Basics and differences of Distributed Systems and Parallel Computing

1. What are distributed systems?

 It is basically an information highway. The basic units are called nodes. Their hardware is separate but they are connected to a common network. They all work together like 3 people riding a tri-cycle.

  

2. Why distributed systems?

To understand why we use distributed systems let's go back to the tricycle example. Suppose those 3 people have to go from A to B. If only one person is pedaling it would take much time but if all 3 are pedaling we would reach our destination faster. Similarly in distributed systems, the tasks can be divided between the nodes. Moreover, data distribution and many other applications are possible.

3. What is parallel computing?

It is basically better and more efficient use of hardware. To explain it imagine a toll booth with only one booth. Cars will take time to get cleared and will keep on piling. If we increase the number of booths multiple cars can now be checked and can be on their way faster. In parallel computing, the toll booths are processors and instructions to be executed are cars. Adding more processors and dividing the tasks increases the efficiency of the overall hardware.

4. Why parallel computing?

Due to the single instruction at a time or known as serial computing the computing time was increasing as the stack of instructions increased. To solve this issue parallel computing was developed. The overall computing speed and efficiency of hardware increased due to it.

5. Difference between distributed systems and parallel computing.

 1. In parallel computing it is a single device that does multiple tasks with the help of many processors whereas in distributed systems multiple computers perform multiple tasks over the common network.

 2. Parallel computing has shared or distributed memory whereas distributed system only has distributed memory.

 3. In parallel computing all the processors and interconnected with each other through a bus while in distributed systems all nodes are connected to a common network.

 4. Parallel computing basically increases the performance of the system so that it can execute multiple instructions at once whereas distributed systems increase the scalability of the system, reduce errors and has increased resource sharing capacities.

6. Types of distributed systems

1. Client-Server Systems -

 Request response system

2. Peer-to-Peer Systems - 

Programs are evenly divided between nodes

3. Middleware - 

Increase interoperability

4. Three-tier -

 Information about the client or task is saved in the middle tier rather than with the client.

5. N-tier - 

Request response between server/applications and devices on a network.

7. Types of parallel computing

 1. Bit-level parallelism-

 Improving hardware

 2. Instruction-level parallelism-

 Reorganization of instructions to improve efficiency

 3. Task parallelism

 Division of tasks to improve efficiency

 4. Data-level parallelism

 Instructions from one stream operate on several data. 

8. Advantages of Distributed Systems

 1. Saves time

 2. Saves money

 3. Solves more complex programs

9. Advantages of parallel computing

 1. Reliability

 2. Scalability

 3. Flexibility

 4. Fast speed

 5. High performance 

Thank you for reading our blog 

Blog by-

Ayush Phadtare

Chashu Agrawal

Sakshi Awale

Gayatri Aurangbadkar