Enterprise wide applications such as a BPM workflow require a distributed workflow architecture in order for the workflow engine to reach peak performance.

Distributed Workflow Architecture

Distributed Workflow Architecture

In companies around the world there are BPM application performing as they were designed to do yet their outputs vary greatly.  Facts are that as BPM applications process workflow instances, some process a few while others process very great quantities of instances.

While there can be many contributing factors, today I would like to focus on the hardware part of the equation.  Servers like any other piece of equipment struggle to perform efficiently as the load increases.  As the server struggles the applications running on that server or communicating with it tend to falter due to the lack of data processing resources.  With enterprise wide applications such as a BPM workflow application, it is critical for peak performance that the workflow engine be fully distributed.  Otherwise, the BPM application will continue to increase workflow processing steps until system failure due to lack of resources.

Below is a reporting from an experiment we performed for a client.  There are a series of graph to illustrate data taken from a 2 Quad core CPU, 16 GB Ram server.  It hosts the application and also executes workflows generated by the application.  The graph below shows data taken from a 25 hour period of time.

Distributed Workflow Architecture

Single Server Processed Instances Distributed Workflow Architecture

With cDevWorkflow’s engine, workflow processing can be easily distributed and run on separate servers to balance the load.  By distributing workflow processing to a separate server, there is a significant increase in the number of instances that can be processed.

distributed workflow architecture

Two Servers Processed Instances Distributed Workflow Architecture

With even larger installations, more servers can be added to further distribute workflow processing loads.  With 2 servers to process the workflow instances, here is what the data looks like:

distributed workflow architecture

Three Servers Processed Instances Distributed Workflow Architecture

Displayed with the information graphed above we increased the number of servers to 2 servers, the number of processed instances has increased substantially.  Given today’s server infrastructure, whether its physical servers, virtual machines or cloud machines, any number of servers can be configured with cDevWorkflow to distribute and process the load.

A small sampling of the Workflow Technology for BPM Solutions contained within our cDevWorkflow product offering can be found here: Workflow Technology That Works.

Learn more about our Process Automation & IT, QA Services or Software Development products and solutions on the Web!  Visit us at: Innovative Process Solutions

Distributed Workflow Architecture Distributed Workflow Architecture Distributed Workflow Architecture Distributed Workflow Architecture Distributed Workflow Architecture Distributed Workflow Architecture Distributed Workflow Architecture Distributed Workflow Architecture