Distributed Data Processing

Distributed data processing refers to data processing carried out in a distributed system, where each technological or functional node can independently process local data and make decisions. Nodes in a distributed system can exchange information through communication channels to process data or obtain results of mutual interest.

A distributed system is a set of independent computers that appear to users as a single unified system.

Key Points

There are two important aspects of this definition:

  1. The hardware components of the distributed system are autonomous.
  2. Users perceive the system as unified.

When building distributed systems, there are several tasks to ensure:

  • Sharing of user access to system resources
  • Transparency of the system
  • Openness of the system
  • Scalability of the system

Open Distributed System

An open distributed system provides services that are called using standard interfaces, described by Interface Definition Language (IDL). The description accurately reflects the names of available functions, types of parameters, return values, and exceptional situations that may arise during operation.

Scalability

System scalability can be measured in several ways:

  1. Size: The ease of connecting additional users and resources to the system.
  2. Geographical: Users and resources can be spread out across different locations.
  3. Administrative: The system can be easily managed across multiple organizations.

Information Distributed Systems

Information distributed systems are structured in layers:

  1. Presentation Layer
  2. Application Logic Layer
  3. Resource Management Layer

Layers can be abstract but are often visible as separate subsystems in software.

Information systems must communicate with other systems, involving transforming and presenting information to users. The components handling these activities are part of the presentation layer. Clients can be external and independent; for instance, a network navigator is a client but not its presentation layer. In client/server systems, clients and presentation layers can merge, with programs acting as both.

Information systems perform data processing through programs that execute operations requested by clients via the presentation layer. These programs are often referred to as services offered by the system. Depending on the complexity, this layer may also be called business processes, business logic, business rules, or simply a server, all pertaining to specific implementations.

An information system requires data to function, stored in databases and file systems. The resource management layer handles various data sources, allowing for the recursive construction of information systems from other systems as components.

In practical applications, the conceptual designs of layers can be integrated in various configurations, referred to as tiers. There are four main types of information systems categorized by the number of tiers:

  • One-tier
  • Two-tier
  • Three-tiered
  • Multilayer systems

The number of links in a program with client-server architecture depends on the integration level of the three program layers. A single-layer program tightly connects all three layers, where the presentation layer contains details about the database structure, with the application layer intertwined with both the presentation and service layers, often running on the same computer.