This language-agnostic proposal focuses upon concepts and strategies critical to the design and implementation of asynchronous systems and data processing layers. Key components include a survey of implementation strategies for non-blocking edge tiers, patterns for building out a distributed worker / processing tier, along with several horror stories of cascading failures and their resolution.
With systems such as Grid Engine, Condor and others, it is relatively easy these days for organizations to create robust distributed compute farms. See how the Grid::Request Perl module can make the authoring, submission and control of large distributed jobs easy and in a scheduler agnostic manner.
Many people view topics like Map/Reduce and queue systems as advanced concepts that require in-depth knowledge and time consuming software setup. Gearman is changing all that by making this barrier to entry as low as possible with an open source, distributed job queuing system.