deserialise an object from a path file
Make a local filesystem path based on a 'tag' to temporarily store the serialised object.
Get an object that has been distributed so as to be available for tasks in the current job.
Distribute an object to be available for tasks in the current job.
serialise an object to a path
Faciliate making an object available to all tasks (mappers, reducers, etc). Use XStream to serialise objects to XML strings and then send out via Hadoop's distributed cache. Two APIs are provided for pushing and pulling objects.