Locks on objects are needed to guarantee the consistency of the data.
However, locks need memory and with too many locks, we can get a lock escalation which will prevent other sessions from doing their work because they will go blocked.

What if we could have fewer locks being held even for larger transactions? This would allow more sessions to access, concurrently, the table.

Now, imagine this is turned on by default on Azure SQL DB.

Let's take a look at the optimized locking feature and how that differs from what we are used to.

Is it any good? Or bad? Maybe it depends!
You need to join me in this session to find out.