oamyjin/RadioSaber

RadioSaber is State-of-the-Art 5G RAN slicing algorithm that achieves high spectrum efficiency, ensures weighted fairness across slices, and allows slices to customize scheduling policies.

Instructions (to group members)

Always check at least that there are no compilation errors. Simply remove the LTE-Sim executable and do make -j8 and see if the compilation succeeds.

Git operations

For small changes that do not need review, directly modify the master branch. However, always pull before pushing. Do git fetch origin and git merge origin/master and solve any potential conflicts, then do git push -u origin.

For large changes that need review, do it on a different branch.

git fetch origin
git merge origin/master  # Update to the latest master branch
git checkout -b [feature-branch]  # Create and checkout to a new branch

Replace [feature-branch] with a customized branch name that is instructive. Make your changes in that branch. Fetch and merge the master branch frequently to keep your branch up-to-date and avoid too much conflicts. When you are done with your modifications, do git push -u origin [feature-branch]. Then go to GitHub and make a pull request.

Style guide

Regardless of what format the simluator previously follows, we will stick to Google C++ format for any further changes made. One should install clang-format and format the code prior to any code commit.

sudo apt install clang-format  # Install clang formatting tool
git clang-format --style=file  # Do this prior to any commit

Week of 10/30

Added customizable objective (metric) to the RadioSaber scheduler. This can be accessed via SingleCellWithI and setting the inter-slice scheduler argument. 9 stands for the original RadioSaber (with spectral efficiency as objective), 91 stands for RadioSaber with proportional fairness, and 92 stands for RadioSaber with M-LWDF (modified largest weighted delay first).

Added a trial experiment script, adapted from RadioSaber but with different interslice metrics.

Week of 11/06

Prepared configurations for the comparison test. Refactored and implemented the plots for throughput, RB/s, flow completion time, and queuing delay. To run the scripts:

cd ran-sched-experiments/contradict-objective
./run.sh                                   # Create all logs (test included)
./plot.py --test                           # Plot test log
./plot.py --intra-alg all --flow-type all  # Plot all experiment logs