0.9.0pre1:
-* PEERINFO:
- - trust: need *fast* way to check/update trust in peers
- (async peerinfo would not be right; certainly not with the current API)
-* TOPOLOGY:
- - move code to use hash table instead of linked list
- - check if new HELLO learned is different from old HELLO
- before resetting entire state!
- - the code uses the term 'blacklist' for both peers that are forbidden
- to connect (i.e. F2F mode) as well as peers that we currently
- won't try to actively connect to ourselves (since we just tried);
- This is confusing. We need two distinct terms (greylist?).
- - instead of periodically discarding blacklisted entries,
- simply add task that is triggered at the right time (earlier free,
- more balanced load)
-
- - needs more testing (especially F2F topology)
- - needs to re-try connecting after disconnect (currently, it
- initially triggers a connection request, but if that connection
- fails / goes down, it does not retry in a timely fashion;
- cause seems to be the 'blacklist_after_attempt' being set to 1h,
- which is rather long -- and should probably be adjusted based on
- the number of connections / known peers)
- - If the topology daemon crashes, peers that were put on the
- blacklist with transport will never be removed from it (until
- transport service dies); we should use the blacklist notification
- API to learn about the exact set of blacklisted peers at all times
- (FIXME: the transport_api implementation of blacklisting
- also does not work nicely for this since it won't let us know about
- disconnect-reconnect events and the implicit whitelisting
- that might happen here; that's not so bad since we will
- re-blacklist on pre-connect attempts anyway, so this is
- a minor issue; OTOH, we might want to be more explicit about
- allowing/forbidding connects on pre-connect to avoid
- entering connect attempts to just be blacklisted shortly afterwards).
-* DATASTORE:
- - API lacks cancellation methods (needed? or is disconnect enough?)
- - may also want to integrate request queuing here instead of gnunet-service-fs_drq.c
* FS: [CG]
- - support recursive download even if filename is NULL and we hence
- do not generate files on disk (use temp_filename)
- bound parallelism (# fs downloads)
- distinguish in performance tracking and event signalling between
downloads that are actually running and those that are merely in the queue
- persistence support (publish, unindex, search, download)
- - active migration support (in fs or in datastore or new daemon?)
- - gnunet-service-fs (hot-path routing, load-based routing, nitpicks)
+ - gnunet-service-fs (hot-path routing, load-based routing, nitpicks)
- [gnunet-service-fs.c:208]: member 'LocalGetContext::results_bf_size' is never used
- [gnunet-service-fs.c:501]: member 'PendingRequest::used_pids_size' is never used
- [gnunet-service-fs.c:654]: member 'ConnectedPeer::last_client_replies' is never used
- [gnunet-service-fs.c:700]: member 'ConnectedPeer::last_client_replies_woff' is never used
- GAP improvements:
+ active reply route caching design & implementation of service; gap extension!
+* MIGRATION [CG]
+ - on-demand encoding => move logic to block-library!?
+ - peer selection => how to consider latency/bw/etc?
+ - content transmission => how often the same block?
+ - how to select delay before next migration?
+ - migration to us
+ - testing
+ - integrate with FS or not? (peer list, index/on-demand encoding, block code,
+ inbound priority assignment; all would be easier with tight integration!)
* TBENCH: [MW]
- good to have for transport/DV evaluation!
* DV: [Nate]
- implement DV library (looks done)
- implement DV transport plugin
- implement testcases
- - implement performance tests
+ - implement performance tests (needs tbench)
+* TOPOLOGY:
+ - needs more testing (especially F2F topology) & transport blacklisting
* UTIL:
- only connect() sockets that are ready (select()) [Nils]
[On W32, we need to select after calling socket before doing connect etc.]
+ download
+ search
+ unindex
+* MIGRATION:
+ - improved content selection (not just 'get_random')
0.9.0pre3:
* Determine RC bugs and fix those!
* MYSQL database backends: [CG]
- datacache
- datastore
+* FS:
+ - reconstruct IBLOCKS from DBLOCKS if possible (during download; see FIXME in fs_download)
+
0.9.0:
* new webpage:
Optimizations:
* TCP:
- should use hash map to look up sessions
-* PEERINFO:
- - api creates many, many short-lived TCP connections; either some
- clients should use the API differently or we need to change the
- API to enable re-use of connections to the service
* STATISTICS:
- should use BIO instead of mmap
* TRANSPORT:
- need to periodically probe latency/transport cost changes & possibly switch transport
- - instantly filter addresses from *other* peers that
- are *equal* to our own address + port (i.e., localhost:2086). We
- no longer filter those for outgoing (helps with loopback testing
- and keeps the code clean), but we should filter strictly *impossible*
- incoming addresses! This is for efficiency, not correctness.
- - should use hash map to look up Neighbours
+ - should use hash map to look up Neighbours (service AND plugins!)
* HOSTLIST:
- 'server' uses 'GNUNET_PEERINFO_iterate', should probably switch to notification API
- (for more instant / up-to-date hostlists at lower cost) [OPTIMIZATION]
+ (for more instant / up-to-date hostlists at lower cost)
* DATASTORE (?):
- check for duplicates on insertion (currently, same content is frequently
stored again [seen with KBLOCKS and SBLOCKS]!)
+* PEERINFO:
+ - merge multiple HELLOs of the same peer in the transmission queue
+ (theoretically reduces overhead; bounds message queue size)
+ - merge multiple iteration requests over "all" peers in the queue
+ (theoretically reduces overhead; bounds messgae queue size)
Minor features:
* TCP:
- add stats (# bytes available, # bytes used, # PUTs, # GETs, # GETs satisfied)
* FS:
- support inline data in directories for recursive file downloads (fs_download)
-
+* BLOCKS:
+ - testcase would be nice...