This is a project which tries to benchmark and compare multiple public cloud providers. It also tries to identify the bottleneck resources in the cloud platform for applications.
Put your keys in the ./config/ folder. For Azure you would need to create two keys:
- cloud.pem
- cloud.key
Go to the script folder and use the run.sh script. The default help is pretty self explanatory:
Please add the configs that you want to run to the printConfigs function inside the script. The syntax is:
configFor "Experiment" "InstanceType" "Machine Count" "Disk Type"
Experiment: any one of: tpcds, tpch, tera, spark
Instance type: any of the instance types in Amazon
Instance count: number of instances in the cluster
Disk type: ebs or empty string
By default the disks are set to be 2x250GB of gp2 type per instance.
Feel free to change that within this file.