- BMv2 supported
- All needed python packets installed
-
Run Planter with
-m(manually config) mode:Input:
python3 Planter.py -mOutput:
(base) changgang@linux:~/Planter$ python3 Planter.py -m ____ ___ __ /\ _`\ /\_ \ /\ \__ \ \ \L\ \//\ \ __ ___\ \ ,_\ __ _ __ \ \ ,__/ \ \ \ /'__`\ /' _ `\ \ \/ /'__`/\`'__\ \ \ \/ \_\ \_/\ \L\.\_/\ \/\ \ \ \_/\ __\ \ \/ \ \_\ /\____\ \__/.\_\ \_\ \_\ \__\ \____\ \_\ \/_/ \/____/\/__/\/_/\/_/\/_/\/__/\/____/\/_/ Version 0.0.1 Please set the following directories: - Where is your data folder? (default = '/home/Data')Press
enterif default is correct. Otherwise, input the correct directory and pressenter:Output:
- Where is your Planter folder? (default = '/home/Planter')Press
enterif default is correct. Otherwise, input the correct directory and pressenter:Output:
- Which model do you want to plant? (default = DT | options = -h)Input
RFpressenter:Output:
- Which type of model you want to plant? (default = 1 | options = -h)Input
EBpressenter:Output:
- Which dataset do you want to use? (default = UNSW_5_tuple | options = -h)Input
Irispressenter:Output:
- Number of features? (default = 5)Input
4pressenter:Output:
src/configs/Planter_config.json is generated - Use the testing mode or not? (default = y)ℹ️ Testing mode randomly selects 2000 samples from the dataset for fast training.
Input
npressenter:Output:
The shape of the dataset - train x: (105, 4) train y: (105,) test x: (45, 4) test y: (45,) Add the following path: /home/Planter/src/models/RF/Type_EB - Number of trees? (default = 5)Input
5pressenter:Output:
- Number of depth? (default = 4)Input
4pressenter:Output:
- Number of leaf nodes? (default = 1000)Input
1000pressenter:Output:
(The output matrix gives the classification performance of the scikit-learn model on python.)
precision recall f1-score support 0 1.0000 1.0000 1.0000 13 1 0.9500 0.9500 0.9500 20 2 0.9167 0.9167 0.9167 12 accuracy 0.9556 45 macro avg 0.9556 0.9556 0.9556 45 weighted avg 0.9556 0.9556 0.9556 45 Classification results are downloaded to log as src/logs/log.json The table for Tree: 4 is generated Generating vote to class table...Done Begin transfer: Feature table 0 Input table has: 80 entries and: 6 different ranges and output TCAM entry has 15 80th testing sample with correct matches: 100.0 % and 0 errors. Begin transfer: Feature table 1 Input table has: 45 entries and: 5 different ranges and output TCAM entry has 11 45th testing sample with correct matches: 100.0 % and 0 errors. Begin transfer: Feature table 2 Input table has: 70 entries and: 7 different ranges and output TCAM entry has 16 70th testing sample with correct matches: 100.0 % and 0 errors. Begin transfer: Feature table 3 Input table has: 26 entries and: 7 different ranges and output TCAM entry has 12 26th testing sample with correct matches: 100.0 % and 0 errors. Ternary_Table is generated Exact_Table is generated /home/Planter/src/configs/Planter_config.json is generated - Test the table or not? (default = y)Input
ypressenter:Output:
(The output matrix gives the classification performance from the pipeline logic by using the generated table entries on python.)
Test the exact feature table, extract code, and decision table (feel free if the acc to sklearn is slightly lower than 1) switch_prediction: 1, test_y: 1, with acc: 0.956, with acc to sklearn: 1.0, with error: 0.0, M/A format macro f1: 0.955, macro f1: 0.955 The accuracy of the match action format of Random Forest is 0.9555555555555556 precision recall f1-score support 0 1.0000 1.0000 1.0000 13 1 0.9500 0.9500 0.9500 20 2 0.9167 0.9167 0.9167 12 accuracy 0.9556 45 macro avg 0.9556 0.9556 0.9556 45 weighted avg 0.9556 0.9556 0.9556 45 src/configs/Planter_config.json is generated Exact match entries: 187 Ternary match entries: 54 - Which architecture do you use? (default = tna | options = -h)Input
v1modelpressenter:Output:
- Which is the use case? (default = standard_classification | options = -h)Input
performancepressenter:Output:
Add the following path: /home/Planter/src/use_cases/standard_classification Add the following path: /home/Planter/src/architectures/v1model Generating p4 files and load data file... Done - What is the target device? (default = Tofino | options = -h)Input
bmv2pressenter:Output:
- Which type of mode do you want to choose? (default = software | options = -h)Input
softwarepressenter:Output:
Dump the targets info to src/configs/Planter_config.json Add the following path: /home/Planter/src/targets/bmv2/software - Send packets to which port? (default = 'eth0'):Input
eth0pressenter:Output:
- Please input your password for 'sudo' command:Input
your_own_sudo_passwordpressenter:Output:
(The output matrix gives the classification result of generated P4 on BMv2 in emulated Mininet environment.)
ip link show *** Killing stale mininet node processes pkill -9 -f mininet: *** Shutting down stale tunnels pkill -9 -f Tunnel=Ethernet pkill -9 -f .ssh/mn rm -f ~/.ssh/mn/* *** Cleanup complete. rm -f *.pcap rm -rf build pcaps logs mkdir -p build pcaps logs p4c-bm2-ss --p4v 16 --p4runtime-files build/RF_performance_Iris.p4.p4info.txt -o build/RF_performance_Iris.json RF_performance_Iris.p4 sudo python3 ../../utils/run_exercise.py -t topology.json -j build/RF_performance_Iris.json -b simple_switch_grpc Reading topology file. Building mininet topology. Configuring switch s1 with file s1-commands.txt s1 -> gRPC port: 50051 ********** h1 default interface: eth0 10.0.1.1 08:00:00:00:01:01 ********** Starting mininet CLI ====================================================================== Welcome to the BMV2 Mininet CLI! ====================================================================== Your P4 program is installed into the BMV2 software switch and your initial runtime configuration is loaded. You can interact with the network using the mininet CLI below. To inspect or change the switch configuration, connect to its CLI from your host operating system using this command: simple_switch_CLI --thrift-port <switch thrift port> To view a switch log, run this command from your host OS: tail -f /home/changgang/Documents/GitHub/Planter/src/targets/bmv2/software/model_test/test_environment/logs/<switchname>.log To view the switch output pcap, check the pcap files in /home/changgang/Documents/GitHub/Planter/src/targets/bmv2/software/model_test/test_environment/pcaps: for example run: sudo tcpdump -xxx -r s1-eth1.pcap To view the P4Runtime requests sent to the switch, check the corresponding txt file in /home/changgang/Documents/GitHub/Planter/src/targets/bmv2/software/model_test/test_environment/logs: for example run: cat /home/changgang/Documents/GitHub/Planter/src/targets/bmv2/software/model_test/test_environment/logs/s1-p4runtime-requests.txt mininet> Predicted load table time ... (0.037s) ============================================================================================== = Two steps to exit Planter: 1. input 'exit', press 'return'; 2. press ctrl + c = ============================================================================================== Test the switch model, the result should be the same as previous python test Switch model 45th prediction: 1, test_y: 1, with acc: 0.956, with acc to sklearn: 1.0, with error: 0.0, M/A format macro f1: 0.9556, macro f1: 0.9556 precision recall f1-score support 0 1.0000 1.0000 1.0000 13 1 0.9500 0.9500 0.9500 20 2 0.9167 0.9167 0.9167 12 accuracy 0.9556 45 macro avg 0.9556 0.9556 0.9556 45 weighted avg 0.9556 0.9556 0.9556 45 ======================================= Test Finished ======================================== mininet> Process <load data> cost 0.0073s Process <train model> cost 0.0154s Process <convert model> cost 0.0182s Process <python-based test> cost 0.0779s