A high-performance coordination service for write-intensive workloads.
|apps||apps: added libpath by -L option to pass compile|
|conductor||accord: added apps directory and queue code|
|include||accord: Added acrd_tx_atomic_inc|
|lib||rename proto.h to accord_proto.h|
|libacrd||libacrd: removed needless file|
|test||accord: Added acrd_tx_atomic_inc|
|.gitignore||accord: added apps directory and queue code|
|Makefile||accord: added apps directory and queue code|
|TODO||TODO: got together TODO lists|
** Accord project ** NOTE: Now, this project is marked as deprecated, because we've observed that write-intensive workload is much less important than read-intensive workload for coordination service. Please use ZooKeeper. * Overview ---------- - Accord is a coordination service(like ZooKeeper). It features : 1. Accord focuses on write-intensive workloads unlike ZooKeeper. ZooKeeper forwards all write requests to a master server. It can be bottleneck in write-intensive workloads. The below benchmark demonstrates that the write-operation throughput of Accord is much higher than one of ZooKeeper (up to 20 times better throughput at persistent mode, and up to 18 times better throughput at in-memory mode). 2. More flexible transaction support. Not only write, del operations, but also cmp, copy, read operations are supported in transaction operations. 3. In-memory mode and persistent mode support. 4. Message size is unbounded, and partial update is supported. * Project Page ---------- - Getting started / Performance benchmark is avalable from our project page. Please visit http://www.osrg.net/accord * Directory structure ---------- - conductor : Server-side daemon - libacrd : Client-side library - test : Test and benchmark programs. Files in this directory are useful as sample programs of Accord. - include : Header files - lib : Utility library * API Documentation ---------- - Please see include/accord.h or http://www.osrg.net/accord/api.txt - Programs in ACRD_ROOT/test are very useful to understand how to use accord API.