2. - In distributed computing system provides two concepts:
1. The sharing of computer resources simultaneously by many users
2. The accessing of computers from a place different from the main computer
room
- In 1980 used workstation, it is a single user computer
- The early 1960 and 1970 there are two network technology released:
1. LAN(Local Area Network)-building and campus, data transmission rate is 10
megabits per second
2. WAN(Wide Area Network)-cites and countries, data transmission rate is 64
kilobits per second
-Other technology
3. ATM(Asynchronous Transfer mode)- It is a very high speed networking
possible, data transmission rate is 1.2 gigabits per second
3. System models
1. Minicomputer Model
2. Workstation Model
3. Workstation server Model
4. Processor Pool Model
5. Hybrid Model
All these models are used for building distributed computing systems
13. 3. Workstation Server model:
-workstation model is a network of a personal
Workstations, each with its own disk and local
file system
-diskless workstation-without local disk
-diskful workstation- its own local disk
-diskless workstation is more popular than
diskful workstation
-making the Workstation Server model more
Popular than the Workstation model for building distributed computing systems
19. Issues in design of Distributed Systems
a. Transparency
1. Access Transparency
2. Location Transparency
i. Name Transparency
ii. User Mobility
3. Replication Transparency
4. Failure Transparency
5. Migration Transparency
6. Concurrency Transparency
7. Performance Transparency
8. Scaling Transparency
20. 1. Transparency
Transparency ensures that the complexities of the distributed system are hidden from users, making interactions
seamless. Types include:
Access Transparency: Users access resources consistently, regardless of their physical location or access methods.
Location Transparency: Resources can be accessed without knowledge of their physical or network location.
Name Transparency: Resources retain the same name even if their location changes.
User Mobility: Users can move across locations without affecting access.
Replication Transparency: Users remain unaware of multiple copies of data managed by the system.
Failure Transparency: The system continues to function smoothly during failures, hiding recovery processes.
Migration Transparency: Resources or processes can be moved within the system without user disruption.
Concurrency Transparency: Allows multiple users to access resources concurrently without conflicts.
Performance Transparency: Ensures consistent performance regardless of the system’s workload.
Scaling Transparency: Adapts to system growth or shrinkage without affecting usability.
2. Reliability
Reliability ensures the system remains operational despite faults or failures. Key aspects include:
Fault Avoidance: Reduces the likelihood of failures by using robust hardware and software.
Fault Tolerance: Enables the system to continue functioning during failures using:
Redundancy Techniques: Duplicating critical components (e.g., data, servers).
Distributed Control: Avoiding single points of failure by distributing control across the system.
Fault Detection and Recovery:
Atomic Transactions: Ensures operations either fully complete or do not execute at all.
Stateless Server: Simplifies recovery by not maintaining client state.
Acknowledgment and Retransmission: Ensures communication reliability through timeout-based retransmission of
messages.
21. 3. Flexibility
Flexibility allows the system to adapt and evolve with minimal effort:
Ease of Modification: Updates, patches, or fixes can be applied without disrupting the system.
Ease of Enhancement: New features or upgrades can be integrated seamlessly to meet changing
needs.
4. Performance
Performance optimization ensures efficient resource usage and fast response times. Key practices
include:
Batch Operations: Group tasks together to reduce processing overhead.
Caching: Store frequently accessed data closer to the user to reduce delays.
Minimize Data Copying: Avoid redundant data transfers to save bandwidth.
Minimize Network Traffic: Optimize communication between nodes to reduce latency.
Fine-Grain Parallelism: Use multicore processors to execute smaller tasks in parallel for better speed.
5. Scalability
Scalability ensures the system can grow (or shrink) without losing efficiency or usability. Strategies
include:
Avoid Centralized Entities: Distribute responsibilities to prevent bottlenecks and single points of
failure.
Avoid Centralized Algorithms: Use decentralized algorithms to support scalability.
Perform Most Operations on Clients: Offload tasks to client-side processing to reduce server load.
