P2P File Sharing System

A peer-to-peer file sharing system built in C++, similar to BitTorrent. Peers can share and download files directly from each other. A central server acts as a tracker — it only stores metadata (who has what), never the actual file data.

Architecture

[Server - Tracker]       ← maintains file registry only
      ↑        ↑
  [Peer 1]  [Peer 2]     ← transfer files directly between each other

• Server: Tracks which peers have which files/chunks. Stores nothing on disk.
• Client (Peer): Shares local files and downloads files directly from other peers.

How It Works

1. A peer registers its files with the server (share command).
2. Another peer asks the server for available files (show command).
3. The server responds with peer IPs and chunk locations.
4. The downloading peer connects directly to source peers and downloads chunks in parallel.
5. Chunks are reassembled into the final file locally.

Files are split into 1024-byte chunks. Different chunks can be downloaded simultaneously from different peers, improving speed.

Requirements

• g++ with C++11 support
• POSIX-compatible OS (Linux / macOS)

Build

git clone <repo-url>
cd p2p-file-sharing
make

This produces two binaries: server and client.

Running

Start the server (Terminal 1)

./server

Output: Server started on port 8080

Start a peer (Terminal 2)

./client 8082

Start another peer (Terminal 3)

./client 8083

Each peer must use a different port number.

Peer Commands

Command	Description
show	List all files available to download
share N file1 file2 ...	Share N files from the current directory
download filename	Download a file from other peers

Example

Peer 1 — share a file:

share 1 notes.pdf

Peer 2 — list and download:

show
download notes.pdf

Notes

• The file you want to share must be in the same directory from which you run ./client.
• Server data is in-memory only — restarting the server clears all registered files. Peers must re-register their files after a server restart.
• Downloaded files are saved to disk permanently and are not affected by server restarts.
• If a download is interrupted, a .log file tracks completed chunks so the download can resume.

Deploying on Multiple Machines

Edit SERVER_IP in both server.hpp and peers.hpp:

#define SERVER_IP "your.server.ip"

Then recompile and run ./server on the server machine and ./client <port> on each peer machine.

Peers must have publicly reachable IPs (or be on the same LAN) since file transfers are direct peer-to-peer — they do not go through the server.

Protocol

Request	Description
Register Request	Peer tells server which files it wants to share
File List Request	Peer asks server for all available files
File Locations Request	Peer asks server for IPs/ports of peers that have a specific file
Chunk Register Request	Peer notifies server after receiving a new chunk
File Chunk Request	Peer requests a specific chunk directly from another peer

Cleanup

make clean