lixiao-c / ycsb-generator

Generate YCSB dataset and output to file

YCSB dataset generator

Generate YCSB datasets according to different worloads, and output to file.

The origin YCSB repository is [https://github.com/brianfrankcooper/YCSB]

Getting Started

1. Download Redis and install [https://github.com/redis/redis]

   git clone https://github.com/redis/redis
   cd redis
   make

2. Write wordload cnofiguration file in ./workloads . We can set some parameters in this file, including key distribution, key length, and etc. The guidelines can be found in [https://github.com/brianfrankcooper/YCSB/wiki/Implementing-New-Workloads]

3. Generate dataset

   First, start redis server

   cd redis
   src/redis-server --bind 127.0.0.1 --port 6379

   Then start ycsb.sh, which needs installing java and maven

   cd YCSB-generator
   ./ycsb.sh <selected_workload> <loadnum> <runnum> <loadfile> <runfile>

   The dataset is stored in loadfile and runfile. The format is

   insert <key>
   update <key>
   read <key>
   

Changing skewness of Zipfian distribution

The default skewness of YCSB is 0.99. If you want to generate datasets with different skewness, please modify these files below. You should replace 0.99 in these file with you wanted skewness.

core/src/main/java/site/ycsb/generator/ZipfianGenerator.java
core/src/main/java/site/ycsb/generator/ScrambledZipfianGenerator.java
core/src/main/java/site/ycsb/workloads/RestWorkload.java

Enjoy!