How do I optimize writes to an AWS MongoDB?

To summarize here, a write operation which doesn't affect another read operation can be written into a separate collection.

The situation is obvious and the reason you stated is also correct. MongoDB has collection(Database) level locking. So this is more of a design level problem.

The basic solution here would be to categorize your data which needs to be written. You need to separate out similar kind of data based on your queries and create different collections.

For example,
There are five datatypes or categories. Store them in different collections.

Datatype1 -> Collection1,
Datatype2 -> Collection2,
..
So on.

To summarize here, a write operation which doesn't affect another read operation can be written into a separate collection.

P.S : I am not a MongoDB user, but have attended workshop on it and discussed about situations like this.