diff options
| author | Alan Conway <aconway@apache.org> | 2010-11-05 01:55:18 +0000 |
|---|---|---|
| committer | Alan Conway <aconway@apache.org> | 2010-11-05 01:55:18 +0000 |
| commit | d1826cf61cb8b57f2264cbf6cfb88756416c4ff8 (patch) | |
| tree | 2736f2f931fd0598a432fc055492e6228da1828c /cpp | |
| parent | 7bed64eb2467485a5e05e79342076b011dc36930 (diff) | |
| download | qpid-python-d1826cf61cb8b57f2264cbf6cfb88756416c4ff8.tar.gz | |
New cluster plan: added transaction support
git-svn-id: https://svn.apache.org/repos/asf/qpid/trunk/qpid@1031396 13f79535-47bb-0310-9956-ffa450edef68
Diffstat (limited to 'cpp')
| -rw-r--r-- | cpp/design_docs/new-cluster-plan.txt | 4 |
1 files changed, 4 insertions, 0 deletions
diff --git a/cpp/design_docs/new-cluster-plan.txt b/cpp/design_docs/new-cluster-plan.txt index b5cd353fc0..781876e55a 100644 --- a/cpp/design_docs/new-cluster-plan.txt +++ b/cpp/design_docs/new-cluster-plan.txt @@ -344,6 +344,10 @@ of the independence of operations on separate queues. Multicast should encode messages in fixed size buffers (64k)? Can't assume we can send message in one chunk. For 0-10 can use channel numbers & send whole frames packed into larger buffer. +** TODO [#B] Transaction support. +Extend broker::Cluster interface to capture transaction context and completion. +Sequence number to generate per-node tx IDs. +Replicate transaction completion. ** TODO [#B] Batch CPG multicast messages The new cluster design involves a lot of small multicast messages, they need to be batched into larger CPG messages for efficiency. |