TODO 15.4 KB
Newer Older
1
$Id$
2
3
4
Legend:
SPEC!!  - Not specified
SPEC    - Spec not finalized
5
6
7
N       - nick claims
R       - arma claims
P       - phobos claims
8
9
10
11
12
13
14
        - Not done
        * Top priority
        . Partially done
        o Done
        D Deferred
        X Abandoned

Nick Mathewson's avatar
Nick Mathewson committed
15
16
    . <nickm> "Let's try to find a way to make it run and make the version
       match, but if not, let's just make it run."
17
18
    - <arma> "should we detect if we have a --with-ssl-dir and try the -R
      by default, if it works?"
19

Nick Mathewson's avatar
Nick Mathewson committed
20
Items for 0.1.2.x:
Roger Dingledine's avatar
Roger Dingledine committed
21
22
  - Bug: We are willing to build circuits using not-up-to-date descriptors,
    when we get a request and we've been idle a long time.
Roger Dingledine's avatar
Roger Dingledine committed
23
24
25
  - We should ship with a list of stable dir mirrors -- they're not
    trusted like the authorities, but they'll provide more robustness
    and diversity for bootstrapping clients.
Nick Mathewson's avatar
Nick Mathewson committed
26
27
28
29
30
31
  - Servers are easy to setup and run: being a relay is about as easy as
    being a client.
    - Reduce resource load
      - look into "uncounting" bytes spent on local connections. so
        we can bandwidthrate but still have fast downloads.
      - Write limiting; separate token bucket for write
32
      o dir answers include a your-ip-address-is header, so we can
Nick Mathewson's avatar
Nick Mathewson committed
33
34
35
        break our dependency on dyndns.
      - Count TLS bandwidth more accurately
      - Write-limit directory responses.
36
37
38
N   . Improve memory usage on tight-memory machines.
      - Directory-related fixes.
        o Remember offset and location of each descriptor in the cache/journal
39
        o When sending a big pile of descs to a client, don't shove them all
40
41
42
43
44
45
          on the buffer at once. Keep a list of the descriptor digests for
          the descriptors we still want to send.  We might end up truncating
          some replies by returning fewer descriptors than were requested (if
          somebody requests a desc that we throw away before we deliver it),
          but this happens only when somebody wants an obsolete desc, and
          clients can already handle truncated replies.
46
        o But what do we do about compression? That's the part that makes
47
          stuff hard.
48
49
          o Implement compress/decompress-on-the-fly support.
          o Use it for returning lists of descriptors.
50
          o Use it for returning lists of network status docs. (This will
51
            take a hybrid approach; let's get the other bits working first.)
52
53
          o Make clients handle missing Content-Length tags.  (Oh, they do.)
            o Verify that this has happened for a long time.
54
55
        o Try a similar trick for spooling out v1 directories.  These we
          _uncompress_ on the fly.
56
        - Look into pulling serverdescs off buffers as they arrive.
57
58
59
60
        . Mmap cache files where possible.
          o Mmap cached-routers file; when building it, go oldest-to-newest.
          - More unit tests and asserts for cached-routers file: ensure digest
            for the right router.  Verify dl by digest, fp, etc.
61
          . Make sure cached-routers values and offsets are correct in the
62
            presence of windows FS insanity.
63
64
          - Save and mmap v1 directories; store them zipped, not
            uncompressed.
65
66
          - Store networkstatus docs zipped, not uncompressed.  Maaaybe mmap
            them too.
67
        o Be a little more OO to save memory in frequently
Nick Mathewson's avatar
Nick Mathewson committed
68
          replicated structs.
69
          o Split circuit_t into origin circuits and or circuits
70
            o Move as many fields as reasonable out of base class.
71
72
            o Re-pack structs to avoid wasted bytes.
          o Split connection_t based on type field.
73

Nick Mathewson's avatar
Nick Mathewson committed
74
    - "bandwidth classes", for incoming vs initiated-here conns.
75
    o Asynchronous DNS
76
77
78
79
80
      . And test it
      . Make it work on windows.
        o Implement
        - Test
      - Make it the default on platforms where it works
81

Nick Mathewson's avatar
Nick Mathewson committed
82
83
84
85
86
  - Security improvements
    - Directory guards
R   - remember the last time we saw one of our entry guards labelled with
      the GUARD flag. If it's been too long, it is not suitable for use.
      If it's been really too long, remove it from the list.
Nick Mathewson's avatar
Nick Mathewson committed
87

Nick Mathewson's avatar
Nick Mathewson committed
88
  - Make reverse DNS work.
Nick Mathewson's avatar
Nick Mathewson committed
89

Nick Mathewson's avatar
Nick Mathewson committed
90
91
92
93
94
95
96
97
98
  - Performance improvements
    - Better estimates in the directory of whether servers have good uptime
      (high expected time to failure) or good guard qualities (high
      fractional uptime).
      - AKA Track uptime as %-of-time-up, as well as time-since-last-down.
    - Clients should prefer to avoid exit nodes for non-exit path positions.
      (bug 200)
    - Have a "Faster" status flag that means it. Fast2, Fast4, Fast8?

99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
  - A more efficient dir protocol.
    - Clients stop dumping old descriptors if the network-statuses
      claim they're still valid.
      - Later, servers will stop generating new descriptors simply
        because 18 hours have passed.
    - Authorities should fetch the network-statuses amongst each
      other, consensus them, and advertise a communal network-status.
      This is not so much for safety/complexity as it is to reduce
      bandwidth requirements for Alice.
      - How does this interact with our goal of being able to choose
        your own dir authorities? I guess we're now assuming that all
        dir authorities know all the other authorities in their "group"?
    - Should we also look into a "delta since last network-status
      checkpoint" scheme, to reduce overhead further?
    - Extend the "r" line in network-status to give a set of buckets (say,
      comma-separated) for that router.
      - Buckets are deterministic based on IP address.
      - Then clients can choose a bucket (or set of buckets) to
        download and use.

Nick Mathewson's avatar
Nick Mathewson committed
119
  - Critical but minor bugs, backport candiates.
120
121
    - If the client's clock is too far in the past, it will drop (or
      just not try to get) descriptors, so it'll never build circuits.
Nick Mathewson's avatar
Nick Mathewson committed
122
R   - Failed rend desc fetches sometimes don't get retried.
Roger Dingledine's avatar
Roger Dingledine committed
123
    . If we fail to connect via an exit enclave, (warn and) try again
Nick Mathewson's avatar
Nick Mathewson committed
124
      without demanding that exit node.
Roger Dingledine's avatar
Roger Dingledine committed
125
126
      - And recognize when extending to the enclave node is failing,
        so we can abandon then too.
Nick Mathewson's avatar
Nick Mathewson committed
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
R   - non-v1 authorities should not accept rend descs.
    - We need a separate list of "hidserv authorities" if we want to
      retire moria1 from the main list.
    - support dir 503s better
      o clients don't log as loudly when they receive them
      - they don't count toward the 3-strikes rule
        - should there be some threshold of 503's after which we give up?
      - think about how to split "router is down" from "dirport shouldn't
        be tried for a while"?
      - authorities should *never* 503 a cache, but *should* 503 clients
        when they feel like it.
      - update dir-spec with what we decided for each of these
    - provide no-cache no-index headers from the dirport?

  - Windows server usability
    - Solve the ENOBUFS problem.
      - make tor's use of openssl operate on buffers rather than sockets,
        so we can make use of libevent's buffer paradigm once it has one.
      - make tor's use of libevent tolerate either the socket or the
        buffer paradigm; includes unifying the functions in connect.c.
    - We need a getrlimit equivalent on Windows so we can reserve some
      file descriptors for saving files, etc. Otherwise we'll trigger
      asserts when we're out of file descriptors and crash.
M   - rewrite how libevent does select() on win32 so it's not so very slow.
      - Add overlapped IO
152

Nick Mathewson's avatar
Nick Mathewson committed
153
  - When we connect to a Tor server, it sends back a cell listing
154
155
    the IP it believes it is using. Use this to block dvorak's attack.
    Also, this is a fine time to say what time you think it is.
Nick Mathewson's avatar
Nick Mathewson committed
156
    o Verify that a new cell type is okay with deployed codebase
Roger Dingledine's avatar
Roger Dingledine committed
157
    . Specify HELLO cells
158
159
    - Implement

Nick Mathewson's avatar
Nick Mathewson committed
160
N - Exitlist should avoid outputting the same IP address twice.
Nick Mathewson's avatar
Nick Mathewson committed
161

Nick Mathewson's avatar
Nick Mathewson committed
162
N - Write path-spec.txt
163
164
  - Break the dir v1 stuff out of tor-spec.txt into dir-spec-v1.txt,
    and change the new dir-spec.txt to dir-spec-v2.txt.
Nick Mathewson's avatar
Nick Mathewson committed
165
166
167
168

  - Packaging
    - Tell people about OSX Uninstaller
    - Quietly document NT Service options
Nick Mathewson's avatar
Nick Mathewson committed
169

Nick Mathewson's avatar
Nick Mathewson committed
170
171
172
173
174
  - Docs
    - More prominently, we should have a recommended apps list.
      - recommend gaim.
      - unrecommend IE because of ftp:// bug.
    - torrc.complete.in needs attention?
175

Nick Mathewson's avatar
Nick Mathewson committed
176
Topics to think about during 0.1.2.x development:
Nick Mathewson's avatar
Nick Mathewson committed
177
178
179
180
181
  * Figure out incentives.
    - (How can we make this tolerant of a bad v0?)
  * Figure out non-clique.
  * Figure out China.
  - Figure out avoiding duplicate /24 lines
Nick Mathewson's avatar
Nick Mathewson committed
182
183
184
185
  - Figure out partial network knowledge.
  - Figure out hidden services.

Minor items for 0.1.2.x as time permits.
186
187
188
  - Rate limit exit connections to a given destination -- this helps
    us play nice with websites when Tor users want to crawl them; it
    also introduces DoS opportunities.
189
  - The bw_accounting file should get merged into the state file.
Nick Mathewson's avatar
Nick Mathewson committed
190
  - Streamline how we define a guard node as 'up'.
Nick Mathewson's avatar
Nick Mathewson committed
191
  - Better installers and build processes.
192
    - Commit edmanm's win32 makefile to tor cvs contrib, or write a new one.
Nick Mathewson's avatar
Nick Mathewson committed
193
  - Christian Grothoff's attack of infinite-length circuit.
Nick Mathewson's avatar
Nick Mathewson committed
194
195
196
197
    the solution is to have a separate 'extend-data' cell type
    which is used for the first N data cells, and only
    extend-data cells can be extend requests.
    - Specify, including thought about anonymity implications.
Nick Mathewson's avatar
Nick Mathewson committed
198
  - Display the reasons in 'destroy' and 'truncated' cells under some
Roger Dingledine's avatar
Roger Dingledine committed
199
    circumstances?
Nick Mathewson's avatar
Nick Mathewson committed
200
201
202
  - We need a way for the authorities to declare that nodes are
    in a family.  Also, it kinda sucks that family declarations use O(N^2)
    space in the descriptors.
203
204
205
  - If the server is spewing complaints about raising your ulimit -n,
    we should add a note about this to the server descriptor so other
    people can notice too.
206
207
  - rate limit the number of exit connections to a given destination, to
    help with DoS/crawling issues.
Roger Dingledine's avatar
Roger Dingledine committed
208
209
  - cpu fixes:
    - see if we should make use of truncate to retry
Nick Mathewson's avatar
Nick Mathewson committed
210
    - kill dns workers more slowly
211
  . Directory changes
Roger Dingledine's avatar
Roger Dingledine committed
212
    . Some back-out mechanism for auto-approval
Roger Dingledine's avatar
Roger Dingledine committed
213
      - a way of rolling back approvals to before a timestamp
214
        - Consider minion-like fingerprint file/log combination.
Roger Dingledine's avatar
Roger Dingledine committed
215
  - packaging and ui stuff:
216
217
    . multiple sample torrc files
    . figure out how to make nt service stuff work?
218
      . Document it.
Nick Mathewson's avatar
Nick Mathewson committed
219
    - Vet all pending installer patches
220
221
222
      - Win32 installer plus privoxy, sockscap/freecap, etc.
      - Vet win32 systray helper code

Nick Mathewson's avatar
Nick Mathewson committed
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
  - Improve controller
      - change circuit status events to give more details, like purpose,
        whether they're internal, when they become dirty, when they become
        too dirty for further circuits, etc.
        - What do we want here, exactly?
        - Specify and implement it.
      - Change stream status events analogously.
        - What do we want here, exactly?
        - Specify and implement it.
      - Make other events "better".
      - Change stream status events analogously.
        - What do we want here, exactly?
        - Specify and implement it.
      - Make other events "better" analogously
        - What do we want here, exactly?
        - Specify and implement it.
      . Expose more information via getinfo:
        - import and export rendezvous descriptors
        - Review all static fields for additional candidates
      - Allow EXTENDCIRCUIT to unknown server.
      - We need some way to adjust server status, and to tell tor not to
        download directories/network-status, and a way to force a download.
      - It would be nice to request address lookups from the controller
        without using SOCKS.
      - Make everything work with hidden services
Roger Dingledine's avatar
Roger Dingledine committed
248
249
250
251
252
  - Directory system improvements
    - config option to publish what ports you listen on, beyond
      ORPort/DirPort.  It should support ranges and bit prefixes (?) too.
      - Parse this.
      - Relay this in networkstatus.
253

254
Future version:
255
256
257
258
259
260
261
262
263
264
265
  . Update the hidden service stuff for the new dir approach.
    - switch to an ascii format, maybe sexpr?
    - authdirservers publish blobs of them.
    - other authdirservers fetch these blobs.
    - hidserv people have the option of not uploading their blobs.
    - you can insert a blob via the controller.
    - and there's some amount of backwards compatibility.
    - teach clients, intro points, and hidservs about auth mechanisms.
    - come up with a few more auth mechanisms.
  - auth mechanisms to let hidden service midpoint and responder filter
    connection requests.
266
267
268
269
270
271
  - Bind to random port when making outgoing connections to Tor servers,
    to reduce remote sniping attacks.
  - Have new people be in limbo and need to demonstrate usefulness
    before we approve them.
  - Clients should estimate their skew as median of skew from servers
    over last N seconds.
Nick Mathewson's avatar
Nick Mathewson committed
272
  - Make router_is_general_exit() a bit smarter once we're sure what it's for.
273
274
275
276
  - Audit everything to make sure rend and intro points are just as likely to
    be us as not.
  - Do something to prevent spurious EXTEND cells from making middleman
    nodes connect all over.  Rate-limit failed connections, perhaps?
Nick Mathewson's avatar
Nick Mathewson committed
277
  - Automatically determine what ports are reachable and start using
278
279
    those, if circuits aren't working and it's a pattern we recognize
    ("port 443 worked once and port 9001 keeps not working").
280
281
282
283
284
285
286
  - Limit to 2 dir, 2 OR, N SOCKS connections per IP.
  - Handle full buffers without totally borking
  - Rate-limit OR and directory connections overall and per-IP and
    maybe per subnet.
  - Hold-open-until-flushed now works by accident; it should work by
    design.
  - DoS protection: TLS puzzles, public key ops, bandwidth exhaustion.
287
    - Specify?
288
  - tor-resolve script should use socks5 to get better error messages.
Nick Mathewson's avatar
Nick Mathewson committed
289
290
  - hidserv offerers shouldn't need to define a SocksPort
    * figure out what breaks for this, and do it.
291
292
  - tor should be able to have a pool of outgoing IP addresses
    that it is able to rotate through. (maybe)
293
294
  - let each hidden service (or other thing) specify its own
    OutboundBindAddress?
295
296
  - Have a mode that doesn't write to disk much, so we can run Tor on
    flash memory (e.g. Linksys routers).
297
298
299
300
301
302
303
304
305
306
307
308
309

Blue-sky:
  - Patch privoxy and socks protocol to pass strings to the browser.
  - Standby/hotswap/redundant hidden services.
  - Robust decentralized storage for hidden service descriptors.
  - The "China problem"
  - Allow small cells and large cells on the same network?
  - Cell buffering and resending. This will allow us to handle broken
    circuits as long as the endpoints don't break, plus will allow
    connection (tls session key) rotation.
  - Implement Morphmix, so we can compare its behavior, complexity, etc.
  - Other transport. HTTP, udp, rdp, airhook, etc. May have to do our own
    link crypto, unless we can bully openssl into it.
310
311
312
313
314
315
  - Need a relay teardown cell, separate from one-way ends.
    (Pending a user who needs this)
  - Handle half-open connections: right now we don't support all TCP
    streams, at least according to the protocol. But we handle all that
    we've seen in the wild.
    (Pending a user who needs this)
316

317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
Non-Coding:
  - Mark up spec; note unclear points about servers
  - Mention controller libs someplace.
P - flesh out the rest of the section 6 of the faq
  . more pictures from ren. he wants to describe the tor handshake
NR- write a spec appendix for 'being nice with tor'
  - tor-in-the-media page
  - Remove need for HACKING file.
  - Figure out licenses for website material.

Website:
  - and remove home and make the "Tor" picture be the link to home.
  - put the logo on the website, in source form, so people can put it on
    stickers directly, etc.
R - make a page with the hidden service diagrams.

  - ask Jan to be the translation coordinator? add to volunteer page.
334
  o track down the patch for cross-compiling.
335