Metrics - Latency / Throughput

Experiments -

1. One PUT request at a time from one client
2. Multiple PUT requests at a time from different clients 
(assuming that one client's requests are serialized)

Similarly as above
3. One GET request at a time from one client
4. Multiple GET requests at a time from different clients 

5. One PUT request followed by a GET request from the same client
6. One PUT request followed by a GET request from a different client

In all the above, vary the number of requests, the number of clients and the size of the requests.

7. GET key not found
8. GET key cache hit
9. GET key cache miss

TODO:

1. Read/Write workload with only leader.

2. Time to detect leader crash and new leader election

3. Throughput with increasing quorum size (for 100% writes)