NGN Network Slicing Exam

Overview

This project implements a network slicing architecture using Mininet, Ryu SDN controller, and WebSocket-based communication for dynamic flow allocation and monitoring. The system supports both basic and slicing modes for traffic management, enabling bandwidth reservation and QoS enforcement.

---

Instructions and File Descriptions

SSH into the VM

To connect to the VM (password: vagrant):

ssh -X -p 2222 vagrant@localhost

---

Topology

The file topology.py defines the network topology, which is dynamically loaded from a YAML file (topology.yaml). The topology includes hosts, switches, and links with configurable bandwidth.

To run the topology in Mininet:

sudo python3 topology.py

This will:

• Load the topology from topology.yaml.
• Save host MAC addresses and switch link information to /tmp/host_info.json and /tmp/switch_links_info.json.
• Start a WebSocket server for Mininet commands on ws://127.0.0.1:9876.

---

Controller

The file flow_allocator_controller.py is the Ryu-based SDN controller. It manages:

• Flow allocation and deletion.
• Bandwidth reservation using QoS queues.
• Dynamic topology updates.

To start the controller, use the provided shell script:

sudo ./start_controller.sh

This will also start a WebSocket server for flow management on ws://127.0.0.1:8765.

---

Automatic Allocation

Run the tester.py script to allocate flows interactively:

sudo python3 tester.py

This script:

• Reads host MAC addresses from /tmp/host_info.json.
• Allows you to select source and destination hosts.
• Sends WebSocket requests to allocate flows.

---

WebSocket Communication

The system uses WebSocket servers for communication:

• Mininet WebSocket Server (ws://127.0.0.1:9876): Executes commands on Mininet hosts.
• Controller WebSocket Server (ws://127.0.0.1:8765): Handles flow allocation, deletion, and monitoring.

---

Traffic Testing

The system supports two modes for traffic testing, configured via the .env file:

TEST_MODE=slicing
# TEST_MODE=basic

Slicing Mode

In slicing mode, traffic is managed with bandwidth reservations. To test slicing:

1. Run the topology:

   sudo python3 topology.py

2. Start the controller

   ./start_controller.sh

3. Allocate flows using the tester.py script:

   sudo python3 tester.py

4. Run the pre-built iperf command to generate traffic in the tester.py.

5. View throughput graphs generated by graph_mesh_slice.py.

Basic Mode

In basic mode, traffic is managed without slicing. Follow the same steps as slicing mode but set TEST_MODE=basic in the .env file. Graphs will be generated by graph_mesh_basic.py.

---

Path Finder

The file path_finder.py implements the Widest Path algorithm for routing traffic between hosts. Instead of seeking the shortest path, it finds the path where the minimum bandwidth link (the bottleneck) has the maximum possible capacity - essentially selecting the path where the weakest link is as strong as possible. The path finder integrates with the controller to enforce appropriate QoS policies based on the selected path.

---

Monitoring and Visualization

• Throughput Graphs: Generated by graph_mesh_slice.py or graph_mesh_basic.py and saved in the netbench directory.
• Flow Reservations: View active reservations using the show command in the CLI.

---

CLI Commands

The CLI provides the following commands:

Command	Description
allocate	Allocate a new flow between two hosts.
delete	Delete an existing flow.
dump	Dump flows from a specific switch.
show	Show the flow reservation table.
ping	Test connectivity between two hosts.
iperf	Run iperf tests based on the TEST_MODE.
help	Show the help menu.
clear	Clear the screen.
exit	Exit the CLI.

---

Notes

• Ensure Mininet and the Ryu controller are running before testing.
• Open at least 3 terminal sessions for running the topology, controller, and CLI.
• Clean up QoS configurations using the clean_ovs_qos() function in topology.py if needed.
• If everything seems broken or your topology crashed, then just run a cleanup and you will be safe 😄 :

  sudo mn -c

All set! 😄