1 Implementable right now (but not necessarily important), with caveats
2 (unavailable components that will limit what can be implemented right
3 away), in order in which they will likely be done:
4 * TESTING-TESTBED [Nate]
9 * FRAGMENTATION [Ji Lu]
11 * MySQL / Postgres plugins (datastore, datacache)
18 - implement transport API to pretty-print transport address
19 + resolve cyclic dependency (gnunet-peerinfo -> transport_api -> peerinfo!)
20 + gnunet-transport-service crashes when asked to resolve address (!)
21 + asynchronous transport lookup and non-CPS iterator in peerinfo require
22 some work to ensure printing in the proper order...
24 - SET_KEY is triggered every 2.5 minutes (after session is up); should
25 just do PING (not SET_KEY + PING)
26 - "Forcing disconnect of XXX due to inactivity" -- can happen every few MS!?
27 (disconnect does not really succeed, or what?)
28 Also, core does not seem to actually decrement the number of active
29 sessions even if no connections exist for a while! (likely related...)
30 - transport reports bw quota violations on big-endian systems (core or
33 - need function to test "equivalency" of HELLOs; use in topology!
36 + active reply route caching design & implementation of service,
38 - gnunet-publish cannot be aborted using CTRL-C
39 - on some systems, keyword search does not find locally published content
40 (need testcase of command-line tools!)
41 - 2-peer download is still too slow (why?)
42 - advanced FS API parts
43 + namespaces: fundamental namespace API
44 + search: SBlocks, KSBlocks, probes, notify FS-service of known results
46 + location URIs (publish, search, download)
47 + indexing: index-failure-cleanup
48 + download: management/bounding of parallel downloads (for recursive dl.)
49 + datastore reservation (publishing)
50 + gnunet-publish (URI args)
51 + gnunet-download (directory)
52 + gnunet-search (options, incl. namespace search)
53 + gnunet-download (many options)
54 + gnunet-directory (man page, options)
55 + gnunet-pseudonym (all of it)
56 + gnunet-service-fs (remove failing on-demand blocks, hot-path routing, load-based routing, nitpicks)
57 - implement adv. FS testcases
59 + insert: sblocks, loc uris
61 + search: sblocks, skblocks, loc uris
64 - implement FS performance tests
69 - [./fs/gnunet-service-fs.c:208]: (style) struct or union member 'LocalGetContext::results_bf_size' is never used
70 - [./fs/gnunet-service-fs.c:501]: (style) struct or union member 'PendingRequest::used_pids_size' is never used
71 - [./fs/gnunet-service-fs.c:654]: (style) struct or union member 'ConnectedPeer::last_client_replies' is never used
72 - [./fs/gnunet-service-fs.c:669]: (style) struct or union member 'ConnectedPeer::avg_delay' is never used
73 - [./fs/gnunet-service-fs.c:675]: (style) struct or union member 'ConnectedPeer::avg_priority' is never used
74 - [./fs/gnunet-service-fs.c:688]: (style) struct or union member 'ConnectedPeer::pending_requests' is never used
75 - [./fs/gnunet-service-fs.c:694]: (style) struct or union member 'ConnectedPeer::last_p2p_replies_woff' is never used
76 - [./fs/gnunet-service-fs.c:700]: (style) struct or union member 'ConnectedPeer::last_client_replies_woff' is never used
83 * Module features to implement:
85 + good to have for DV evaluation!
88 + implement DV service [Nate & CG]
89 + implement DV library [Nate]
90 + implement DV transport plugin [Nate & CG]
91 + implement testcases [Nate]
92 + implement performance tests [Nate]
94 - design & implement new setup tool
97 + how to integrate scheduler with GTK event loop!
99 - 'server' uses 'GNUNET_PEERINFO_iterate', should probably switch to notification API
100 (for more instant / up-to-date hostlists at lower cost) [OPTIMIZATION]
102 - does not seem to work with timeouts (especially if service is not running)
104 - trust: need *fast* way to check/update trust in peers
105 (async peerinfo would not be right)
106 - scheduler should change OS process priority based on task priority;
107 should make better use of task priorities in general
108 - only connect() sockets that are ready (select()) [Nils]
109 [On W32, we need to select after calling socket before
112 - add stats (# bytes available, # bytes used, # PUTs, # GETs, # GETs satisfied)
114 - persistence support (publish, unindex, search, download)
115 - non-anonymous FS service (needs DHT)
116 + basic DHT integration
117 + CS-DHT-functions (DHT-put of LOC)
118 + P2P-functions (DHT-get)
119 - active migration support (in fs or in datastore?)
121 - needs more testing (especially F2F topology)
122 - needs to re-try connecting after disconnect (currently, it
123 initially triggers a connection request, but if that connection
124 fails / goes down, it does not retry in a timely fashion;
125 cause seems to be the 'blacklist_after_attempt' being set to 1h,
126 which is rather long -- and should probably be adjusted based on
127 the number of connections / known peers)
128 - If the topology daemon crashes, peers that were put on the
129 blacklist with transport will never be removed from it (until
130 transport service dies); we should use the blacklist notification
131 API to learn about the exact set of blacklisted peers at all times
132 (FIXME: the transport_api implementation of blacklisting
133 also does not work nicely for this since it won't let us know about
134 disconnect-reconnect events and the implicit whitelisting
135 that might happen here; that's not so bad since we will
136 re-blacklist on pre-connect attempts anyway, so this is
138 - the code uses the term 'blacklist' for both peers that are forbidden
139 to connect (i.e. F2F mode) as well as peers that we currently
140 won't try to actively connect to ourselves (since we just tried);
141 This is confusing. We need two distinct terms.
142 - move code to use hash table instead of linked list
143 - instead of periodically discarding blacklisted entries,
144 simply add task that is triggered at the right time (earlier free,
146 - check if new HELLO learned is different from old HELLO
147 before resetting entire state!
150 * Module features to implement:
152 + good to have for DV/DHT evaluation!
154 + implement DHT service (needs DV, DATACACHE)
155 + implement DHT library
156 + implement testcases
157 + implement performance tests
160 * Plugins to implement:
161 - UDP backend (RC-pre2)
162 + Fragmentation library
164 - HTTP backend (RC-pre2)
166 - need to get rid of synchronous API for service starts (cause all kinds of problems)
167 [=> eliminate for need to tell ARM about service starts most of the time!] [Safey]
168 - better tracking of which config changes actually need to cause process restarts by ARM.
169 - listen for requests to discover dependencies between services (and avoid
170 having to explicitly program start requests)
171 - better crash management (attach debugging support, capture and analyze
172 debug output, detect random vs. deterministic crashes)
179 + see discussions @ FISL about integration with event loop!
180 * Determine RC bugs and fix those!
182 - modify configuration to allow controlling connections for non-local starts
183 - testbed creation with topology (needs working F2F topology) [Nate]
184 - testbed with churn [Nate]
185 - implement testcases for distributed testing [Nate]
186 - test basic peer re-configure [Nate]
187 - test topology creation [Nate]
188 - test churn generation [Nate]
189 - consider changing API for peer-group termination to
190 call continuation when done
196 - update webpage documentation
198 - expand bibliography
199 - convert documentation pages to books
200 - update books (especially for developers)
201 - create good Drupal theme for GNUnet
202 - make a NICE download page and figure out how to
203 enable developers to publish TGZs nicely
204 - port "contact" page
205 - add content type for "todo" items?
207 * Plugins to implement:
208 - MySQL database backends
211 - Postgres database backends
215 * Determine RC bugs and fix those!
220 - SMTP transport backend
221 - HTTPS transport backend
222 + improved HTTPS support in MHD
224 - Implement method of learning our external addresses from
225 other peers; need some kind of threshold-based
226 scheme, limiting both the total number of addresses that we accept
227 this way as well as requiring multiple confirmations; also, we
228 should possibly try to confirm that the given address works for
229 us ourselves (loopback-style) before adding it to the list
230 + we may be able to simplify WELCOME messages (no need to add
231 addresses there anymore, but may help to learn them there anyway...).
232 + we probably want some kind of voting/counting for learning IP addresses
233 (maybe including IP addresses in ads proportional to how often others
234 report them? we at least need some protection against >64k HELLOs!),
235 + provide a way to give the user a list of "learned" IP addresses and
236 a way to easily "veto" addresses off the list!
237 => If MiM attacker uses vetoed address, blacklist the specific IP for
238 the presumed neighbour!
239 - implement gnunet-transport (transport configurator / tester)
240 - UPnP-based IP detection
241 (Note: build library always, build service when libxml2/etc. are available)
243 - Remove KBlocks in gnunet-unindex (see discussion with Kenneth Almquist on gnunet-devs in 9/2009)
244 - Allow checking of presence of search results and/or content via command-line tools
245 (add options to gnunet-search / gnunet-download to limit search to local peer)
247 - expire 'ancient' HELLOs (those without valid addresses AND that
248 we have not 'used' (for their public keys) in a while; need a way
249 to track actual 'use')
250 - make sue we also trigger notifications whenever HELLOs expire
256 - should use hash map to look up sessions
258 - api creates many, many short-lived TCP connections; either some
259 clients should use the API differently or we need to change the
260 API to enable re-use of connections to the service
262 - should use BIO instead of mmap
264 - need to periodically probe latency/transport cost changes & possibly switch transport
265 - instantly filter addresses from *other* peers that
266 are *equal* to our own address + port (i.e., localhost:2086). We
267 no longer filter those for outgoing (helps with loopback testing
268 and keeps the code clean), but we should filter strictly *impossible*
269 incoming addresses! This is for efficiency, not correctness.
270 - should use hash map to look up Neighbours
274 - repeatedly resolve hostname and look up interfaces to determine our own IP
275 - [./transport/plugin_transport_tcp.c:391]: (style) struct or union member 'Plugin::address_update_task' is never used (related to issue above)
277 - [./transport/gnunet-service-transport.c:173]: (style) struct or union member 'TransportPlugin::rebuild' is never used (related to TCP not refreshing external addresses?)