Project 14: Link Failure Detection and Recovery using SDN

1. Problem Statement

In traditional networking, link failures often lead to significant downtime and require manual reconfiguration or slow legacy protocols. This project demonstrates how Software-Defined Networking (SDN) can automate fault tolerance and path restoration. Using a Triangle Topology, the goal is to implement a system where a central POX Controller dynamically detects a link failure between two switches via OpenFlow and automatically reroutes traffic through a backup path to maintain connectivity.

---

2. Topology & Infrastructure

The network is emulated in Mininet and follows a triangle design to provide redundancy.

• Hosts: h1 (10.0.0.1), h2 (10.0.0.2)
• Switches: s1, s2, s3 (Open vSwitches)
• Controller: Remote POX Controller (127.0.0.1)

Path Logic:

• Primary Path: h1 <-> s1 <-> s2 <-> h2
• Backup Path: h1 <-> s1 <-> s3 <-> s2 <-> h2

---

3. Setup and Execution Steps

1. Initialize the POX Controller: In Terminal 1, run the controller with the necessary modules for discovery and loop prevention:

   ./pox.py forwarding.l2_learning openflow.discovery openflow.spanning_tree --hold-down=2

2. Launch the Network Topology: In Terminal 2, execute the custom Mininet script:

   sudo python3 topo.py

3. Monitor Traffic with Wireshark: Open Wireshark on the loopback (lo) interface and filter by openflow_v1 to observe the control channel.

4. Execute the Test:

   • Start a ping from h1 to h2.
   • Manually bring down the primary link using link s1 s2 down.
   • Observe the automatic recovery and traffic rerouting.

---

4. Proof of Execution

A. Connectivity & Recovery (Ping/iperf Results)

Initially, the pings show low latency. When the s1-s2 link is disabled, the pings pause briefly during the STP Convergence period and then resume automatically via the backup switch s3.

Analysis: The Round Trip Time (RTT) increases after recovery (e.g., from 0.05ms to ~10ms+) because the data packets must now travel through an additional switch (s3).

B. Flow Table Verification

We used ovs-ofctl to inspect the Flow Tables on s1 to prove the controller modified the data plane rules.

• Pre-Failure: Flow rule matches Destination IP and outputs to the port connected to s2.
• Post-Failure: Flow rule is updated by the controller to output to the port connected to s3.

C. Wireshark Logs (Link Detection)

The "Detection" phase is validated by the OFPT_PORT_STATUS message. This is an asynchronous message sent by the switch to the POX controller to report that a link is down.

---

5. Conclusion

The project successfully demonstrates the core advantage of SDN: Centralized Control Plane Logic. By separating the control logic from the physical switches, the POX controller was able to detect a link failure in real-time and push new flow rules to the switches, achieving recovery with zero manual configuration.