This is a very good debate, especially if being fought between a Basis/DBA and an ABAP/functional Consultant.
Practically tuning is a thankless job, people (Business/Functional Consultant) remember you if and only if a program is running since hours. And the program comes back on the tuning anvil.
Now when a program comes for tuning what are the dimensions it do need to be checked and rectified for performance? Whether it is the resource utilization by the program or the time taken by the program? This is something where I feel that the resource utilization of the program should be considered as the prime factor.
But as a common practice, a program comes back to tuning Lab when the program is identified as long time taking. So the normal inclination and pressure is always on the version which takes shorter time, even if the new version is more resource hungry and causes resource crisis for other programs who are also getting executed on the same time.
Some of my senior, friends and colleagues always are in favour of the version which consumes less time,,,,
And functional/business consultant always does so because their SLA gets hampered every time the new tuned (so-called) version of the program is returned back calling it inefficient/potentially dangerous for production environment.
They always says that if someone has the money to buy enough fuel it is better to buy a car with high pickup power. But always they forget that they need to drive the car on a road, not on a racing track and running high pickup/power car (requiring more fuel) not only consumes more fuel but also risk lives.
So, not the expertise in running different performance tool neither an in-depth knowledge of the technology, but changing the mindset of the business/functional people is the challenge of a system tuning expert.
This is my first post in year 2010, happy new decade to all…..