Software transactional memory

Introduction

Software Transactional Memory, or STM, is an abstraction for concurrent communication. The main benefits of STM are composability and modularity. That is, using STM you can write concurrent abstractions that can be easily composed with any other abstraction built using STM, without exposing the details of how your abstraction ensures safety. This is typically not the case with other forms of concurrent communication, such as locks or MVars.

STM for Haskell was first implemented in GHC 6.4. There exist STM implementations for other compilers.

Resources

• Beautiful Concurrency - a chapter from the book Beautiful Code

• Research Papers on STM

• Documentation for the Control.Concurrent.STM primitives GHC

• Documentation for the Control.Concurrent.STM module in GHC

• Programming in the Age of Concurrency: Software Transactional Memory - interview of Simon Peyton-Jones and Tim Harris.

Projects using STM

• Conjure (a bittorrent client) is using STM quite heavily.

• The HAppS MACID monad runs on top of STM. It is open source and available via darcs at http://happs.org. http://Pass.net is a production application that uses this infrastructure.

• Shellac is a framework for building read-eval-print style shells. It uses STM, but in a pretty limited way.

• Pugs (an implementation of Perl 6 in Haskell) uses STM, and also exposes STM to the Perl 6 programmer via a similar interface to STM Haskell.

Examples

• Simple STM example

• Graceful exit

• Advanced STM Monad

• Examples from this wiki: Search for "concurrent.stm" at the top of this page.