How would you optimize concurrency and multithreading in a high-scale application at Persistent Systems?
