The old restore supports adding prefix and suffix to the backup data. This allows us to restore a key space (say [a, b)) to a new key space (say [awesome_a, awesome_b).

This feature is useful in:

• Performance testing: We do not want to throw away existing loaded data in the cluster because loading a huge DB takes time;
• Cross-verification: Users may want to verify the restored data in the

In the current version of FASTER C++, on Linux, we use libaio for async IO handling. It is a known issue that libaio is not very efficient. Recently, io_uring is released with Linux kernel 5.1, which advertises to be a high performance aysnc IO library. It would be useful to try it and see if we can improve disk performance on Linux by replacing libaio with it.

More details for io_uring

Currently we don't have any mechanism to limit the maximum number of clients that could be handled simultaneously.

This feature should be designed properly. Here is some clue: https://redis.io/topics/clients#maximum-number-of-clients

Right now, PebblesDB uses a lot of memory for the TableCache (caching metadata) and for the bloom filters used for each sstable.

We want to add a command line option for PebblesDB which would limit the total amount of memory used by PebblesDB for the TableCache and bloom filters.

When using the specified amount of memory, preference should be given first to the table cache, and then bloom

ObservableMap can be built on top of Map. All the documentation needed to built this type is available on the documentation site.

I do not have any concrete requirements on how the APIs would look like but Observable types should make it simple to attach listeners and react to changes in the database.

Another option

We have a lot of async code now, but are still using some futures, converting them all (or as much as possible) would make the code clearer.

We currently create an iterator count the number of entries to returns to compute the number of entries a database stores.

There is a better way that is not O(n) which is using the num_entries field of the Stat struct that is returned by [the mdb_stat function](https://