Distributed Object Based Design: Vertical Fragmentation of Classes

• Authors: C. I. Ezeife;Ken Barker
• Venue: Distributed and Parallel Databases
• Year: 1998

Processing costs in distributed environments is most oftendominated by the network communications required for interprocesscommunication. It is well known from distributed relational databasedesign research that careful placement ofdata “near” the users orprocessors where it is used is mandatory or system performance will suffer greatly. Data placement in relational database systems iscomparatively simple because the data is flat, structured, and passive.Objects are characterized by an inheritance hierarchy (other hierarchiescould also be considered including, class composition and execution),unstructured (possibly dynamic data), and contain a behavioralcomponent that defines how the “data” is accessed by encapsulatingit within the object per se. Algorithms currently exist for fragmentingrelations, but the fragmentation and allocation of objects is still a relatively untouched field of study.Similar to relations, objects can be fragmented both horizontally andvertically. Vertical fragmentation must minimize application execution time by splitting a class so that all class attributes and methodsfrequently accessed together are grouped together into a single fragment.This paper adopts a classification of classes into four main models, andcontributes by proposing algorithms for vertically fragmenting thefour realizable class models consisting of simple or complex attributes combined with simple or complex methods. Vertical fragmentation entailssplitting classes into a set of “smaller” equivalent classes (actually fragments of the class‘ extent) that can later beplaced precisely where they are used.Our approach consists of grouping into a fragment, all attributes andmethods of the class frequently accessed together by applicationsrunning on either this class, its subclasses, its containing classesor its complex method classes.