Learning Performance Engineering through Applications

In 1986 D. Ferrari, in a position paper published on IEEE Transactions on Software Engineering, emphasized that the field of Computer Systems Performance Evaluation (CSPE) since its birth had grown in isolation from the other areas of computer science. It was pointed out that the drawbacks of such insularity outweigh its advantages. Some urgent actions to improve the situation were listed. Among them, on the educational side, the integration of performance evaluation concepts in the computer systems basic courses and a more application-oriented approach of the teaching have been advocated. Another problem highlighted was the existence of a time lag between the available performance evaluation techniques and the type of problems created by the new system architectures. In 2003 Ferrari revisited and updated the concepts described in the first paper. Some limited positive results have been perceived. But, what has happened during the last 14 years, about these and other problems related to the education and practice of performance engineering? The old problems, at least partially, have been solved? What are the new problems that arose?