Distributed Hash Table

# Distributed Hash Table PRD

## Overview
Build a Kademlia-style distributed hash table that uses XOR distance for routing, k-buckets for peer management, iterative node lookups for key discovery, and replication for fault-tolerant decentralized key-value storage across a network of nodes.

## Requirements
- 160-bit node IDs generated by SHA-1 hashing of the node's address
- XOR distance metric for determining closeness between node IDs and keys
- k-buckets routing table: k=20, 160 buckets organized by XOR distance prefix
- Node lookup: recursive lookup with alpha=3 parallel queries to find closest nodes
- Key-value store operations: PUT, GET, and DELETE
- Key replication: store values on the k closest nodes to the key for redundancy
- Iterative node lookup: find the k closest nodes to any target ID
- Kademlia RPCs: PING, STORE, FIND_NODE, FIND_VALUE
- UDP or TCP transport for inter-node communication
- Node bootstrap: join the network via a known bootstrap node
- Routing table refresh: periodic random lookups to keep buckets populated
- Stale node eviction: ping nodes before removing them from buckets
- Key republishing: periodic re-storage to maintain replication as nodes join and leave
- Node join and leave handling with routing table updates
- Handle network churn: nodes joining and leaving dynamically
- Concurrent request handling for parallel RPCs
- CLI for node operations: start, join, put, get, info, and peers
- Metrics: routing table size, stored key count, and RPC counts
- Simulate a 10+ node network locally for testing

## Tech Stack
- TypeScript / Node.js
- dgram (UDP) or net (TCP) module for networking
- No external DHT or distributed systems libraries — raw implementation required

## Scoring Criteria
- **Functional (40%)**: Nodes discover each other, keys store and retrieve across the network, lookups converge
- **Quality (20%)**: Correct XOR routing, proper k-bucket management, resilient to node failures
- **Fidelity (25%)**: All features including replication, refresh, eviction, churn handling, and metrics
- **Speed (15%)**: Time bonus