summaryrefslogtreecommitdiff
path: root/README_v6
diff options
context:
space:
mode:
authorerdgeist <>2009-10-08 20:27:51 +0000
committererdgeist <>2009-10-08 20:27:51 +0000
commit9be794395fd6cd83d5b580ec433edc060182a9b7 (patch)
tree44971b153e96621af5a51a5a33dbfd5c3202a9d9 /README_v6
parent9eb860f0413624da02e2dcd8a8dc04fb3ae1ed99 (diff)
v6 works since january
Diffstat (limited to 'README_v6')
-rw-r--r--README_v69
1 files changed, 1 insertions, 8 deletions
diff --git a/README_v6 b/README_v6
index 6a5a82a..7fac79a 100644
--- a/README_v6
+++ b/README_v6
@@ -1,8 +1 @@
1Q: Why is there no v6-support in opentracker? IPv6 is implemented in opentracker now. You can chose whether your tracker runs in v4 or v6 mode in Makefile. YMMV.
2
3A: Although I tried very hard, implementing v6 right now would be a terrible waste of bandwidth, there is no compact format for v6 addresses, so instead of
4answering "d5:peers6:AAAAPPe" I'd have to send "d5:peersld2:ip39:AAAA:AAAA:AAAA:AAAA:AAAA:AAAA:AAAA:AAAA4:port2:PPPPeee" for a single peer. Even if there was a
5compact mode, v6 addresses still would eat up thrice the memory, v4 addresses take. This, however, wouldn't be a show stopper.
6
7Other problems concern efficient peer selection for obviously v6-capable peers and how to select peers for non-v6 clients. v6 addresses eat up more memory on the
8host, too ;)