/ca-plus

Primary LanguagePython

Dataset Release for CA++: Enhancing Carrier Aggregation Beyond 5G (MobiCom'23)

In this repo, we release datasets and analysis codes for experiments.

If you use them in your publications, please cite our MobiCom'23 paper,

@inproceedings{li2023ca++, title={CA++: Enhancing Carrier Aggregation Beyond 5G}, author={Li, Qianru and Zhang, Zhehui and Liu, Yanbing and Tan, Zhaowei and Peng, Chunyi and Lu, Songwu}, booktitle={Proceedings of the 29th Annual International Conference on Mobile Computing and Networking (MobiCom'23)}, year={2023} }

Instructions

Users can follow instructions here to use data and generate experiment results. We have $DATA=./dataset and $SRC=./code.

Evaluation with real-world datasets

  • Figure 10:

    • (a)-(d): Run the following command to get results. Note that city is recognized by column "region" in the file, as we have ATT: R1-C1, R3-C2; T-Mobile: R12-C1; Verizon: R6-C1

    $ python $SRC/measurement_time.py $DATA/{att,tmobile,verizon}/{A,V,T}_instances.csv {A,T,V}

    • (e): Download HST dataset[1], and then unzip it into any folder. Assume the data folder is $hst_data. Run command:

    $ find $hst_data -type f | xargs -I {} python $SRC/measurement_time_hst.py {} $DATA/HST/enb_dict_hst.txt

  • Figure 11:

    • (a) Perform different operatons for for 5G and HST data respectively:

      5G data: Run command $ python $SRC/conn_time_5g.py $DATA/{att,tmobile,verizon}/{A,V,T}_instances.csv

      HST data: Download data file from HST dataset[1], and then unzip it into any folder. Assume the data folder is $hst_data. Then run command $ find $hst_data -type f | xargs -I {} python $SRC/conn_time_hst.py {} $DATA/HST/enb_dict_hst.txt

    • (b) Perform different operatons for for 5G and HST data respectively:

      5G data: Run command $ python $SRC/measurement_report.py $DATA/{att,tmobile,verizon}/{A,V,T}_instances.csv {A,T,V}

      HST data: Download data file from HST dataset[1], and then unzip it into any folder. Assume the data folder is $hst_data. Then run command $ find $hst_data -type f | xargs -I {} python $SRC/signaling_hst.py {} $DATA/HST/enb_dict_hst.txt

  • Figure 12

    • (a)(b): $ python $SRC/throughput_map.py $DATA/att/A-C1_grid_cell_set_ca_em_0405-1226.csv {output}

    • (c)(d): Use data points in $DATA/att/A-R1_grid_avg_bandwidth_ca_em_submission_0405-1226.csv

  • Figure 13: $ python $SRC/channel_width_boost.py $DATA/dataset/{att,tmobile,verizon}/instances_5g.csv {A,T,V}

  • Figure 14: $ python $SRC/throughput_boost_{att,tmo,vrz}.py $DATA/{att,tmobile,verizon}/{A-C1,A-C2,V-C1,T-C1}_grid_cell_set_ca_em_0405-1226.csv {output}

  • Figure 16: We need four sets of operations to tune various policies:

    • PCell threshold A2/A3/A5

      • Specify the delta using corresponding variables (a3_delta, a2_delta, a5_delta) in main function. For example, a3_delta = 0, a2_delta = -10, a5_delta = 0 means reduce A2 threshold by 10.
      • Run command e.g. $ python $SRC/tune_policy_pcell_v.py $DATA/verizon/V-C1_grid_cell_set_ca_em_0405-1226.csv $DATA/verizon/V-C1_rss_cell_0405-0127_311480_0.0005_0_0.csv $DATA/verizon/V-C1_grid_cellset_performance_cfg.csv {output}

    • LTE SCell threshold A1

      • Specify the delta using variable lte_delta.
      • Run command e.g. $ python $SRC/tune_policy_scell_lte_a.py $DATA/att/A-C2_grid_cell_set_ca_em_0405-1226.csv $DATA/att/A-C2_rss_cell_0405-0127_310410_0.0005_0_0.csv $DATA/att/A-C2_grid_cellset_performance_cfg.csv {output}

    • NR SCell threshold B1

      • Specify the delta using variable b1_nr_delta.
      • Run command e.g. $ python $SRC/tune_policy_scell_nr_a.py $DATA/att/A-C2_grid_cell_set_ca_em_0405-1226.csv $DATA/att/A-C2_rss_cell_0405-0127_310410_0.0005_0_0.csv $DATA/att/A-C2_grid_cellset_performance_cfg.csv {output}

    • To disable 5G cells

      • Run command e.g. $ python $SRC/tune_policy_disable_nr.py $DATA/verizon/V-C1_grid_cell_set_ca_em_0405-1226.csv $DATA/verizon/V-C1_rss_cell_0405-0127_311480_0.0005_0_0.csv $DATA/verizon/V-C1_grid_cellset_performance_cfg.csv {output}

Code for inter-channel wireless quality inference

  • Run $SRC/inference/main.m to get recovery error with CA++

[1] Processed raw modem traces and generated text logs for the original dataset from Wang, Jing, et al. "An active-passive measurement study of tcp performance over lte on high-speed rails." The 25th Annual International Conference on Mobile Computing and Networking. 2019.