update readme

README.md

@@ -5,10 +5,12 @@
 This may not be the best distributed cache, but it is a distributed cache.
 
 Features:
-- LRU (least-recently-used) eviction policy
-- Get/Put operations run in O(1) time
-- Eviction runs in O(1) time
+- LRU eviction policy
+- Get/Put operations and eviction run in O(1) time
 - Client uses consistent hashing to uniformly distribute requests and minimize required re-mappings when servers join/leave the cluster
-- Fault tolerance handled via single-leader replication
-- Bully election algorithm 
-
+- Bully election algorithm used to elect a leader node for the cluster
+- Follower nodes monitor heartbeat of leader and run a new election if it goes down
+- Leader node monitors heartbeats of all nodes in the cluster, keeping a list of active reachable nodes in the cluster updated in real-time
+- Client monitors the leader's cluster config for changes and updates its consistent hashing ring accordingly
+- No single point of failure. The distributed election algorithm allows any nodes to arbitrarily join/leave cluster at any time, and there is always guaranteed to be a leader tracking the state of nodes in the cluster to provide to clients for consistent hashing.
+- Time to update cluster state after node joins/leaves cluster is <= 1 second