Module-3

1. Explain Distributed Scheduling and Characteristics of Good Scheduling Algorithms

Distributed Scheduling

Definition

• Distributed scheduling means dividing and assigning processes (tasks) to different

computers in a distributed system.

• The main aim is to:

o Use all computers effectively.

o Balance the workload.

o Avoid idle systems.

Approaches of Distributed Scheduling

1. Task Assignment

• Tasks are assigned to specific nodes.

• Improves the overall system performance.

2. Load Balancing

• Work is evenly distributed among all nodes.

• Prevents any computer from becoming overloaded while others remain free.

3. Load Sharing

• Ensures no node remains idle.

• Shares work from busy nodes to idle nodes.

Characteristics of Good Scheduling Algorithms

1. The user should not have to provide details of processes or resources.

2. Must allow process migration (moving a process from one node to another).

3. Must take scheduling decisions quickly with minimum overhead.

4. Must remain stable and avoid excessive process migration (thrashing).

5. Should use heuristics (practical rules) to obtain near-optimal scheduling results.

2. Explain Distributed Shared Memory (DSM), its architecture, advantages and disadvantages.

Definition

• Distributed Shared Memory (DSM) is a method of managing memory across multiple

computers (nodes) in a network.
 • It makes all nodes appear to use a single shared memory, even though each node has its
own physical memory.

• Programs access memory as if it were shared directly, so programmers do not need to

manage message passing.

• DSM creates a common virtual address space, and data is transferred between nodes using
consistency rules.

Components of DSM Architecture

1. Processes

• User applications running on different computers.

• Access memory as if it is shared.

2. Memory Manager

• Present on every node.

• Manages memory requests.

• Transfers data between nodes.

• Maintains memory consistency.

3. Shared Virtual Memory

• Provides a single virtual address space.

• Makes distributed memory appear as one shared memory.

• Data is transferred between nodes using consistency protocols.

Types of Distributed Shared Memory

1. On-Chip Memory

• Memory is built directly inside the CPU chip.

• Connected directly to the CPU address lines, making it very fast.

• Costly and complex to design.

2. Bus-Based Multiprocessors

• A bus (parallel wires) connects the CPU and memory.

• Special algorithms prevent multiple CPUs from accessing the same memory simultaneously.

• Cache memory reduces unnecessary network traffic.

3. Ring-Based Multiprocessors

• There is no central memory.

• All nodes are connected in a ring structure.

• Data is transferred using a token.

• The address space is divided into shared areas across the ring.

Advantages of DSM

1. Simpler for Programmers – No need to manage data transfer manually.

2. Easier Portability – Programs can run on different systems using a common interface.

3. Locality of Data – Data is moved in blocks, improving future access.

4. On-Demand Data Movement – Only required data is transferred.

5. Larger Virtual Memory – Total memory is the sum of all nodes' memory.

6. Better Performance – Faster data access than message passing.

7. Flexible Communication – Nodes can easily join or leave the system.

8. Supports Process Migration – Processes can move between nodes using the shared address
space.

Disadvantages of DSM

1. Slower Access – Data access is slower than in non-distributed systems.

2. Consistency Issues – All copies of shared data must remain consistent.

3. Asynchronous Messaging – Internally uses message passing, which may be less efficient.

4. Data Redundancy – Multiple copies of the same data may exist.

5. Lower CPU Performance – CPU performance may decrease due to DSM management
overhead.
3. Explain Types of Distributed Shared Memory.

Types of Distributed Shared Memory (DSM)

1. On-Chip Memory

• Memory is built directly inside the CPU chip.

• Connected directly to the CPU address lines, making it very fast.

• Very costly and complex to design.

2. Bus-Based Multiprocessors

• A bus (parallel wires) connects the CPU and memory.

• Special algorithms prevent multiple CPUs from accessing the same memory simultaneously.

• Cache memory is used to reduce unnecessary network traffic.

3. Ring-Based Multiprocessors

• There is no central memory.

• All nodes are connected in a ring structure.

• Data is transferred using a token.

• The address space is divided into shared areas across the ring.

4. Explain the mechanisms for building Distributed File Systems. (repeated )

Distributed File System (DFS) – Pointwise

Definition

• A Distributed File System (DFS) allows files to be stored on multiple computers (servers).

• Users can access these files as if they were stored on their local machine.

• It provides a single, unified view of files, even though they are stored in different locations.

• DFS makes file sharing, collaboration, and remote access easier and more efficient.

Applications of DFS

1. NFS (Network File System)

 o Developed by Sun Microsystems.

o Allows a computer to access files on another computer over a network.

2. CIFS (Common Internet File System)

o Microsoft's extension of SMB.

o Mainly used in Windows systems for file and printer sharing.

3. SMB (Server Message Block)

o Developed by IBM.

o Enables sharing of files and resources such as printers over a network.

4. Hadoop Distributed File System (HDFS)

o Used in big data applications.

o Stores and processes very large datasets across multiple servers.

5. NetWare

o A network operating system developed by Novell.

o Provides file sharing, multitasking, and DFS services.

Working of DFS

1. Standalone DFS Namespace

• DFS root is stored on a single computer.

• Does not use Active Directory (AD).

• Rarely used because it has limited advantages and poor fault tolerance.

2. Domain-Based DFS Namespace

 • DFS root is stored in Active Directory (AD).

• Makes DFS available to all users on the network.

• Files are accessed through paths such as \domainname\dfsroot.

• Provides scalability, fault tolerance, and easy management.

Advantages of DFS

1. Multiple users can simultaneously access and share files.

2. Provides transparency, so users do not need to know the actual file location.

3. Improves availability through file replication.

4. Improves performance by balancing workload and reducing file access time.

5. Increases network efficiency by distributing files across multiple servers.

Disadvantages of DFS

1. Security risks if network nodes or communication links are not secure.

2. Possibility of data/message loss during file transfer.

3. Complex management compared to a single local file system.

[Link] Multimedia File Systems and File Placement Techniques.

Multimedia Systems

Definition

• A multimedia system is a computer-based system designed to create, store, integrate, and

manage different types of digital media such as text, audio, video, images, and animation.

• It supports applications like online learning, video conferencing, gaming, streaming

platforms, and digital libraries.

• Unlike a normal computer, a multimedia system can combine and process two or more
types of media simultaneously.

Multimedia File System

Definition
A Multimedia File System is a file system designed to store, organize, and manage multimedia data
such as text, images, audio, video, and animation. It provides continuous, real-time access to
multimedia files for smooth playback and efficient storage.

Features of Multimedia File Systems

1. Supports storage of audio, video, images, and animation.

2. Provides continuous data transfer for uninterrupted playback.

3. Supports large file storage.

4. Reduces latency and improves data access speed.

5. Supports real-time multimedia applications.

6. Handles different multimedia file formats such as JPEG, PNG, MPEG, AVI, WAV, and MP3.

File Placement Techniques

Definition

File placement is the strategy used to decide where files or data blocks should be stored in a
distributed file system. The aim is to reduce access time, balance the load, and increase reliability.

Techniques of File Placement

1. Centralized File Placement

• Files are stored on a single central server.

• Easy to manage.

• May become a performance bottleneck.

2. Distributed File Placement

• Files are stored across multiple servers (nodes).

• Improves performance and availability.

• Balances the workload among servers.

3. Replicated File Placement

• Multiple copies of files are stored on different servers.

• Improves reliability and fault tolerance.

• Ensures files remain available even if one server fails.

4. Partitioned (Block-Based) File Placement

• Large files are divided into smaller blocks.

 • Blocks are stored on different servers.

• Example: Hadoop Distributed File System (HDFS) stores files in blocks across multiple
servers for better speed and reliability.

[Link] Block Caching with suitable example. (repeated )

Block Caching

Definition

Block caching is a technique in which frequently accessed data blocks are stored in a cache (main
memory or local storage). This reduces repeated access to the disk or server, resulting in faster data
retrieval and improved system performance.

Working of Block Caching

1. A process requests a data block.

2. The system first checks whether the block is available in the cache.

3. If the block is found (cache hit), it is returned immediately.

4. If the block is not found (cache miss), it is fetched from the disk or server.

5. The fetched block is stored in the cache for future access.

Types of Block Caching

1. Client-Side Caching

• Data blocks are cached on the client's local machine.

• Reduces repeated requests to the server.

2. Server-Side Caching

• Frequently accessed data blocks are cached on the server.

• Improves response time for multiple clients.

3. Distributed Caching

• Cache is shared across multiple servers or nodes.

• Improves scalability and load balancing.

Example

Suppose a user watches a video stored on a distributed file system.

• The first time the video is played, the required data blocks are loaded from the server.

• These blocks are stored in the local cache.

 • If the user plays the video again, the data is read directly from the cache instead of
requesting it from the server.

• This results in faster playback and reduced network traffic.

Advantages

• Improves system performance.

• Reduces data access time (latency).

• Decreases network traffic.

• Reduces server workload.

• Optimizes resource utilization.

Disadvantages

• Requires additional cache memory.

• Cache consistency must be maintained.

• May return stale (outdated) data if the cache is not updated properly

[Link] how placing files on multiple disks improves Multimedia File System performance
(repaeated )

Definition

Placing files on multiple disks means storing multimedia files or their data blocks across more than
one storage disk instead of a single disk. This improves the speed, reliability, and performance of
multimedia file systems, especially for large audio and video files.

How Multiple Disks Improve Performance

1. Parallel Data Access

• Different parts of a file are stored on different disks.

• Multiple disks can read data simultaneously.

• Increases data transfer speed.

2. Reduced Access Time

• File requests are distributed across several disks.

• Reduces waiting time for users.

• Improves response time.

3. Load Balancing

• Workload is shared among multiple disks.

 • Prevents a single disk from becoming overloaded.

• Ensures efficient utilization of storage resources.

4. Higher Throughput

• Multiple disks process several read/write requests at the same time.

• More multimedia files can be served simultaneously.

5. Continuous Multimedia Playback

• High-speed data transfer prevents interruptions.

• Supports smooth playback of audio and video without buffering.

6. Improved Reliability

• Data can be replicated across multiple disks.

• If one disk fails, data can still be accessed from another disk.

7. Better Scalability

• Additional disks can be added as storage requirements increase.

• Supports growing multimedia applications