ReplicaLocator provides a way to find the set of owning nodes for a given (token,
replication strategy, table) tuple. It does so by either using the precomputed
token ranges, or doing the computation on the fly (precomputation is configurable).
Represents a sequence of replicas for a given token and strategy,
ordered according to the ring order (for token-ring tables) or with the
order defined by tablet data (for tablet tables).
A token ring is a continuous hash ring. It defines association by hashing a key
onto the ring and then walking the ring in one direction.
Cassandra and Scylla use it for determining data ownership which allows for efficient load balancing.
The token ring is used by the driver to find the replicas for a given token.
Each ring member has a token (i64 number) which defines the member’s position on the ring.
The ring is circular and can be traversed in the order of increasing tokens.
TokenRing makes it easy and efficient to traverse the ring starting at a given token.