k-rpc
Low-level implementation of the k-rpc protocol used by the BitTorrent DHT.
npm install k-rpc
Read BEP 5 and BEP 44 for more background info.
Usage
var krpc =var rpc =var target = 'aaaabbbbccccddddeeeeffffaaaabbbbccccdddd' 'hex'// query the BitTorrent DHT to find nodes near the target bufferrpc{console}{console}
API
var rpc = krpc([options])
Create a new rpc instance. Options include
// per peer query timeout defaults to 2stimeout: 2000// an array of bootstrap nodes. defaults to the BitTorrent bootstrap nodesnodes: 'example.com:6881'// how many concurrent queries should be made. defaults to 16concurrency: 16// how big should be routing buckets be. defaults to 20.k: 20// the local node id. defaults to 20 random bytesid:// Length of ID in bytes, defaults to 20 (sha1)idLength: 20// optional k-rpc-socket instancekrpcSocket:
rpc.id
Buffer containing the local node id.
rpc.nodes
Routing table populated by running rpc.populate
. This is a k-bucket instance.
rpc.populate(target, query, [callback])
Populate the rpc.nodes
routing table with nodes discovered by looking for other peers close to our own local node id using the specified query. The internal routing table will be used for subsequent closest queries to take load of the bootstrap nodes.
// send a find_node queryrpc
You should call this method as soon as possible to spread out query load in the DHT.
Callback is called with (err, numberOfReplies)
.
rpc.closest(target, query, onreply, [callback])
Find peers close the specified target buffer whilst sending the specified query. onreply
will be called with (reply, node)
for every reply received and the callback is called with (err, totalNumberOfReplies)
.
// find peers sharing a torrent info_hashrpc{if messager && messagervalues console}
If a closest query is being executed while a population request in being run the closest query will take priority.
You can return false
from onreply to stop the query. This is useful if you are only looking for a single peer for example.
{consolereturn false}
rpc.query(node, query, callback)
Query a single node. If the node has a token it is set as a.token
in the query automatically.
Callback is called with (err, reply)
.
rpc.queryAll(nodes, query, onreply, callback)
Query multiple nodes with the same query. query.a.token
will be set as the corresponding nodes token when querying.
Callback is called with (err, numberOfReplies)
and onreply
will be caleld with (reply, node)
as the nodes reply.
rpc.destroy()
Destroy the underlying rpc socket.
rpc.on('query', query, node)
Emitted when a query is received.
rpc.response(node, query, response, [nodes], [callback])
Send a response to a node for a specific query. If you pass in an array of nodes {id: nodeId, host: someHost, port: somePort}
they will be added to the response.
rpc.error(node, query, error, [callback])
Send an error response for a query.
rpc.on('ping', oldNodes, swapNew)
Emitted when the bucket is full. Try and oldNodes
and if one
of them fails call swapNew
with that node to swap if for a newer one
License
MIT