Source code for the 'Energy Smart SFC Orchestrator (ESSO)' paper

Note: The terms NFV-POD and CO are used interchangeably in this repository.

The run_simulation.ipynb notebook demonstrates how to run a simulation.

Repository directory/file structure:

• data: contains rocketfuel raw data, code to generate SFC requests, and generated datasets used in the paper

  • Each dataset is contained in its own directory under data. Each dataset directory (data/<dataset-name>) must contain the following files to run a simulation

    • co_topology.dat: contains the inter-NFV-POD/inter-CO topology, CO2 emission rates, and flag to denote whether renewable energy is available at a NFV-POD
    • timeslots.dat: contains SFC arrival events and their details (number of VNF, their type, bandwidth, latency) per timeslots.
    • vnf_types.dat: contains information about VNF types. The flavor-id listed in this file is referenced from the timeslots.dat file.
    • init_topology: this file is generated by running src/process_topology.o, which parses the co_topology.dat file to generate an extended topology containing both inter- and intra-NFV-POD network. This network represents the overall network topology along with the inernal topology (server racks, ToR switch) of each NFV-POD. The run_simulation.ipynb demonstrates this process as well.

    File formats with samples are explained in detail in data/README.md. The data directory also contains a file named greencap.dat. This file lists randomized renewabel energy data for 24 hours. In each run of the simulation the file is shuffled, and renewable energy for NFV-PODs are selected from it based on the flag included in the co_topology.dat file.

• runs: conatins files and directories generated during a simulation (additional details in run_simulation.ipynb)
• src: contains some randomly typed files!