Re: Performance testing after modifying application.
The term "Performance Regression" is one of those that makes perfect sense in a theoretical context, yet to seasoned practitioners instantly raises questions. I like it for exactly that reason (there's a "prove-it" challenge) and have been using it for the past 6 months in my new company as we work with the global consultancies. It's similar to the more general "Agile Testing", which raised similar flags a year or so ago, yet today, dominates many threads and discussions and from what I could tell, StarEast this week.
Regression testing is repeated at frequent intervals, and in Agile methods, usually EVERY interval. Performance testing typically gets done a few times in a project (contrary to best practices and advice) due to lack of available trained people, appropriate test hardware and the test software needed. What we've seen with cloud testing is that the trained people can now have nearly immediate access to low cost test tools and test hardware, so building "Performance Regression" into development and operational plans isn't only feasible, it's obtainable, and soon, a standard operating practice.