summaryrefslogtreecommitdiff
path: root/src/transport/gnunet-service-tng.c
diff options
context:
space:
mode:
authorChristian Grothoff <christian@grothoff.org>2019-02-23 10:47:41 +0100
committerChristian Grothoff <christian@grothoff.org>2019-02-23 10:47:41 +0100
commit3d0f1dd3805bfef30ff7a7f8e246a926b7fa7838 (patch)
treee56f8a29f2e7d23f260f451bc4f965237cecfb6b /src/transport/gnunet-service-tng.c
parent3f6d37c74ee2357cc2cdae7f0022ff422f0a92c6 (diff)
notes
Diffstat (limited to 'src/transport/gnunet-service-tng.c')
-rw-r--r--src/transport/gnunet-service-tng.c13
1 files changed, 13 insertions, 0 deletions
diff --git a/src/transport/gnunet-service-tng.c b/src/transport/gnunet-service-tng.c
index a32d1872c..d392b3a46 100644
--- a/src/transport/gnunet-service-tng.c
+++ b/src/transport/gnunet-service-tng.c
@@ -33,6 +33,19 @@
* transport-to-transport traffic)
*
* Implement next:
+ * - address validation: what is our plan here?
+ * #1 Peerstore only gets 'validated' addresses
+ * #2 transport needs another API to "trigger" validation!
+ * API may be used by core/application or communicators;
+ * => use yet another lib/MQ/connection?
+ * #3 transport should use validation to also establish
+ * effective flow control (for uni-directional transports!)
+ * #4 UDP broadcasting logic must be extended to use the new API
+ * #5 only validated addresses go to ATS for scheduling; that
+ * also ensures we know the RTT
+ * #6 to ensure flow control and RTT are OK, we always do the
+ * 'validation', even if address comes from PEERSTORE
+ * #7
* - ACK handling / retransmission
* - track RTT, distance, loss, etc.
* - DV data structures: