I have a theoretical question: What is the relation between load and response time? Is there any relation?
I am think of the following example: The response time is mostly constant for a system. But when the load gets over a critical point, the response time will increase dramatically, because of scheduling and other overhead.
Am I correct? Is there any literature about it?
thanx.
Yes, is that correct. You are talking about load sensitivity that is
“how the response time varies with the load” (folwer). It is a performance factor.
Take a look at this book “Patterns of enterprise application architecture”, there is a free chapter at:
http://books.google.com/books?id=FyWZt5DdvFkC&pg=PT35&lpg=PT35&dq=martin+folwer+scalability+load+response+time&source=bl&ots=eFxv1wUq7A&sig=H0rJicgI2JGEaOeYZEX4FGQhUd8&hl=pt-BR&ei=vqt4TpfTC4LYgQe3z5nRDQ&sa=X&oi=book_result&ct=result&resnum=1&ved=0CB0Q6AEwAA#v=onepage&q&f=false