My colleague said that after adding a disk array, It would take 10 minutes to insert 50 thousand records cyclically. Now it takes more than 20 seconds.
There are so many differences!
Generally, when a performance problem occurs, additional hardware investment will improve the performance, which is simple, clear, and straightforward. However, this seems to be regarded as a temporary solution, and the problem still exists, and we will be despised by our technical-only people: What is the ability to rely on hardware? The real problem exists in the code! Code improvement is a good skill!
But it's a huge leap from 10 minutes to more than 20 seconds! No matter how you step on it, it is an indisputable fact that people are incapable of improving their performance. If I am the boss, the customer only needs to pay some money and add a disk array card to solve the problem perfectly. Why not? What a good solution! How long does it take to perform a code review and refactoring? Not counting the time for re-debugging, testing, and deployment! Is it better to say it?
In addition, the so-called technical improvement, such as the half bucket of water, can only sound nonsense: using stored procedures... Loop insertion, only one database connection is enabled... it doesn't matter (it turns out that my method has no effect ). I should try to find a piece of tofu and hit it to death.
Later, I barely searched for information about the disk array card. The reason why the disk array card improves performance is that it has a high-speed cache. It takes over the hard disk and loads common data to the high-speed cache for the host to read. When the host writes data, it also writes data to the high-speed cache of the array card. This means that the write is successful and then the result is returned. As a matter of fact, it was written to the hard disk slowly after the time of the array.