STEP 0: Reserve POWDER Resources

• BARE MINIMUM
  • 1 d740/d430 compute node
  • 1 x310 SDR (ideally browning cellsdr-x310)
  • 1 B210 SDR (ideally bookstore)
  • UL Frequency Range (ideally 2500-2510 MHz)
  • DL Frequency Range (ideally 2620-2630 MHz)Reserve srsLTE experiment resources

STEP 1: Setup Experiment Profiles

• ota_srslte profile
• single-pc profile

STEP2: Connect to Nodes

• In Local_Scripts run node_start.sh to open up shells for EPC, eNB, UE, KafkaServer

  • This asks for your POWDER username, location of data center, PC number(s) for the EPC(s)/eNB(s), PC number for the Kafka server, and location of UE(s) (bookstore, meb, etc).

• Clone this repo onto the node

STEP 3: Spin up Kafka server

NOTE: By default, log retention is 1 hour. Edit log.retention.ms in Kafka/config/server.properties file to change this.

• In Node_Scripts run in order:
  • Install_dependencies.sh
    • If choosing to run InfluxDB on the same node, the option to do so is here
  • Kafka_Scripts/kafka _up.sh – This starts ZooKeeper/Kafka daemons. kafka_down.sh terminates the servers.
• Due to networking issues on POWDER, run this command: sudo ip route add 155.98.47.0/24 via 155.98.36.204

STEP 4: Spin up srsLTE

NOTE: Set your node identifier to be of the form: [node type][#] (Ex. ue1)

• For each UE
  • On the EPC node: add entries for distinct keys and IMSIs in /etc/srslte/user_db.csv
  • On each UE: change the key and IMSI to ensure there are no duplicates
• For each node (EPC/eNB/UE)
  • In Node_Scripts/ run Install_dependencies.sh
    • Select "No" for the InfluxDB installation
  • Node_Scripts/srsLTE_Scripts run srsLTE_top.sh
  • Provide the Kafka server public IP (the value for advertised.listeners in the server.properties file)

STEP 5: Setup InfluxDB for visualization

• Login to the setup page using port 8086 at the IP of the node running InfluxDB (Kafka server node by default)
• Setup basic credentials for this database (username, password, name of DB)
• In Local_Scripts/Data_Processor/Reader/influx_reader.py change the token, org, and bucket variables
  • token comes from InfluxDB GUI: Data > Token > User's Token > Copy to clipboard
  • org is what you've named the DB
  • bucket is what you've named the bucket
• In InfluxDB GUI under Boards
  • Import the UE/eNB dashboards
    • Create Dashboard > Import Dashboard
      • Dashboard JSON files are in InfluxDB/ as enb_dash.json and ue_dash.json

STEP 6: Consume/process data locally

• In Local_Scripts/Data_Processor run consumer_top.sh and specify the Kafka server IP
  • This script creates a consumer for each topic and reads in the log files