Parallel & Distributed Computing Overview
Brief Overview
This document provides a concise overview of grid and cloud architectures, distributed
systems middleware, and key concepts like TPS, EAI, and memory models.
Parallel & Distributed Computing
Grid Computing
- Definition: Collection of heterogeneous, loosely coupled resources.
- Model: Unified grid interface.
- Key: Middleware manages resources.
Cloud Computing
- Delivery of scalable services over internet.
- Pay-as-you-go model.
Distributed Information Systems
- TPS: Reliable transactions (BEGIN, COMMIT, ABORT).
- EAI: Connects different applications using middleware.
Distributed Systems
- Collection of autonomous computers acting as one system.
- Middleware acts as OS.
Goals:
- Accessibility
- Transparency
- Openness
- Scalability
Scaling Techniques:
- Latency hiding
- Distribution
- Replication
Types:
- Distributed computing
- Cloud systems
- Information systems
- IoT systems
Advantages:
- Resource sharing
- High availability
- Scalability
Disadvantages:
- Security risks
- Complexity
- Network issues
Memory Architectures:
- Shared Memory (easy, less scalable)
- Distributed Memory (scalable, complex)
- Hybrid (very scalable, very complex)
Multithreading:
- Threads share memory within a process
- Faster than processes
- Models: Many-to-One, One-to-One, Many-to-Many
End of Document