This was a long chapter! We looked at Celery as a powerful package for writing distributed applications in Python. We then looked at Python-RQ, a lightweight and simpler alternative. Both packages use a distributed task queue architecture, which is a multimachine implementation of the same system that is used to distribute work that we saw in Chapter 3, Parallelism in Python.

Pyro was then introduced as an alternative approach to both Celery and Python-RQ. Pyro has a very different philosophy that is firmly rooted in the proxy pattern and remote-procedure-call (RPC) architecture for distributed systems.

Both approaches have their merits and their strengths, and undoubtedly you will find yourselves preferring one or the other.

The next chapter will look at one way to deploy our distributed applications to the cloud—it is going to be a fascinating read.