docker-nextepc

Extension of this original repository as an extension of this initial repository.

Any credit should be given to the original authors.

Disclaimer: Due to some hardware related inconviniences this project has not been properly tested and it is very probable that it has some issues related to the authomatic attachment of UEs to eNBs. Given that case, some proposals are described below.

How to run (On Linux)

Make sure git and docker-compose are installed

sudo apt-get install -y git docker-compose

Start by cloning this Github repository and moving inside the repository folder

sudo git clone https://github.com/Joanguitar/docker-nextepc
cd docker-nextepc

Build the docker

sudo docker-compose build

Launch the docker

sudo docker-compose up -d

At this point everything should be up and running.

Once you're done, stop the docker or the logging files will plage your disk

sudo docker-compose down

Things you can do

I compiled a list of usefull scripts inside the commands folder that you can access as cd commands from the main git folder.

To access the bash terminal of a container you can run

sudo bash access.sh <container>

The simulation container names are ue1, ue2, enb, hss, pgw, sgw, mme, pcrf. This will allow you to run useful commands for networking experimentation like ifconfig, ping, iperf3 or tshark for example.

You can simultaneouly capture all interesting traffic using

sudo bash capture_traffic.sh <time>

The variable stands for the amount of time that's going to be captured. The captures are saved in the "capture" folder and can be openned with Wireshark.

The NextEPC GUI is running under localhost:3000 with credentials

• User: admin
• Pass: 1423

Here you can modify the subscriptions for the UEs, but be aware that most of the features shown there like traffic shaping are still not implemented.

Considerations when running in a virtual machine

If you're running this in a virtual machine, you have to be aware that this is a memory hungry docker that needs to run without delays in the process. So I recommend allocating more memory than the one by default and enabling the use of multiple cores, else you may see random disconnections of the UE. I got it running with 4GB of allocated memory and 4CPU cores without any problem.

What's different

• Added a second eNB container (enb2)
• Changed references of enb to enb1
• Changed dependency of ue2 to enb2
• Registred both UEs

Configuration

The SIM card provisioned in the virtual UE1 (from srsUE) and the EPC is using the following parameters :

• IMSI=001010000000001
• Ki=c8eba87c1074edd06885cb0486718341
• OPc=17b6c0157895bcaa1efc1cef55033f5f

The SIM card provisioned in the virtual UE2 (from srsUE) and the EPC is using the following parameters :

• IMSI=001010000000002
• Ki=c8eba87c1074edd06885cb0486718342
• OPc=17b6c0157895bcaa1efc1cef55033f6f

These parameters can be changed in run_ue1.sh and run_ue2.sh of the Virtual-UE-eNB folder.

Porposals of fixing UE attament errors

First proposal

• Add another network to the UE pool in the file nextepc.conf - addr: 45.46.0.1/16 between the lines 340 and 341.
• Duplicate the file route_ue.sh and rename the copies as route_ue1.sh and route_ue2.sh.
• Change the line 6 in route_ue2.sh to ping -i 10 45.46.0.1
• In the file docker-compose.yml, change the volumes of the containers ue1 and ue2 to make reference to route_ue1.sh and route_ue2.sh respectivelly instead of route_ue.sh.

Second proposal

• Set explicitelly the virtual radio networks in the file docker-compose.yml , as it is done in the version of (https://github.com/ravens/docker-nextepc). Make sure to create two different networks: 45.45.0.1/16 for ue1 and enb1, and 45.46.0.1/16 for ue2 and enb2.
• Change the line 6 in route_ue2.sh to ping -i 10 45.46.0.1
• In the file docker-compose.yml, change the volumes of the containers ue1 and ue2 to make reference to route_ue1.sh and route_ue2.sh respectivelly instead of route_ue.sh.