Once implementation is available, wherever appropriate, replace futures::lock::Mutex with futures::lock::RwLock

[jstuczyn]

Basically wait a week or so for 0.6.0 release of https://github.com/asomers/futures-locks that will use tokio 0.2.0.
Why we want to change it to RwLock rather than Mutex? As there are way many more reads of the data compared to writes. It is read every single client request (for EVERY client) while it is written to only once per single client - during registration. So that would be a significant performance gain.

Edit: This issue refers to uses in sfw-provider as well as in nym-client

[jstuczyn]

Or just replace it with tokio::sync::RwLock
Also, use tokio::sync::Mutex in place of futures::lock::Mutex for consistency sake.

[jstuczyn]

Future note to whoever is going to be implementing this. tokio::sync::RwLock (from what I've tried) does not seem to have an easy access to poll method - it might be problematic if for some reason we wanted that. It might be worth to instead wait for proper futures::lock::RwLock (as with their current futures::lock::Mutex