distributed-router-sim-Multithreaded-

A toy project simulating a distributed routing protocol using multithreading and TCP sockets. Built for learning, but actually quite fun.

A multithreaded routing protocol simulator written in raw Python, simulating dynamic network topologies, inter-node message passing via TCP sockets, and routing updates using the Bellman-Ford algorithm.
Originally built for a systems programming course (COMP3221, University of Sydney), later extended and hacked around — this is not production code, just an overengineered toy.

🚀 Features

• 🧅 Multi-threaded message handling using threading.Thread, Lock, Event, and Queue
• 🌐 TCP socket-based communication between neighbors
• 🛡️ Periodic routing broadcasts (configurable intervals)
• ⚙️ Interactive command support:
  • FAIL, RECOVER, CHANGE: simulate network changes
  • MERGE, SPLIT: mutate the topology (experimental)
  • QUERY, QUERY PATH: view current shortest paths
  • CYCLE DETECT: detect routing loops
  • RESET, BATCH: reset state or run scripted input

📁 Project Structure

.
├── main.py             # Core simulation engine and threads
├── Routing.sh          # Launches multiple nodes
├── test.sh             # Start nodes with necessary command
├── A.txt – F.txt       # Network config files per node

Each .txt file defines neighbors in the form:

<neighbor node> <cost> <port>

Example (A.txt):

2
B 1.0 6001
C 2.0 6002

🛠 How to Run

To run all nodes at once (A–F) with demo setting:

./test.sh

Or manually run a single node:

python3 main.py A 6000 A.txt 2 1
python3 main.py <node name> <port> <config file name> <seconds before initial routing print> <broadcast update interval>

Where:

• A = node ID
• 6000 = this node's TCP port
• A.txt = neighbor config file
• 2 = seconds before initial routing print
• 1 = broadcast update interval in seconds

🧪 Sample Commands

You can input commands directly into a node terminal:

QUERY E
QUERY PATH A F
CHANGE C 9
FAIL D
RECOVER D
MERGE B C
CYCLE DETECT
RESET

Challenses that I solved during the projects:

• Race conditions, dead locks, dead locks caused by recursively calling functions, lots lots of error handling for multi-thread and socket system.
• Race condition caused by socket not yet start --need to wait until thread been ready by calling Event Ready Check.
• This is not a production system — it has race conditions, shared state soup, and some commands like MERGE/SPLIT are partially functional.
• Sockets is not reusable after failure.
• All threading coordination (including deadlock resolution) is manual.
• But hey — it works. Mostly.

📚 Academic Context

Originally written for COMP3221 (Computer Systems Programming) at the University of Sydney, but the final version was heavily modified for fun, exploration, and late-night debugging pain.

👤 Author

Aaron Zhang