Distributed Database Design: Lecture Notes on Design Problems, Strategies, Fragmentation and Allocation – Chapter 3

Summary

Distributed database design involves deciding where to place data and programs in a computer network. Students learn about design problems, strategies, fragmentation and allocation of database fragments.

Design Problem

The design problem of distributed systems focuses on placing data and programs across network sites. The distribution of applications is not considered. The distribution of data is studied instead.

Design Strategies

Distribution design is central to distributed database management systems. It involves fragmenting relations and allocating fragments. The chapter concentrates on these two aspects.

Fragmentation

Fragmentation decides the unit of distribution: whole relations or fragments. Fragments of relations are usually the best unit. Correctness rules for fragmentation include completeness, reconstruction and disjointness.

Replication and Allocation

Replication decides which fragments to store as multiple copies. Allocation decides where to store fragments. The goal is to find the optimal distribution of fragments to sites.

Key terms

Fragmentation

The process of dividing relations into smaller sub-relations for distribution.

Allocation

The process of deciding where to store database fragments.

Replication

The process of storing multiple copies of database fragments.

Horizontal Fragmentation

Dividing a relation into subsets of tuples based on predicates.

Vertical Fragmentation

Dividing a relation into subsets of attributes.

Optimality

The goal of minimizing cost and maximizing performance in fragment allocation.

Questions this document helps with

• What are the main aspects of distribution design in distributed database management systems?
• How do you decide whether to use whole relations or fragments for distribution?
• What factors influence the allocation of database fragments to network sites?

Summary written with AI from the document's text.