Valinor-o offline timestamp processing scripts for Valinor framework

This is the last step in processing the collected timestamps by Valinor-n or Valinor-h frameworks. The provided Jupyter notebooks read from redis and process the burstiness of timestamp records.

More information can be found at in our NSDI paper.

Also checkout Valinor super repository.

Requirements for Valinor offline analysis scripts

• Redis server running on localhost
• Python3, pip3, and pip3 dependencies listed below:
  • matplotlib
  • redis
  • scipy
  • numpy
  • jupyter-lab
  • statsmodels
  • hurst

Installation and Running

• Install python with all the dependencies listed above.
• Run jupyter-lab in the background.
• Copy the redis-dump file containing the timestamps to the default Redis location and start redis server. Refer to load_redis.sh script for an example of how to load redis-server.
• We have provided an example template file for jupyter-notebook that runs the majority of analysis on timestamp records. Simply run the blocks on the notebooks to load the data from redis and process them.
• Some of the processing scripts require large amounts of memory. We recommend >16GB RAMs on the system.

Author

Erfan Sharafzadeh

2020-2023