For concurrent access to the history store, we are possibly moving towards using a pool of cursors/resources. We already have a pool of LAS cursors at the moment which can be further enhanced.
This ticket will address the following:

• Understand the existing design for LAS cursor pool access.
• Given that History Store (HS) will not have a sweep and hence will not need exclusive write access, with that constraint gone, figure out improvements we can do to the pool.
• Explore using atomics to control the access to the pool rather than locks.