BitTorrent Protocol

Simulation with MPI

This project simulates the BitTorrent protocol for P2P file sharing using MPI. The simulation captures core BitTorrent mechanics, distributing file segments across multiple nodes, allowing efficient and decentralized file sharing without relying on a central server.

Overview

Component	Description
Tracker	Acts as a central node, managing file segment locations across the network and guiding peers.
Clients	Download and upload file segments with other clients (coperative data-sharing network).

Core Logic

1. Tracker Responsibilities

The tracker (MPI rank 0) coordinates data exchange but does not store file data itself:

• File Segment Indexing: Maintains a record of segment locations across peers.
• Peer Connections: Provides peers holding requested segments when clients inquire.
• Segment Updates: Tracks segment availability updates from clients, ensuring current data.

This setup allows the tracker to efficiently connect clients, minimizing redundant downloads.

2. Client Role and Threads

Each client operates with two active threads to balance its dual role as data consumer and provider:

Thread	Description
Download Thread	Manages segment requests from peers and builds the file.
Upload Thread	Responds to requests from other clients, sharing owned segments.

This thread setup balances network traffic by allowing each client to both download and upload data.

3. Downloading Mechanism

The download process optimizes network efficiency:

• Segment Requests: Clients check with the tracker for peers holding needed segments.
• Non-Sequential Retrieval: Clients download available segments first, minimizing network wait times.
• Load Balancing: By varying peer sources, clients distribute load evenly across the network.

4. Updating the Tracker

Clients periodically report their segment updates to the tracker to maintain real-time segment availability across the network.

Update Step	Description
Status Reporting	Clients report newly acquired segments to the tracker.
Network Updates	The tracker informs other clients of updated segment availability.

5. Completion and Network Persistence

After completing their downloads, clients remain active as seeds, helping maintain data availability for others:

Completion Step	Description
Seeding	Completed clients serve segments to new peers, sustaining availability.
Shutdown	The tracker signals when all clients have completed downloads, allowing an orderly exit.

Fault Tolerance and Efficiency

The simulation includes several strategies to ensure efficient and resilient data sharing:

Feature	Description
Redundant Connections	Multiple peers are accessible for each segment, preventing interruptions.
Load Balancing	Distributes requests, avoiding overload on any single client.
Non-Sequential Retrieval	Maintains high transfer speeds without waiting on specific peers.

Simulation Constraints and Simplifications

The simulation abstracts certain real-world complexities:

Simplification	Description
Simulated Data Transfer	Segments represented as hash values, focusing on protocol, not file transfer.
Controlled Environment	Peer requests managed via MPI messages, simplifying protocol behavior testing.

Benefits of MPI for Simulation

• Parallel Processing: Multiple clients operate simultaneously, simulating real-time P2P interactions.
• Scalability: MPI’s process-based architecture supports scaling of client interactions.
• Structured Testing: Simplified testing of protocol operations and validations in a controlled environment.