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:
9 * FRAGMENTATION [Ji Lu]
10 * HTTP transport [Matthias]
15 * MySQL / Postgres plugins (datastore, datacache)
19 - "Forcing disconnect of XXX due to inactivity" -- can happen every few MS!?
20 (disconnect does not really succeed, or what?)
21 Also, core does not seem to actually decrement the number of active
22 sessions even if no connections exist for a while! (likely related...)
23 - transport reports bw quota violations on slow systems (core or
26 - utilization can (easily, restart?) go out of control (very large), causing
27 content expiration job to go crazy and delete everything!
29 - gnunet-publish seg faults if given a directory (!)
30 - on some systems, keyword search does not find locally published content
31 (need testcase of command-line tools! - also good to cover getopt API!)
32 [could be related to datastore issue above!]
33 - 2-peer download is still too slow (why?)
34 - advanced FS API parts
35 + pick correct filenames for recursive downloads (mkdir, .gnd)
36 + support recursive download even if filename is NULL and we hence
37 do not generate files on disk (use temp_filename)
38 + bound parallelism (# fs downloads)
39 + distinguish in performance tracking and event signalling between
40 downloads that are actually running and those that are merely in the queue
41 + gnunet-service-fs (hot-path routing, load-based routing, nitpicks)
42 - [gnunet-service-fs.c:208]: member 'LocalGetContext::results_bf_size' is never used
43 - [gnunet-service-fs.c:501]: member 'PendingRequest::used_pids_size' is never used
44 - [gnunet-service-fs.c:654]: member 'ConnectedPeer::last_client_replies' is never used
45 - [gnunet-service-fs.c:669]: member 'ConnectedPeer::avg_delay' is never used
46 - [gnunet-service-fs.c:675]: member 'ConnectedPeer::avg_priority' is never used
47 - [gnunet-service-fs.c:688]: member 'ConnectedPeer::pending_requests' is never used
48 - [gnunet-service-fs.c:694]: member 'ConnectedPeer::last_p2p_replies_woff' is never used
49 - [gnunet-service-fs.c:700]: member 'ConnectedPeer::last_client_replies_woff' is never used
56 - trust: need *fast* way to check/update trust in peers
57 (async peerinfo would not be right; certainly not with the
60 - scheduler should change OS process priority based on task priority;
61 should make better use of task priorities in general
62 - only connect() sockets that are ready (select()) [Nils]
63 [On W32, we need to select after calling socket before
66 - need function to test "equivalency" of HELLOs (or integrate with "merge"?); use in PEERINFO
68 - design & implement new setup tool
70 - good to have for transport/DV evaluation!
73 - implement DV service
74 - implement DV library
75 - implement DV transport plugin
77 - implement performance tests
79 - does not seem to work with timeouts (especially if service is not running)
81 - needs more testing (especially F2F topology)
82 - needs to re-try connecting after disconnect (currently, it
83 initially triggers a connection request, but if that connection
84 fails / goes down, it does not retry in a timely fashion;
85 cause seems to be the 'blacklist_after_attempt' being set to 1h,
86 which is rather long -- and should probably be adjusted based on
87 the number of connections / known peers)
88 - If the topology daemon crashes, peers that were put on the
89 blacklist with transport will never be removed from it (until
90 transport service dies); we should use the blacklist notification
91 API to learn about the exact set of blacklisted peers at all times
92 (FIXME: the transport_api implementation of blacklisting
93 also does not work nicely for this since it won't let us know about
94 disconnect-reconnect events and the implicit whitelisting
95 that might happen here; that's not so bad since we will
96 re-blacklist on pre-connect attempts anyway, so this is
97 a minor issue; OTOH, we might want to be more explicit about
98 allowing/forbidding connects on pre-connect to avoid
99 entering connect attempts to just be blacklisted shortly afterwards).
100 - the code uses the term 'blacklist' for both peers that are forbidden
101 to connect (i.e. F2F mode) as well as peers that we currently
102 won't try to actively connect to ourselves (since we just tried);
103 This is confusing. We need two distinct terms (greylist?).
104 - move code to use hash table instead of linked list
105 - instead of periodically discarding blacklisted entries,
106 simply add task that is triggered at the right time (earlier free,
108 - check if new HELLO learned is different from old HELLO
109 before resetting entire state!
111 - datastore reservation (publishing)
112 - search: availability probes
113 - persistence support (publish, unindex, search, download)
114 - active migration support (in fs or in datastore or new daemon?)
116 + active reply route caching design & implementation of service; gap extension!
118 - how to integrate scheduler with GTK event loop!?
122 - good to have for DV/DHT evaluation!
124 - implement DHT service (needs DV, DATACACHE)
125 - implement DHT library
126 - implement testcases
127 - implement performance tests
131 - need to get rid of synchronous API for service starts (cause all kinds of problems)
132 [=> eliminate for need to tell ARM about service starts most of the time!] [Safey]
133 - better tracking of which config changes actually need to cause process restarts by ARM.
134 - listen for requests to discover dependencies between services (and avoid
135 having to explicitly program start requests)
136 - better crash management (attach debugging support, capture and analyze
137 debug output, detect random vs. deterministic crashes)
140 - location URIs (publish, search, download)
141 - non-anonymous FS service (needs DHT)
142 + DHT integration for search
143 + CS-DHT-functions (DHT-put of LOC)
144 + P2P-functions (DHT-get)
145 - collection API & tests
146 + gnunet-pseudonym (collection support)
147 - implement FS performance tests
154 * Determine RC bugs and fix those!
156 - modify configuration to allow controlling connections for non-local starts
157 - testbed creation with topology (needs working F2F topology)
159 - implement testcases for distributed testing
160 - test basic peer re-configure
161 - test topology creation
162 - test churn generation
163 - consider changing API for peer-group termination to
164 call continuation when done
166 - finalize API design
169 - integration with transport service
170 * MYSQL database backends: [CG]
176 - expand bibliography
177 - convert documentation pages to books
178 - update books (especially for developers)
179 - create good Drupal theme for GNUnet
180 - make a NICE download page and figure out how to
181 enable developers to publish TGZs nicely
182 - port "contact" page
183 - add content type for "todo" items?
184 * POSTGRES database backends: [CG]
187 * Determine RC bugs and fix those!
191 - SMTP transport backend
192 - HTTPS transport backend
193 + improved HTTPS support in MHD
195 - Implement method of learning our external addresses from
196 other peers; need some kind of threshold-based
197 scheme, limiting both the total number of addresses that we accept
198 this way as well as requiring multiple confirmations; also, we
199 should possibly try to confirm that the given address works for
200 us ourselves (loopback-style) before adding it to the list
201 + we may be able to simplify WELCOME messages (no need to add
202 addresses there anymore, but may help to learn them there anyway...).
203 + we probably want some kind of voting/counting for learning IP addresses
204 (maybe including IP addresses in ads proportional to how often others
205 report them? we at least need some protection against >64k HELLOs!),
206 + provide a way to give the user a list of "learned" IP addresses and
207 a way to easily "veto" addresses off the list!
208 => If MiM attacker uses vetoed address, blacklist the specific IP for
209 the presumed neighbour!
210 - implement gnunet-transport (transport configurator / tester)
211 - UPnP-based IP detection
212 (Note: build library always, build service when libxml2/etc. are available)
214 - Remove KBlocks in gnunet-unindex (see discussion with Kenneth Almquist on gnunet-devs in 9/2009)
215 - Allow checking of presence of search results and/or content via command-line tools
216 (add options to gnunet-search / gnunet-download to limit search to local peer)
218 - expire 'ancient' HELLOs (those without valid addresses AND that
219 we have not 'used' (for their public keys) in a while; need a way
220 to track actual 'use')
221 - make sue we also trigger notifications whenever HELLOs expire
228 - should use hash map to look up sessions
230 - api creates many, many short-lived TCP connections; either some
231 clients should use the API differently or we need to change the
232 API to enable re-use of connections to the service
234 - should use BIO instead of mmap
236 - need to periodically probe latency/transport cost changes & possibly switch transport
237 - instantly filter addresses from *other* peers that
238 are *equal* to our own address + port (i.e., localhost:2086). We
239 no longer filter those for outgoing (helps with loopback testing
240 and keeps the code clean), but we should filter strictly *impossible*
241 incoming addresses! This is for efficiency, not correctness.
242 - should use hash map to look up Neighbours
244 - 'server' uses 'GNUNET_PEERINFO_iterate', should probably switch to notification API
245 (for more instant / up-to-date hostlists at lower cost) [OPTIMIZATION]
249 - repeatedly resolve hostname and look up interfaces to determine our own IP
250 - [./transport/plugin_transport_tcp.c:391]: (style) struct or union member 'Plugin::address_update_task' is never used (related to issue above)
252 - [./transport/gnunet-service-transport.c:173]: (style) struct or union member 'TransportPlugin::rebuild' is never used (related to TCP not refreshing external addresses?)
254 - add stats (# bytes available, # bytes used, # PUTs, # GETs, # GETs satisfied)
256 - support inline data in directories for recursive file downloads (fs_download)