diff --git a/docs/VOLUME.md b/docs/VOLUME.md index 4bc6a3c..36b529e 100644 --- a/docs/VOLUME.md +++ b/docs/VOLUME.md @@ -18,7 +18,7 @@ The First test on a Riak Cluster has been based on the following configuration: - AAE set to passive - sync writes enabled (on both backends) - An object size of 8KB -- A pareto distribution of requests with a keyspace of 50M keys +- A pareto distribution of requests with a keyspace of 200M keys - 5 GETs for each UPDATE - 4 hour test run @@ -36,7 +36,7 @@ Tail latency under load is around in leveled is less than 5% of the comparable v leveled Results | eleveldb Results :-------------------------:|:-------------------------: -![](../test/volume/cluster_one/output/summary_leveled_5n_45t_fixed.png "LevelEd") | ![](../test/volume/cluster_one/output/summary_leveldb_5n_45t_fixed.png "LevelDB") +![](../test/volume/cluster_one/output/summary_leveled_5n_45t_ii.png "LevelEd") | ![](../test/volume/cluster_one/output/summary_leveldb_5n_45t_ii.png "LevelDB") ### Lies, damned lies etc @@ -64,7 +64,7 @@ Also tested was this same d2.2xlarge cluster, but without sync_on_write. Result leveled Results | eleveldb Results :-------------------------:|:-------------------------: -![](../test/volume/cluster_two/output/summary_nosync_d2_leveled_fixed.png "LevelEd") | ![](../test/volume/cluster_two/output/summary_nosync_d2_leveldb_fixed.png "LevelDB") +![](../test/volume/cluster_two/output/summary_nosync_d2_leveled_ii.png "LevelEd") | ![](../test/volume/cluster_two/output/summary_nosync_d2_leveldb_ii.png "LevelDB") This test showed a 26.7% improvement in throughput when using leveled. The improvement in tail latency in this test had leveled at about 25% of the tail latency of leveldb - so still a substantial improvement, although not of the same order of magnitude as the sync_on_write test with i2.2xlarge and leveldb. @@ -84,7 +84,7 @@ The test has been based on the following configuration: - AAE set to passive - sync writes disabled (on both backends) - An object size of 8KB -- A pareto distribution of requests with a keyspace of 50M keys +- A pareto distribution of requests with a keyspace of 200M keys - 5 GETs for each UPDATE - 6 hour test run @@ -111,13 +111,13 @@ The difference in volatility is clear when graphing the test performance: leveled Results | eleveldb Results :-------------------------:|:-------------------------: -![](../test/volume/cluster_three/output/summary_5n_100t_i2_nosync_leveled.png "LevelEd") | ![](../test/volume/cluster_three/output/summary_5n_100t_i2_nosync_leveldb.png "LevelDB") +![](../test/volume/cluster_three/output/summary_5n_100t_i2_nosync_leveled_ii.png "LevelEd") | ![](../test/volume/cluster_three/output/summary_5n_100t_i2_nosync_leveldb_ii.png "LevelDB") Repeating this test over 8 hours with a smaller keyspace and a uniform (not pareto) key distribution yields similar results. This change in test parameters was done to reduce the percentage of not_founds towards the end of the test (down to about 20%). leveled Results | eleveldb Results :-------------------------:|:-------------------------: -![](../test/volume/cluster_three/output/summary_5n_100t_i2_nosync_leveled_uniform.png "LevelEd") | ![](../test/volume/cluster_three/output/summary_5n_100t_i2_nosync_leveldb_uniform.png "LevelDB") +![](../test/volume/cluster_three/output/summary_5n_100t_i2_nosync_leveled_uniform_ii.png "LevelEd") | ![](../test/volume/cluster_three/output/summary_5n_100t_i2_nosync_leveldb_uniform_ii.png "LevelDB") ### Lies, damned lies etc diff --git a/test/volume/cluster_one/output/summary_leveldb_5n_45t_ii.png b/test/volume/cluster_one/output/summary_leveldb_5n_45t_ii.png new file mode 100644 index 0000000..c86c966 Binary files /dev/null and b/test/volume/cluster_one/output/summary_leveldb_5n_45t_ii.png differ diff --git a/test/volume/cluster_one/output/summary_leveled_5n_45t_ii.png b/test/volume/cluster_one/output/summary_leveled_5n_45t_ii.png new file mode 100644 index 0000000..090fcf7 Binary files /dev/null and b/test/volume/cluster_one/output/summary_leveled_5n_45t_ii.png differ diff --git a/test/volume/cluster_three/output/summary_5n_100t_i2_nosync-leveldb_ii.png b/test/volume/cluster_three/output/summary_5n_100t_i2_nosync-leveldb_ii.png new file mode 100644 index 0000000..6f26cc1 Binary files /dev/null and b/test/volume/cluster_three/output/summary_5n_100t_i2_nosync-leveldb_ii.png differ diff --git a/test/volume/cluster_three/output/summary_5n_100t_i2_nosync_leveldb_uniform_ii.png b/test/volume/cluster_three/output/summary_5n_100t_i2_nosync_leveldb_uniform_ii.png new file mode 100644 index 0000000..ba594ae Binary files /dev/null and b/test/volume/cluster_three/output/summary_5n_100t_i2_nosync_leveldb_uniform_ii.png differ diff --git a/test/volume/cluster_three/output/summary_5n_100t_i2_nosync_leveled_ii.png b/test/volume/cluster_three/output/summary_5n_100t_i2_nosync_leveled_ii.png new file mode 100644 index 0000000..650bd5d Binary files /dev/null and b/test/volume/cluster_three/output/summary_5n_100t_i2_nosync_leveled_ii.png differ diff --git a/test/volume/cluster_three/output/summary_5n_100t_i2_nosync_leveled_uniform_ii.png b/test/volume/cluster_three/output/summary_5n_100t_i2_nosync_leveled_uniform_ii.png new file mode 100644 index 0000000..9f473c7 Binary files /dev/null and b/test/volume/cluster_three/output/summary_5n_100t_i2_nosync_leveled_uniform_ii.png differ diff --git a/test/volume/cluster_two/output/summary_nosync_d2_leveldb_ii.png b/test/volume/cluster_two/output/summary_nosync_d2_leveldb_ii.png new file mode 100644 index 0000000..38d57d6 Binary files /dev/null and b/test/volume/cluster_two/output/summary_nosync_d2_leveldb_ii.png differ diff --git a/test/volume/cluster_two/output/summary_nosync_d2_leveled_ii.png b/test/volume/cluster_two/output/summary_nosync_d2_leveled_ii.png new file mode 100644 index 0000000..5929ac7 Binary files /dev/null and b/test/volume/cluster_two/output/summary_nosync_d2_leveled_ii.png differ