Project Crashing & Resource Levelling

Project crashing and resource leveling are two methods to speed up a project’s completion but approach the problem from different perspectives.  Project crashing becomes an option when the PM recognizes that a particular effort along the project life cycle can be sped up with the application of additional resources.  Crashing the project is the process of assigning these resources.  This methodology comes with an increased cost to the project.  Depending on the nature of the work the increased cost could become severe.

Resource leveling also speeds up the project, but approaches the issues after identifying that a particular resource is overallocated or nearly overallocated.  Managers typically understand that fatigue can reduce the overall effectiveness of an employee’s output.  A fatigued worker will often reduce both the quality and quantity of his or her output.  This reduction in output can have a significant impact on the project budget and schedule as the overallocation increases the time necessary to complete the task.  Resource leveling is the process by which overallocation is addressed.  Like project crashing it is more costly, but resource leveling is done to address a potential failure or delay due to overallocation whereas project crashing assigns more resources to minimize the time assigned to a particular effort.

Project crashing and resource leveling with regards to software development can be extremely problematic.  For years Sun Microsystems attempted to develop a new file system to be able to handle large (extremely large) data sets.  They assigned a large amount of programmers to the task in an effort to crash through some of the initial hurdles in the project to no avail.  Eventually a small group of dedicated individuals were able to work through the initial issues and work out a file system capable of handling a Zetabyte of data.  That’s 1,000,000,000 Terabytes!  They called it ZFS.

Right now that file system is responsible for making NETFLIX’s vast library function and its robustness as a tool is causing major businesses to switch to file servers running ZFS on the backend.  Sometimes crashing and leveling work, but for some innovative and specialized projects applying these techniques can be detrimental to project completion.  It’s probably best to assign these techniques to routine or near routine tasks only.