PastSet is a tuple-based distributed shared memory system
pyPastSet is a Python implementation of the PastSet tuple-based distributed shared memory and computing system.
In some ways PastSet is similar to the classic Linda tuplespace, but with some significant differences. In PastSet, tuples are generated dynamically based on tuple templates that may also be generated dynamically. Each set of tuples based on identical templates is denoted an element of PastSet. An element may be seen as representing a trace of interprocess communications in the multidimensional space spawned by the tuple template.
In effect, PastSet keeps a sequentially ordered log of all tuples of the same or identical templates that have existed in the system. This also allows the processes to re-read previously read tuples. It is the intention that the added semantics of PastSet will allow programmers to more easily create parallel programs that are not limited to the traditional ‘bag of tasks’ type.
pyPastSet implements the PastSet tuple memory and process distribution model using the powerful Pyro distributed object framework for the core communication and SSH for secure remote process spawning.