fa-5glena-integration

Federated Analytics for B5G/6G networks

Title: Channel-Aware Federated Analytics (FA) in B5G/6G Networks: Dynamic Power Allocation with NS-3 5G-LENA

The project involves developing an integrated framework for simulating FA in real 5G network conditions. The framework applies two algorithms: channel-aware power allocation algorithm to efficiently allocate transmission power for user equipments (UEs) based on distance and channel conditions, and synchronous FA-5GLENA integrated algorithm to integrate the FA with NS-3 5G-LENA simulator and aggregate results for optimized performance within 5G network conditions.

General Project Configuration

1. Download and install Anaconda from the site: https://docs.anaconda.com/anaconda/install/
2. Create environment named "fa_5glena" with python 3.8.20.
3. Inside environment "fa_5glena" create a root directory called "myproject"
4. Inside the root directory "myproject" create subdirectories:
   "FA" - where all FA server and client related files will be placed.
   "shared_memory" - where all shared memory related files will be placed.
   "sim_results" - where network and FA simulation results will be stored.
   After installation of 5G-LENA, a 4th subdirectory named "ns-3-dev" will be generated in this directory.
5. Find other installation steps in each directory's configuration file. See in 5glena, fa, shared_memory directories.
6. The files "net_fetch_sort.cc", "net_select_stats.cc" and "run_project.py" are placed in the root directory "myproject".

This project was created in Anaconda, following the structure: "anaconda3/envs/fa_5glena/myproject", where "myproject" is the root dirctory that contains subdirectories: FA, ns-3-dev, shared_memory and sim_results. It also contains the files "net_fetch_sort.cc", "net_select_stats.cc" and "run_project.py".

Steps to Run the Project

1. Open the Linux terminal (we used gnome-terminal).
2. Navegate to the "shared_memory" subdirectory.
3. Run the following commands to create and load the shared memories:
   g++ -std=c++17 -o shared_cmd.so shared_cmd.cc -lrt
   g++ -std=c++17 -o shared_dr.so shared_dr.cc -lrt
   ./shared_cmd.so
   ./shared_dr.so
   
4. Navegate to "myproject" root directory.
5. Make sure the .cc files "net_fetch_sort.cc" and "net_select_stats.cc" have already been compiled on configuration and respective .so files exist in root directory, before running the project.
6. Run the command "python3 run_project.py" to run the project.

Note: At the end of the simulation, the results are stored in "sim_results" subdirectory.

Validation

The "OptmTx_UMi_algorithm.cc" code runs optimized transmission power allocation algorithm. To run random transmission power allocation scenario, substitute the function "CalculateMinimumTxPowerForUE()" function with code that reads transmission power from "random_txpower_dbm.csv" file found in the repository. To run maximum uniform transmission power allocation, substititute the function "CalculateMinimumTxPowerForUE()" with code that allocates maximum transmission power for all user equipments (UEs).