Operating System
OS Part-1
OS Part-2
- File Concepts and Access methods
- Free Space Management and Allocation methods
- Directory Systems and Protection
- File Organization, Sharing and Implementation issues
- Disk and Drum Scheduling
- I/O Devices Organisation & I/O Buffering
- I/O Hardware, Kernel I/O subsystem and Transforming I/O Requests to Hardware Operations
- Device Drivers and Path Management
- Device Driver Sub Modules and Procedure
- Device Scheduler and Handler
- Interrupt Service Routine (ISR)
- File System in Linux and Windows
OS Part-3
- Process and Process Control Block(PCB)
- Process Scheduling( Preemptive and Non Preemptive)
- Scheduling Algorithms
- Algorithm Evaluation
- Multiple Processor Scheduling
- Real Time Scheduling
- Operations on Processes
- Threads
- Inter-Process Communication
- Precedence Graphs
- Critical Section Problem
- Semaphores
- Classical Problems of Synchronization
- DeadLock
- Deadlock Prevention and Avoidance
- Deadlock Detection and Recovery
- Process Management in Linux
OS Part-4
- Memory Hierarchy in OS
- Concepts of Memory Management
- MFT and MVT
- Logical and Physical Address Space
- Swapping
- Contiguous and Non Contiguous Memory Allocation
- Paging
- Segmentation
- Paging Combined with Segmentation
- Structure and Implementation of Page Table
- Virtual Memory in OS
- Cache Memory Organization
- Demand Paging
- Page Replacement Algorithms
- Allocation of Frames and Thrashing
- Demand Segmentation
OS Part-5
- Distributed Operating System: Introduction and Types
- Distributed OS: Design Issues
- Distributed OS: File System
- Distributed OS: Remote File Access
- Remote Procedure Call(RPC)
- Remote Method Invocation(RMI)
- Distributed Shared Memory
- Parallel Processing and Concurrent Programming
- Security and Threats Protection in Distributed OS
- Security Design Principles and Authentication in Distributed OS
- Sensor Network and Parallel OS
Distributed Shared Memory (DSM) in Operating System
Introduction
Distributed Shared Memory (DSM) ek technique hai jo distributed systems me shared memory ka illusion create karti hai, jisme multiple processes ya computers ek hi memory space share karte dikhai dete hain, lekin physically unki memory alag hoti hai. DSM distributed systems me memory ko efficiently share karne aur inter-process communication (IPC) ko simplify karne ke liye use hoti hai.
Example: Agar ek distributed system me multiple computers connected hain, to DSM ka use karke aise lag sakta hai ki ek hi shared memory available hai, jo ki simpler programming model provide karta hai.
How DSM Works?
DSM virtual memory techniques ka use karke ek logical shared memory space provide karta hai. Ye data ko consistency models aur page replication ke through manage karta hai.
Diagram: DSM Architecture
Har node (computer) ki local memory hoti hai, lekin system ek global shared memory ka illusion create karta hai.
Advantages of DSM
Simplified Programming Model – Developers bina explicit message passing ke memory share kar sakte hain. Reduced Communication Overhead – DSM me automatic data sharing hoti hai, is wajah se manual IPC techniques ki zaroorat nahi hoti. Scalability – System me naye nodes add kiye ja sakte hain bina major architectural changes ke. Fault Tolerance – Agar ek node fail hota hai, to dusre nodes data ko replicate kar sakte hain.
Design Issues in DSM
DSM ka implementation multiple challenges face karta hai:
1. Data Consistency
DSM me ek consistency model define kiya jata hai jo ensure karta hai ki multiple processes correct aur up-to-date data access karein. Strict Consistency – Har process hamesha latest update ko dekhega. Sequential Consistency – Execution ka order same hoga, but timing different ho sakti hai. Causal Consistency – Agar ek process dusre ko affect karta hai, to unka order same hoga. Eventual Consistency – Time ke saath data consistency achieve ho jayegi.
2. Memory Granularity
DSM me memory page-based ya object-based ho sakti hai. Page-based DSM – Memory ko fixed-size pages me divide kiya jata hai. Object-based DSM – Memory ko logical objects ke form me store kiya jata hai.
3. Data Replication & Migration
-
-> Replication – Multiple nodes same data ko store kar sakte hain taaki faster access mile.
-
-> Migration – Data dynamically ek node se doosre node pe shift hota hai taaki load balancing achieve ho.
4. Synchronization
Jab multiple processes ek shared variable ko access karte hain, to synchronization techniques ka use hota hai: Locks – Ek process ko ek memory block par exclusive access milta hai. Semaphores – Critical sections ka access control karne ke liye. Barriers – Processes ko synchronize karne ke liye taaki sab ek specific point pe ek saath pohonchein.
Types of DSM Systems
DSM ko implementation techniques ke basis par classify kiya jata hai:
1. Hardware-based DSM
-
Ye system special hardware support ka use karta hai jisme cache coherence mechanisms implement hote hain.
-
Example: NUMA (Non-Uniform Memory Access) architectures
2. Software-based DSM
-
Ye OS-level aur middleware-level solutions ka use karta hai bina kisi special hardware ke.
-
Example: TreadMarks DSM System
3. Hybrid DSM
-
Ye hardware aur software approaches dono ka combination hota hai.
-
Example: COMA (Cache-Only Memory Architecture)
Page-based DSM vs. Object-based DSM
| Feature | Page-based DSM | Object-based DSM |
|---|---|---|
| Granularity | Fixed-size pages | Logical objects |
| Access Control | Page-level locks | Object-level locks |
| Communication Overhead | High | Low |
| Example | TreadMarks | CORBA |
DSM in Real-World Applications
Parallel Computing – Large-scale computations ke liye DSM ka use hota hai. Cloud Computing – Distributed databases DSM ka use karke efficient memory management karti hain. High-Performance Computing (HPC) – Scientific simulations aur data analysis ke liye DSM ka use hota hai. Distributed Databases – SQL aur NoSQL databases DSM ke principles ka use karti hain.
Conclusion
Distributed Shared Memory (DSM) ek powerful memory management technique hai jo distributed computing ko simpler aur efficient banati hai. Ye parallel processing, scalability, aur high-performance computing me bahut useful hoti hai. DSM ka proper implementation consistency, synchronization aur replication strategies par depend karta hai.