.
├── middleware # Source files and .jar
├── experiments # Log files, processing scripts, plots, images
├── report # Latex source and .pdf
└── README.md # this file
experiments
├── baseline_no_mw_1_server # 2.1
├── parse_memtier.ipynb # ipython notebook that parses MW e client logs
├── logs # logs directory
└── timestamp_dir # timestamp directory (inside all logs and configs)
├── out # processed log files
└── img # plots w
├── baseline_no_mw_2_server # 2.2
├── baseline_mw_1_mw # 3.1
├── baseline_mw_2_mw # 3.2
├── throughput_for_writes # 4.1
├── multigets_sharded # 5.1
├── multigets_non_sharded # 5.2
├── 2k_analysis # 6
└── queue_models # 7
Every log file has a clear structure in its name that identifies the configuration that is being tested, e.g.
ratio1:0_run2_vclients1_workerthreads16_client3_instance1.log
means that this log represents the 2nd repetition of a write-only workload, with 1 vclient, 16 workerthreads and is being sent from the instance 1 of client 3.