isaaczafuta / pydht

Python implementation of the Kademlia DHT data store

pydht

Python implementation of the Kademlia DHT data store.

Useful for distributing a key-value store in a decentralized manner.

To create a new DHT swarm, just call DHT() with the host and port that you will listen on. To join an existing DHT swarm, also provide bootstrap host and port numbers of any existing node. The nodes will discover the rest of the swarm as appropriate during usage.

Example - A two-node DHT:

>>> from pydht import DHT
>>> host1, port1 = 'localhost', 3000
>>> dht1 = DHT(host1, port1)
>>>
>>> host2, port2 = 'localhost', 3001
>>> dht2 = DHT(host2, port2, boot_host=host1, boot_port=port1)
>>>
>>> dht1["my_key"] = [u"My", u"json-serializable", u"Object"]
>>>
>>> print dht2["my_key"]
[u'My', u'json-serializable', u'Object']

pydht's Issues

Successive key retrieval error on 2 node DHT

After storing the key in the DHT (a two node DHT with one the bootstrap node), retrieving the key randomly raises KeyError. For example, if you store a key and then run a program like 5 - 20 times to retrieve that same key it will randomly fail. Why would that be? Is the DHT not 100% reliable / not redistributing keys?

(I'm not an expert on DHTs so maybe this is normal but worth checking out.)

repo description is misleading

The description “Python implementation of the Kademlia DHT data store” is misleading. This is not a compliant implementation.

Please share any docs or examples if available

I really want to work on this project. i am very much interested in this but I dont know how to use it. Please provide any examples or docs if available.

Bootstraping nodes

Hi, thanks for you work.
Does your library support bootstraping nodes from a "tracker"?
I saw in example that you connect directly to a known and already existing node, but what if we don't know the node?