Test Yourself [merged with Fault tolerance]

We used a strategy where the set of tokens, t, is maintained and distributed among the nodes. A pair of successive tokens define a range. The final and first tokens form a wrapping range, spanning the hash space from the greatest value to the lowest value. As a result, the size of the ranges can vary for each node. The token set and ranges fluctuate when more and fewer nodes join and exit the system. As the number of nodes in the system grows, the amount of space required to keep track of each node’s membership.

Identify any problem with this approach?