JAWAHARLAL NEHRU TECHNOLOGICAL UNIVERSITY HYDERABAD MCA 2022 DISTRIBUTED DATABASES Previous Question Paper (February - 2022 - R17)

Data fragmentation is a design technique in distributed databases used to divide a global relation into smaller logical units called fragments. This process improves performance by storing data closer to where it is most frequently accessed and enhances parallelism. The three primary types of fragmentation are:

1. Horizontal Fragmentation: This involves partitioning a relation into subsets of tuples based on a selection predicate. Each fragment contains a subset of the rows of the original table, while all fragments share the same schema. For example, a global Employee table can be fragmented by location:

2. Vertical Fragmentation: This divides a relation into subsets of attributes (columns). Each fragment contains a subset of columns, and to ensure the original relation can be reconstructed, a primary key is included in every fragment. This is useful when different applications require only specific columns. For example, an Employee table can be split into personal details and salary details:

3. Mixed (Hybrid) Fragmentation: This is a combination of horizontal and vertical fragmentation. A relation is first fragmented horizontally and then vertically, or vice-versa. This is used in complex scenarios where data access patterns are highly specific. For example, one might first fragment the Employee table by location, and then further split the Hyderabad fragment into personal and salary components.

Fragmentation must satisfy three rules: Completeness (all data must be mapped to at least one fragment), Reconstruction (the original relation must be recoverable using relational operators like Union or Join), and Disjointness (data should not be redundant unless intentionally replicated).