium to High</i>

278)     <br>

279)     Likely Mentors: <i>Nathan</i>

280)     <p><b>Android Java UI work:</b> Improved home screen to show better
281)     statistics about data transferred (up/down), number of circuits
282)     connected, quality of connection and so on. The "Tether Wifi"
283)     Android application is a good model to follow in how it shows
284)     a realtime count of bytes transferred as well as notifications
285)     when wifi client connect. In addition, better display/handling
286)     of Tor system/error messages would also be very helpful. Finally,
287)     the addition of a wizard or tutorial walkthrough for novice
288)     users to explain to them exactly what or what is not anonymized
289)     or protected would greatly improve the likelihood they will use
290)     Orbot correctly.</p>
291)     
292)     <p><b>Android Java OS/Core app work:</b> Better system-wide
293)     indicator either via the notification bar, "Toast" pop-up dialogs
294)     or some other indicator that an application's traffic is indeed
295)     moving through Orbot/Tor. For instance, right now you need to
296)     first go to a torcheck web service to ensure your browser is
297)     routing via Tor. Orbot should be able to notify you that circuits
298)     are being opened, used, etc. The aforementioned data transfer
299)     tracker might provide this type of awareness as well.</p>
300)     
301)     <p><b>Android Java Library/Community Outreach work:</b> We need
302)     to package a simple library for use with third-party application
303)     to easily enable them to support "Torification" on non-root
304)     devices (aka w/o transparent proxying). This library should
305)     include a wrapper for the Apache HTTPClient library, a utility
306)     class for detecting the state of Orbot connectivity, and other
307)     relevant/useful things an Android app might need to anonymize
308)     itself. This work would include the creation of the library,
309)     documentation, and sample code. Outreach or effort to implement
310)     the library within other open-source apps would follow.</p>
311)     
312)     <p><b>Android OS/C/Linux work:</b> The port of Tor to Android
313)     is basically a straight cross-compile to Linux ARM. There has
314)     been no work done in looking the optimization of Tor within a
315)     mobile hardware environment, on the ARM processor or other
316)     Android hardware, or on mobile networks. It should be noted,
317)     that even without optimization, Tor is handling the mobile
318)     network environment very well, automatically detecting change
319)     in IP addresses, reconnecting circuits, etc across switching
320)     from 2G to 3G to Wifi, and so forth.</p>

321)     </li>
322)     
323)     <li>
324)     <b>Simulator for slow Internet connections</b>

325)     <br>

326)     Priority: <i>Medium</i>

327)     <br>

328)     Effort Level: <i>Medium</i>

329)     <br>

330)     Skill Level: <i>Medium</i>

331)     <br>

332)     Likely Mentors: <i>Steven</i>

333)     <br>

334)     Many users of Tor have poor-quality Internet connections, giving low
335)     bandwidth, high latency, and high packet loss/re-ordering. User
336)     experience is that Tor reacts badly to these conditions, but it is
337)     difficult to improve the situation without being able to repeat the
338)     problems in the lab.

339)     <br>

340)     This project would be to build a simulation environment which
341)     replicates the poor connectivity so that the effect on Tor performance
342)     can be measured. Other components would be a testing utility to
343)     establish what are the properties of connections available, and to
344)     measure the effect of performance-improving modifications to Tor.

345)     <br>

346)     The tools used would be up to the student, but dummynet (for FreeBSD)
347)     and nistnet (for Linux) are two potential components on which this
348)     project could be built. Students should be experienced with network
349)     programming/debugging and TCP/IP, and preferably familiar with C and a
350)     scripting language.
351)     </li>
352)     
353)     <li>
354)     <b>An Improved and More Usable Network Map in Vidalia</b>

355)     <br>

356)     Priority: <i>Low to Medium</i>

357)     <br>

358)     Effort Level: <i>Medium</i>

359)     <br>

360)     Skill Level: <i>Medium</i>

361)     <br>

362)     Likely Mentors: <i>Matt</i>

363)     <br>

364)     One of Vidalia's existing features is a network map that shows the user
365)     the approximate geographic location of relays in the Tor network and
366)     plots the paths the user's traffic takes as it is tunneled through the
367)     Tor network. The map is currently not very interactive and has rather
368)     poor graphics. Instead, we implemented KDE's Marble widget such
369)     that it gives us a better quality map and enables improved interactivity,
370)     such as allowing the user to click on individual relays or circuits to
371)     display additional information. We want to add the ability
372)     for users to click on a particular relay or a country containing one or
373)     more Tor exit relays and say, "I want my connections to exit
374)     from here."

375)     <br>

376)     This project will first involve getting familiar with Vidalia
377)     and the Marble widget's API. One will then integrate the widget
378)     into Vidalia and customize Marble to be better suited for our application,
379)     such as making circuits clickable, storing cached map data in Vidalia's
380)     own data directory, and customizing some of the widget's dialogs.

381)     <br>

382)     A person undertaking this project should have good C++ development
383)     experience. Previous experience with Qt and CMake is helpful, but not
384)     required.
385)     </li>
386)     
387)     <li>
388)     <b>Torbutton equivalent for Thunderbird</b>

389)     <br>

390)     Priority: <i>Medium</i>

391)     <br>

392)     Effort Level: <i>High</i>

393)     <br>

394)     Skill Level: <i>High</i>

395)     <br>

396)     Likely Mentors: <i>Mike</i>

397)     <br>

398)     We're hearing from an increasing number of users that they want to use
399)     Thunderbird with Tor. However, there are plenty of application-level
400)     concerns, for example, by default Thunderbird will put your hostname in
401)     the outgoing mail that it sends. At some point we should start a new
402)     push to build a Thunderbird extension similar to Torbutton.
403)     </li>
404)     
405)     <li>
406)     <b>Improvements for Tor+Vidalia interaction on Linux/Unix platforms</b>

407)     <br>

408)     Priority: <i>Medium</i>

409)     <br>

410)     Effort Level: <i>Medium</i>

411)     <br>

412)     Skill Level: <i>Medium</i>

413)     <br>

414)     Likely Mentors: <i>Erinn, Peter</i>

415)     <br>

416)     Vidalia currently doesn't play nicely with Tor on Linux and Unix platforms.
417)     Currently, on Debian and Ubuntu, there is a configuration mechanism which
418)     allows Vidalia to override Tor's ability to start on boot (by sourcing
419)     <code>/etc/default/tor.vidalia</code> which sets <code>RUN_DAEMON=no</code> at the user's
420)     request), but full implementation of <a href="<gitblob>doc/spec/control-spec.txt">ControlPort</a> 
421)     communication is still required.

422)     <br>

423)     A better solution on Linux and Unix platforms would be to use Tor's
424)     ControlSocket, which allows Vidalia to talk to Tor via a Unix domain socket,
425)     and could possibly be enabled by default in Tor's distribution packages.
426)     Vidalia can then authenticate to Tor using filesystem-based (cookie)
427)     authentication if the user running Vidalia is also in the distribution-specific
428)     tor group.

429)     <br>

430)     This project will first involve adding support for Tor's ControlSocket to
431)     Vidalia. The student will then develop and test this support on various
432)     distributions to make sure it behaves in a predictable and consistent manner on
433)     all of them.

434)     <br>

435)     The next challenge would be to find an intuitive and usable way for Vidalia to be
436)     able to change Tor's configuration (torrc) even though it is located in
437)     <code>/etc/tor/torrc</code> and thus immutable. In Debian and Ubuntu we handle
438)     this with the aforementioned <code>/etc/default/tor.vidalia</code> but this
439)     functionality could (or should) be less distribution-specific. 

440)     <br>

441)     The best idea we've come up with so far is to feed Tor a new configuration via
442)     the ControlSocket when Vidalia starts, but that's bad because if the user is not
443)     using the latest Debian/Ubuntu packages, they may not have disabled Tor's
444)     ability to run on boot and will end up with a configuration that is different
445)     from what they want. The second best idea we've come up with is for Vidalia to
446)     write out a temporary torrc file and ask the user to manually move it to
447)     <code>/etc/tor/torrc</code>, but that's bad because users shouldn't have to
448)     mess with files directly.

449)     <br>

450)     A person undertaking this project should have prior knowledge of various Linux
451)     distributions and their packaging mechanisms as well as some C++ development
452)     experience. Previous experience with Qt is helpful, but not required.
453)     </li>
454)     
455)     
456)     <li>
457)     <b>Usability testing of Tor</b>

458)     <br>

459)     Priority: <i>Medium</i>

460)     <br>

461)     Effort Level: <i>Medium</i>

462)     <br>

463)     Skill Level: <i>Low to Medium</i>

464)     <br>

465)     Likely Mentors: <i>Andrew</i>

466)     <br>

467)     Especially the browser bundle, ideally amongst our target demographic.
468)     That would help a lot in knowing what needs to be done in terms of bug
469)     fixes or new features. We get this informally at the moment, but a more
470)     structured process would be better.

471)     </li>

472)     
473)     <li>
474)     <b>An authenticating IRC proxy</b>

475)     <br>

476)     Priority: <i>Low</i>

477)     <br>

478)     Effort Level: <i>Medium to High</i>

479)     <br>

480)     Skill Level: <i>Medium to High</i>

481)     <br>

482)     Likely Mentors: <i>Sebastian, Weasel, Roger</i>

483)     <br>

484)     The world needs an authenticating irc proxy. As we're periodically
485)     reminded from the Penny Arcade web comic, "Internet user + anonymity =
486)     jerk". With respect to websites we're actually doing ok, since websites
487)     can make their users log in and use other application-level authentication
488)     approaches. But IRC servers are much worse off, because most IRC server
489)     code is poorly written: hard to maintain, and harder to modify. Many
490)     IRC networks now block connections from Tor, and we're basically down to
491)     two holdouts (OFTC and Freenode). This state of affairs means that a lot
492)     of people around the world are thinking "I told you so" about anonymity
493)     online, when in fact the problem is simply lack of technology to make the
494)     problem manageable. We need some way to let the IRC networks distinguish
495)     which users have developed a reputation as not being jerks, so they can
496)     treat the two groups separately. There are some really cool research
497)     designs like <a href="http://www.cs.dartmouth.edu/~nymble/">Nymble</a>,
498)     which aim to let websites blacklist users without needing to learn who
499)     they are.  But Nymble is designed around web interactions. We need to
500)     build the glue around the IRC protocol that would let us plug in a project
501)     like Nymble (or a simpler one to start, as a proof-of-concept). One way
502)     to do that would be to build an IRC proxy that knows how to hear from
503)     IRC clients, knows how to talk to IRC servers, and has an additional
504)     layer that requires the users to authenticate.  Some work on this has
505)     begun by other volunteers, see their progress at <a
506)     href="http://github.com/anonirc/orc">http://github.com/anonirc/orc</a>.
507)     </li>
508)     
509)     <li>
510)     <b>Make torsocks/dsocks work on OS X</b>

511)     <br>

512)     Priority: <i>Medium</i>

513)     <br>

514)     Effort Level: <i>Medium</i>

515)     <br>

516)     Skill Level: <i>Medium</i>

517)     <br>

518)     Likely Mentors: <i>?</i>

519)     <br>

520)     <a href="http://code.google.com/p/torsocks/">Torsocks</a> and <a
521)     href="http://code.google.com/p/dsocks/">dsocks</a> are wrappers that will
522)     run applications, intercept their outgoing network connections, and push
523)     those connections through Tor. The goal is to handle applications that
524)     don't support proxies (or don't supporting them well). To get it right,
525)     they need to intercept many system calls. The syscalls you need to
526)     intercept on Linux differ dramatically from those on BSD. So Torsocks
527)     works fine on Linux, dsocks works ok on BSD (though it may be less
528)     maintained and thus might miss more syscalls), and nothing works well
529)     on both. First, we should patch dsocks to use Tor's <i>mapaddress</i>
530)     commands from the controller interface, so we don't waste a whole
531)     round-trip inside Tor doing the resolve before connecting. Second,
532)     we should make our <i>torify</i> script detect which of torsocks or
533)     dsocks is installed, and call them appropriately. This probably means
534)     unifying their interfaces, and might involve sharing code between them
535)     or discarding one entirely.
536)     </li>
537)     
538)     <li>
539)     <b>Bring up new ideas!</b>

540)     <br>

541)     Don't like any of these? Look at the <a
542)     href="<gitblob>doc/roadmaps/2008-12-19-roadmap-full.pdf">Tor development
543)     roadmap</a> for more ideas, or just try out Tor, Vidalia, and Torbutton,
544)     and find out what you think needs fixing.
545)     Some of the <a href="<gittree>doc/spec/proposals">current proposals</a>
546)     might also be short on developers.
547)     </li>
548)     
549)     </ol>
550)     
551)     <a id="OtherCoding"></a>
552)     <h2><a class="anchor" href="#OtherCoding">Other Coding and Design related ideas</a></h2>
553)     <ol>
554)     <li>Tor relays don't work well on Windows XP. On
555)     Windows, Tor uses the standard <tt>select()</tt> system
556)     call, which uses space in the non-page pool. This means
557)     that a medium sized Tor relay will empty the non-page pool, <a
558)     href="https://trac.torproject.org/projects/tor/wiki/TheOnionRouter/WindowsBufferProblems">causing
559)     havoc and system crashes</a>. We should probably be using overlapped IO
560)     instead. One solution would be to teach <a
561)     href="http://www.monkey.org/~provos/libevent/">libevent</a> how to use
562)     overlapped IO rather than select() on Windows, and then adapt Tor to
563)     the new libevent interface. Christian King made a
564)     <a href="https://svn.torproject.org/svn/libevent-urz/trunk/">good
565)     start</a> on this in the summer of 2007.</li>
566)     
567)     <li>We need to actually start building our <a href="<page
568)     docs/documentation>#DesignDoc">blocking-resistance design</a>. This involves
569)     fleshing out the design, modifying many different pieces of Tor, adapting
570)     <a href="<page projects/vidalia>">Vidalia</a> so it supports the
571)     new features, and planning for deployment.</li>
572)     
573)     <li>We need a flexible simulator framework for studying end-to-end
574)     traffic confirmation attacks. Many researchers have whipped up ad hoc
575)     simulators to support their intuition either that the attacks work
576)     really well or that some defense works great. Can we build a simulator
577)     that's clearly documented and open enough that everybody knows it's
578)     giving a reasonable answer? This will spur a lot of new research.
579)     See the entry <a href="#Research">below</a> on confirmation attacks for
580)     details on the research side of this task &mdash; who knows, when it's
581)     done maybe you can help write a paper or three also.</li>
582)     

583)     <li>Tor 0.1.1.x and later include support for hardware crypto
584)     accelerators via OpenSSL. It has been lightly tested and is
585)     possibly very buggy.  We're looking for more rigorous testing,
586)     performance analysis, and optimally, code fixes to openssl and
587)     Tor if needed.</li>

588)     
589)     <li>Perform a security analysis of Tor with <a
590)     href="http://en.wikipedia.org/wiki/Fuzz_testing">"fuzz"</a>. Determine
591)     if there are good fuzzing libraries out there for what we want. Win fame by
592)     getting credit when we put out a new release because of you!</li>
593)     
594)     <li>Tor uses TCP for transport and TLS for link
595)     encryption. This is nice and simple, but it means all cells
596)     on a link are delayed when a single packet gets dropped, and
597)     it means we can only reasonably support TCP streams. We have a <a
598)     href="https://trac.torproject.org/projects/tor/wiki/TheOnionRouter/TorFAQ#YoushouldtransportallIPpacketsnotjustTCPpackets.">list
599)     of reasons why we haven't shifted to UDP transport</a>, but it would
600)     be great to see that list get shorter. We also have a proposed <a
601)     href="<gitblob>doc/spec/proposals/100-tor-spec-udp.txt">specification
602)     for Tor and
603)     UDP</a> &mdash; please let us know what's wrong with it.</li>
604)     
605)     <li>We're not that far from having IPv6 support for destination addresses
606)     (at exit nodes). If you care strongly about IPv6, that's probably the
607)     first place to start.</li>
608)     
609)     <li>We need a way to generate the website diagrams (for example, the "How
610)     Tor Works" pictures on the <a href="<page about/overview>">overview page</a>
611)     from source, so we can translate them as UTF-8 text rather than edit
612)     them by hand with Gimp. We might want to
613)     integrate this as an wml file so translations are easy and images are
614)     generated in multiple languages whenever we build the website.</li>
615)     
616)     <li>How can we make the various LiveCD/USB systems easier
617)     to maintain, improve, and document?  One example is <a
618)     href="https://amnesia.boum.org/">The (Amnesic) Incognito Live
619)     System</a>.
620)     </li>
621)     
622)     <li>
623)     Another anti-censorship project is to try to make Tor
624)     more scanning-resistant.  Right now, an adversary can identify <a
625)     href="<gitblob>doc/spec/proposals/125-bridges.txt">Tor bridges</a>
626)     just by trying to connect to them, following the Tor protocol,
627)     and seeing if they respond.  To solve this, bridges could <a
628)     href="<svnprojects>design-paper/blocking.html#tth_sEc9.3">act like
629)     webservers</a> (HTTP or HTTPS) when contacted by port-scanning tools,
630)     and not act like bridges until the user provides a bridge-specific key.
631)     To start, check out Shane Pope's <a
632)     href="http://dl.dropbox.com/u/37735/index.html">thesis and prototype</a>.
633)     </li>
634)     
635)     </ol>
636)     
637)     <a id="Research"></a>
638)     <h2><a class="anchor" href="#Research">Research</a></h2>
639)     <ol>
640)     <li>The "end-to-end traffic confirmation attack":
641)     by watching traffic at Alice and at Bob, we can <a
642)     href="http://freehaven.net/anonbib/#danezis:pet2004">compare
643)     traffic signatures and become convinced that we're watching the same
644)     stream</a>. So far Tor accepts this as a fact of life and assumes this
645)     attack is trivial in all cases. First of all, is that actually true? How
646)     much traffic of what sort of distribution is needed before the adversary
647)     is confident he has won? Are there scenarios (e.g. not transmitting much)
648)     that slow down the attack? Do some traffic padding or traffic shaping
649)     schemes work better than others?</li>
650)     <li>A related question is: Does running a relay/bridge provide additional
651)     protection against these timing attacks? Can an external adversary that can't
652)     see inside TLS links still recognize individual streams reliably?
653)     Does the amount of traffic carried degrade this ability any? What if the
654)     client-relay deliberately delayed upstream relayed traffic to create a queue
655)     that could be used to mimic timings of client downstream traffic to make it
656)     look like it was also relayed? This same queue could also be used for masking
657)     timings in client upstream traffic with the techniques from <a
658)     href="http://www.freehaven.net/anonbib/#ShWa-Timing06">adaptive padding</a>,
659)     but without the need for additional traffic. Would such an interleaving of
660)     client upstream traffic obscure timings for external adversaries? Would the
661)     strategies need to be adjusted for asymmetric links? For example, on
662)     asymmetric links, is it actually possible to differentiate client traffic from
663)     natural bursts due to their asymmetric capacity? Or is it easier than
664)     symmetric links for some other reason?</li>
665)     <li>Repeat Murdoch and Danezis's <a
666)     href="http://www.cl.cam.ac.uk/~sjm217/projects/anon/#torta">attack from
667)     Oakland 05</a> on the current Tor network. See if you can learn why it
668)     works well on some nodes and not well on others. (My theory is that the
669)     fast nodes with spare capacity resist the attack better.) If that's true,
670)     then experiment with the RelayBandwidthRate and RelayBandwidthBurst
671)     options to run a relay that is used as a client while relaying the
672)     attacker's traffic: as we crank down the RelayBandwidthRate, does the
673)     attack get harder? What's the right ratio of RelayBandwidthRate to
674)     actually capacity? Or is it a ratio at all? While we're at it, does a
675)     much larger set of candidate relays increase the false positive rate
676)     or other complexity for the attack? (The Tor network is now almost two
677)     orders of magnitude larger than it was when they wrote their paper.) Be
678)     sure to read <a href="http://freehaven.net/anonbib/#clog-the-queue">Don't
679)     Clog the Queue</a> too.</li>
680)     <li>The "routing zones attack": most of the literature thinks of
681)     the network path between Alice and her entry node (and between the
682)     exit node and Bob) as a single link on some graph. In practice,
683)     though, the path traverses many autonomous systems (ASes), and <a
684)     href="http://freehaven.net/anonbib/#feamster:wpes2004">it's not uncommon
685)     that the same AS appears on both the entry path and the exit path</a>.
686)     Unfortunately, to accurately predict whether a given Alice, entry,
687)     exit, Bob quad will be dangerous, we need to download an entire Internet
688)     routing zone and perform expensive operations on it. Are there practical
689)     approximations, such as avoiding IP addresses in the same /8 network?</li>
690)     <li>Other research questions regarding geographic diversity consider
691)     the tradeoff between choosing an efficient circuit and choosing a random
692)     circuit. Look at Stephen Rollyson's <a
693)     href="http://swiki.cc.gatech.edu:8080/ugResearch/uploads/7/ImprovingTor.pdf">position
694)     paper</a> on how to discard particularly slow choices without hurting
695)     anonymity "too much". This line of reasoning needs more work and more
696)     thinking, but it looks very promising.</li>
697)     <li>Tor doesn't work very well when relays have asymmetric bandwidth
698)     (e.g. cable or DSL). Because Tor has separate TCP connections between
699)     each hop, if the incoming bytes are arriving just fine and the outgoing
700)     bytes are all getting dropped on the floor, the TCP push-back mechanisms
701)     don't really transmit this information back to the incoming streams.
702)     Perhaps Tor should detect when it's dropping a lot of outgoing packets,
703)     and rate-limit incoming streams to regulate this itself? I can imagine
704)     a build-up and drop-off scheme where we pick a conservative rate-limit,
705)     slowly increase it until we get lost packets, back off, repeat. We
706)     need somebody who's good with networks to simulate this and help design
707)     solutions; and/or we need to understand the extent of the performance
708)     degradation, and use this as motivation to reconsider UDP transport.</li>
709)     <li>A related topic is congestion control. Is our
710)     current design sufficient once we have heavy use? Maybe
711)     we should experiment with variable-sized windows rather
712)     than fixed-size windows? That seemed to go well in an <a
713)     href="http://www.psc.edu/networking/projects/hpn-ssh/theory.php">ssh
714)     throughput experiment</a>. We'll need to measure and tweak, and maybe
715)     overhaul if the results are good.</li>
716)     <li>Our censorship-resistance goals include preventing
717)     an attacker who's looking at Tor traffic on the wire from <a
718)     href="<svnprojects>design-paper/blocking.html#sec:network-fingerprint">distinguishing
719)     it from normal SSL traffic</a>. Obviously we can't achieve perfect
720)     steganography and still remain usable, but for a first step we'd like to
721)     block any attacks that can win by observing only a few packets. One of
722)     the remaining attacks we haven't examined much is that Tor cells are 512
723)     bytes, so the traffic on the wire may well be a multiple of 512 bytes.
724)     How much does the batching and overhead in TLS records blur this on the
725)     wire? Do different buffer flushing strategies in Tor affect this? Could
726)     a bit of padding help a lot, or is this an attack we must accept?</li>
727)     <li>Tor circuits are built one hop at a time, so in theory we have the
728)     ability to make some streams exit from the second hop, some from the
729)     third, and so on. This seems nice because it breaks up the set of exiting
730)     streams that a given relay can see. But if we want each stream to be safe,
731)     the "shortest" path should be at least 3 hops long by our current logic, so
732)     the rest will be even longer. We need to examine this performance / security
733)     tradeoff.</li>
734)     <li>It's not that hard to DoS Tor relays or directory authorities. Are client
735)     puzzles the right answer? What other practical approaches are there? Bonus
736)     if they're backward-compatible with the current Tor protocol.</li>
737)     <li>Programs like <a

738)     href="<page torbutton/index>">Torbutton</a> aim to hide