KadDHT

KadDHT is a Kademlia-based distributed hash table used by certain peer-to-peer networks to locate resources without a central index. The most notable usage is in the eDonkey network family, including eMule and aMule, where KadDHT provides decentralized file discovery and peer lookup.

In KadDHT, each participating node has a unique identifier mapped in a large key space. Nodes maintain a routing table composed of k-buckets that cover the id space, using the XOR metric to measure distance. To locate a file or sources, a node issues a find_node or find_value operation to peers, gradually narrowing the search to closer nodes until the requested information is found or a value is retrieved.

Data in KadDHT is stored under keys derived from file identifiers; the stored values typically include information about sources (peers hosting the file) and may be time-limited. Users can store values with a TTL; lookups can be recursive or iterative depending on implementation. KadDHT relies on bootstrap nodes to join the network, and nodes periodically refresh contacts to maintain routing information.

Benefits include resilience to server outages and censorship, improved scalability, and decentralized discovery. Limitations include churn sensitivity, potential for routing/table pollution, lack of built-in cryptographic validation of stored data, and privacy concerns due to source lists being discoverable. NAT traversal can also complicate participation for some nodes.

The KadDHT concept has influenced other Kademlia-like DHTs in P2P networks, though specific implementations vary. It is commonly described as an efficiency-oriented DHT designed for file-sharing workloads.