http://lzteddy.com/tags/udp/Recent content in UDP on Zitao LiaoHugo -- gohugo.ioen-usTue, 19 May 2026 00:00:00 +0000http://lzteddy.com/p/cs305-p2p-file-transfer/Fri, 19 Dec 2025 00:00:00 +0000http://lzteddy.com/p/cs305-p2p-file-transfer/<h2 id="background">Background </h2><ul> <li>Time: Dec 2025</li> <li>Context: Group course project for SUSTech CS305 (Computer Networks).</li> <li>Repository: <a class="link" href="https://github.com/Tsurumalu/Computer-Network-Project-P2P" target="_blank" rel="noopener" >Tsurumalu/Computer-Network-Project-P2P</a></li> </ul> <h2 id="goal">Goal </h2><p>Build a <strong>reliable peer-to-peer (P2P) file transfer application</strong> on top of <strong>UDP</strong>. All transport semantics—handshaking, reliable delivery, retransmission, and congestion control—are implemented at the application layer. The system uses a network simulator for testing under packet loss, varying topologies, and peer crashes.</p> <h2 id="system-overview">System Overview </h2><ul> <li><strong>File &amp; chunk model:</strong> Files are split into 512 KiB chunks; each chunk is identified by a SHA-1 hash (20 bytes).</li> <li><strong>Peer model:</strong> Each peer holds a fragment (subset) of chunks and can upload to or download from other peers concurrently.</li> <li><strong>Download flow:</strong> On a <code>DOWNLOAD</code> command, a peer discovers missing chunks via <code>WHOHAS</code> / <code>IHAVE</code>, requests them with <code>GET</code>, then reassembles completed chunks into a pickle-serialized fragment file.</li> <li><strong>Concurrency:</strong> Single-threaded event loop using <code>select</code>; multiple chunk downloads from different peers run in parallel, with per-session upload limits enforced via <code>DENIED</code>.</li> </ul> <h2 id="packet-protocol">Packet Protocol </h2><p>Custom UDP packets (max 1400 bytes) with header fields for type, length, and sequence/ACK number:</p> <table> <thead> <tr> <th>Type</th> <th>Code</th> <th>Role</th> </tr> </thead> <tbody> <tr> <td>WHOHAS</td> <td>0</td> <td>Query which peers hold requested chunks</td> </tr> <tr> <td>IHAVE</td> <td>1</td> <td>Reply listing available chunks</td> </tr> <tr> <td>GET</td> <td>2</td> <td>Request a specific chunk</td> </tr> <tr> <td>DATA</td> <td>3</td> <td>Transfer chunk payload segments</td> </tr> <tr> <td>ACK</td> <td>4</td> <td>Acknowledge received segments</td> </tr> <tr> <td>DENIED</td> <td>5</td> <td>Reject when upload connection limit is reached</td> </tr> </tbody> </table> <h2 id="implementation-highlights">Implementation Highlights </h2><ul> <li><strong>Reliable data transfer (RDT):</strong> Per-packet sequence numbers and ACKs; dynamic timeout from RTT estimation (α = 0.15, β = 0.3); timeout-based and fast retransmit (3 duplicate ACKs).</li> <li><strong>Congestion control (Tahoe-style):</strong> Slow start and congestion avoidance with <code>cwnd</code> and <code>ssthresh</code>; loss triggers window halving and return to slow start. <code>cwnd</code> history is recorded and plotted with matplotlib for presentation.</li> <li><strong>Structured codebase:</strong> Refactored <code>peer.py</code> with <code>PeerState</code>, <code>DownloadSession</code>, <code>UploadSession</code>, and a <code>Peer</code> class driving a non-blocking I/O loop.</li> <li><strong>Robustness:</strong> Handles concurrent multi-peer downloads, connection limits, peer crashes, and out-of-order segment buffering.</li> <li><strong>Testing:</strong> Validated with course-provided pytest suites (handshaking, basic transfer, concurrency, crash scenarios, and advanced topology tests) via <code>grader.py</code> and the network simulator.</li> </ul> <h2 id="tech-stack">Tech Stack </h2><ul> <li>Python 3.12</li> <li>UDP (<code>socket</code>) with <code>select</code> for single-threaded concurrency</li> <li>SHA-1 chunk hashing</li> <li>matplotlib (congestion window visualization)</li> <li>pytest &amp; network simulator (grading and integration tests)</li> <li>Linux / WSL (required runtime environment)</li> </ul>