A collection of user data indexed by row, column, and timestamp. Each table is served using the resources of its parent cluster.

The optional list of row keys that will be used to initially split the table into several tablets (tablets are similar to HBase regions). Given two split keys, s1 and s2, three tablets will be created, spanning the key ranges: [, s1), [s1, s2), [s2, ). Example: * Row keys := ["a", "apple", "custom", "customer_1", "customer_2", "other", "zz"] * initial_split_keys := ["apple", "customer_1", "customer_2", "other"] * Key assignment: - Tablet 1 [, apple) => {"a"}. - Tablet 2 [apple, customer_1) => {"apple", "custom"}. - Tablet 3 [customer_1, customer_2) => {"customer_1"}. - Tablet 4 [customer_2, other) => {"customer_2"}. - Tablet 5 [other, ) => {"other", "zz"}.