kademlia-table v1.0.8
kademlia-table
XOR distance based routing table used for P2P networks such as a Kademlia DHT.
npm install kademlia-table
An extendable implementation of Kademlia and K-Buckets closely following details set out in the Kademlia DHT paper.
Usage
import { KademliaTable } from "kademlia-table";
import { randomBytes } from "node:crypto";
interface Node {
id: string;
// ...properties of Node
}
// Create a new table that stores nodes "close" to the passed in id.
// The id should be uniformily distributed, ie a hash, random bytes etc.
const table = new KademliaTable<'id', Node>(id(), { idKey: 'id' });
// Add a node to the routing table
table.add({ id: id() });
// Get the 20 nodes "closest" to a passed in id
const closest = table.closest(id(), 20);
Extending The Table
The base class is extendable to add custom logic to operations. For example you may want to act on a bucket if it is full when attempting to add a node, such as comparing response times of nodes.
import { KademliaTable } from "kademlia-table";
interface Node {
id: string;
// ...properties of Node
}
class CustomTable extends KademliaTable<'id', Node> {
add(node: Node): boolean {
const result = super.add(node);
if (!result) {
// Do something if bucket is full and retry
return this.add(node);
}
return true;
}
}
API
table = new KademliaTable(id, [configuration])
Create a new routing table.
id
should be a string that is uniformily distributed. configuration
includes:
{
idKey: 'id' // Key used as id
encoding?: "utf8" // Encoding of id strings
bucketSize?: 20 // Max number of nodes in a bucket
}
bool = table.add(node)
Insert a new node. node[idKey]
must be a string of same or shorter length as table[idKey]
.
When inserting a node the XOR distance between the node and the tableidKey is
calculated and used to figure which bucket this node should be inserted into.
Returns true
if the node could be added or already exists.
Returns false
if the bucket is full.
bool = table.has(id)
Returns true
if a node exists for the passed in id
and false
otherwise.
node = table.get(id)
Returns a node or undefined
if not found.
i = table.getBucketIndex(id)
Returns a node's corresponding bucket index.
nodes = table.closest(id, [maxNodes])
Returns an array of the closest (in XOR distance) nodes to the passed in id.
This method is normally used in a routing context, i.e. figuring out which nodes in a DHT should store a value based on its id.
true = table.remove(id)
Remove a node using its id.
table.nodes
Returns all nodes from table as an array. Ordered from closest to furthest buckets.
table.buckets
A fixed size array of all buckets in the table.
number = KademliaTable.getDistance(idA, idB, encoding)
Gets the XOR distance between two id strings.
1 | -1 | 0 = compare(idA, idB) = KademliaTable.createCompare(targetId, encoding)
Creates a function for sorting ids based on distance from a target id going from closest to furthest
License
MIT