Tradeoffs when integrating multiple software components into a highly available application

We analyze nineteen months of field trouble reports for a particular highly available software application. We use this data to understand when it makes sense for the creator of a highly available software application to buy commercial software subcomponents rather than building them in-house. In particular we identify a number of potential negative impacts on system operability which can occur if too many or inappropriately chosen commercial software components are used. Our data indicates that the prudent application builder must carefully weigh these potential negative operability impacts against the well known potential cost savings that can accrue from using commercial software components.