The limited source model can often be used to approximate the behavior of a computer's CPU ( central processing unit). Suppose that 20 terminals (assumed to always be busy) feed the CPU. After the CPU responds to a user, the user takes an average of 80 seconds before sending another request to the CPU (this is called the think time). The CPU takes an average of two seconds to respond to any request. On average, how long will a user have to wait before the CPU acts on the user's request? How will your answer change if there are 30 terminals? What if there are 40 terminals? Of course, you must make appropriate assumptions about the exponential distribution to answer this question.