Paper Details

Communicating Process Architectures (CPA)
 Title: Exception Handling and Checkpointing in CSP
 Conference: Communicating Process Architectures 2012
 Authors: Mads Ohm Larsena, Brian Vinterb
(a) Department of Computer Science, University of Copenhagen
(b) Niels Bohr Institute, University of Copenhagen
 Abstract: This paper describes work in progress. It presents a new way of looking at some of the basics of CSP. The primary contributions is exception handling and checkpointing of processes and the ability to roll back to a known checkpoint. Channels are discussed as communication events which is monitored by a supervisor process. The supervisor process is also used to formalise poison and retire events. Exception handling and checkpointing are used as means of recovering from an catastrophe. The supervisor process is central to checkpointing and recovery as well. Three different kind of exception handling is discussed: fail-stop, retire-like fail-stop, and check pointing. Fail-stop works like poison, and retire-like fail-stop works like retire. Checkpointing works by telling the supervisor process to roll back both participants in a communication event, to a state immediately after their last successful communication. Only fail-stop exceptions have been implemented in PyCSP to this point. 

BibTeX Entry

Full paper