Linear indexing is efficient on static databases, that is, records from the database are rarely inserted or deleted. ISAM improves the performance of linear indexing, and can be used for limited updates of the database. As ISAM uses a two-level linear indexing schema, it would break down for a database where the top-level index is already too big to fit into the memory. Thus, as databases become large, we require better organization methods. One approach proposed in Chapter 6, Exploring Search Options, is that a binary search tree could potentially be utilized for indexing to store the primary and secondary keys. The binary search tree provides an efficient structure to store duplicates, and to perform operations such as deletion and insertion given that sufficient memory is available. However, the only disadvantage with a binary search tree is that it could become unbalanced.
Unbalancing becomes an issue, especially in a scenario when the tree is stored in the disk, as...