sample-dht is a distributed hash table that consists of two components a client which users interact with and 1 or more nodeservers.
The nodeservers are started up as follows:
python dhtnodeserver.py <port>
where port is the port number that you choose to run a nodeserver on.
Any number of nodeserver instances can be started up on a host system.
Clients can now persist and retrieve key value pairs to these nodes as they see fit on put the client retains a key node mapping enabling it to successfully
track what node stores a given key without multiple hops.
A client is instantiated with a list of one or more nodeservers that it interacts with. This can be expressed as follows in code:
client = DHTClient([{"host": "localhost", "port": 2345}, {"host": "localhost", "port": 2346}])
Once a client instance is retrieved users can now call put to persist and getKey to retrieve.
The following shows a sample of how to persist a key value pair
result = client.put("blahKey2","sample value 2")
The following shows a sample of how to retrieve the value of a key
result = client.getKey("blahKey2")
A log file showing the operations of the client and nodeservers can be found at /tmp/dht.log
To be production ready the following non-exhaustive list of items are required:
- A suitable key distribution algorithm to ensure that nodes are not overloaded and to minimize traffic.
- The nodeserver implementation should also implement some LRU caching mechanism to ensure that storage for kv pairs can be reclaimed after
some pre-configured time.
To build a distributable for your environment the following command can be run:
python setup.py bdist
This file was modified by PyCharm 1.2.1 for binding GitHub repository