The main goal of the article is to make centralized databases run faster by figuring out the best way to answer questions (queries) asked by users. They do this by comparing different ways to find the answers and picking the one that uses the least amount of time and computer resources. The key idea is to estimate how much work each way will take, mainly focusing on how many times the computer needs to read from the disk. By picking the most efficient method, it helps save time and resources when dealing with lots of data stored in one central place.