Process Management in Operating Systems

Course Instructor : Dr. Reham El-Shamy

Teaching Assistant : Eng. Ahmed Reda

Prepared By : Arwa Nasser Abdelhamid

Code : 623247
1. Introduction
Operating systems are the backbone of modern computing devices, enabling hardware
and software to work together seamlessly. Among the most essential components of
any OS is process management, a subsystem responsible for handling all running
programs and coordinating their execution. Every task performed on a computer—
browsing the internet, running applications, playing videos, editing documents—relies on
process management to allocate CPU time, memory, I/O resources, and ensure stable
execution without crashes.

Process management ensures that the computer appears fast, responsive, and
multitasking-friendly. For example, when a user watches a YouTube video while
downloading a file and listening to music, the OS must manage all these tasks
simultaneously. This complex coordination happens through process creation,
scheduling, synchronization, communication, and termination.

This paper provides a deep, expanded, and highly detailed exploration of process
management in modern operating systems, including concepts, structures, algorithms,
communication models, synchronization, and real-world implementations in Linux and
Windows. The goal is to produce an academically strong reference suitable for
university-level submission.

2. Defining a Process
A process is an active, running instance of a program. While a program is a passive set
of instructions stored on disk, a process includes:

 Loaded code into memory

 Data and stack sections
 Register values
 Kernel resources
 Open files
 Threads
 Permissions
 Execution context

2.1 Programs vs. Processes (Expanded

Explanation)
A program becomes a process only when executed. For example:

 Opening Google Chrome spawns multiple processes:

o Browser process
 o GPU process
o Renderer processes
o Utility processes

Each acts independently, with its own memory space and permissions, ensuring both
performance and security.

This distinction is crucial because the OS doesn’t schedule programs—it schedules

processes.

3. Process States (Extended)

Every process transitions between states during its lifetime.

3.1 The Five Classic States

1. New: The process is being created by the OS.

2. Ready: Waiting for CPU allocation.
3. Running: CPU actively executes its instructions.
4. Waiting/Blocked: Waiting for I/O events or resource availability.
5. Terminated: Finished execution.

3.2 Advanced/Suspended States

Modern OSs add additional states:

 Suspended Ready: In secondary memory (swapped out).

 Suspended Blocked: Blocked + swapped.
 Zombie: Process finished but PCB still exists.
 Orphan: Parent terminated before child.

These extended states are especially important in Linux and Windows.

4. Process Control Block (PCB) – Deep

Structure
The Process Control Block is the "identity card" and "memory" of each process. It stores
all information needed to resume the process after a context switch.

4.1 Contents of a PCB (Extended List)

 Process ID (PID)
 Parent process ID
 Process priority
 Current state
 CPU registers
 Program counter
 Memory boundaries
 Page tables
 Open file descriptors
 I/O status information
 Accounting information
 Thread list (if multi-threaded)
 Security credentials
 Signal handling configurations

Without the PCB, multitasking would be impossible.

5. Process Operations (Deep Expansion)
5.1 Process Creation
Occurs through system calls such as:

 fork() in Unix
 CreateProcess() in Windows

Creation involves:

1. Allocating memory
2. Initializing PCB
3. Setting default priority
4. Inheriting resources from parent
5. Adding the process to ready queue

5.2 Process Termination

A process may terminate due to:

 Successful completion
 Unhandled exception
 Access violation
 Out-of-memory error
 Manual user kill
 OS resource protection

6. CPU Scheduling (Extended)

Modern computers run thousands of processes. CPU scheduling decides which one
gets CPU time.

6.1 Goals of Scheduling

 Reduce waiting time
 Maximize CPU utilization
 Maximize throughput
 Improve fairness
 Minimize response time (important for user experience)
 Meet real-time deadlines
7. CPU Scheduling Algorithms (Large
Expansion)
Below is an expanded discussion with examples.

7.1 FCFS Example

Processes arrive:

P1 (Burst=24), P2 (Burst=3), P3 (Burst=3)

P1 delays P2 and P3 → high waiting time.

7.2 SJF Example

Processes:

P1(6), P2(8), P3(7), P4(3)

Order: P4, P1, P3, P2

Reduces average waiting dramatically.

7.3 Preemptive vs Non-Preemptive Scheduling
 Preemptive allows interrupting running processes
 Non-preemptive lets processes run until completion

7.4 Priority Scheduling Expanded

Low-priority processes may starve indefinitely.

Aging gradually increases their priority to prevent starvation.

7.5 Round Robin (RR) Expanded

Quantum too small → too many context switches.

Quantum too large → behaves like FCFS.

RR is used in:

 Linux
 macOS
 Android
 Windows time-sharing

7.6 Multilevel Queue Scheduling

Queues:

 System processes (highest priority)

 Interactive tasks
 Batch jobs
 Background services

Each queue has its own algorithm.

7.7 Multilevel Feedback Queue (MLFQ)
The most advanced scheduling model in real OSs.

Processes can move up/down between queues based on behavior.

Used by:

 BSD Unix
 MacOS
 Windows
 Many Linux distros

8. Context Switching (Extended)

8.1 What Happens During a Context Switch?
1. Save current register values
2. Update PCB
3. Choose next process
4. Load next PCB
5. Restore registers
6. Jump to program counter

Context switch time is wasted CPU time → OS must minimize it.

9. Threads and Multithreading (Deep

Expansion)
9.1 Advantages of Threads
 Faster communication
 Lower overhead
 Better CPU utilization
 Useful in parallel programs
9.2 Types of Threads
 User threads: Managed by user-level libraries
 Kernel threads: Managed by OS kernel

10. Interprocess Communication (IPC) –

Big Section
10.1 Why IPC is Important

Processes must coordinate safely.

Example:

Browser process communicates with renderer process through IPC for security
reasons.

10.2 IPC Techniques Expanded

 Pipes
 Named Pipes
 Shared Memory
 Message Queues
 Semaphores
 Signals
 Memory Mapped Files
 Sockets

Shared memory is fastest but needs synchronization.

Message passing is simpler but slower.

11. Synchronization (Extended)

11.1 The Critical Section Problem

Any code accessing shared data must be protected.

Without it → race conditions.

11.2 Tools

 Disabling interrupts (older systems)

 Locks
 Semaphores
 Monitors
 Spinlocks
 Atomic operations

12. Deadlocks (Extended Large Section)

Deadlocks occur when processes form a cycle of resource requests.

12.1 Deadlock Example

P1 holds R1 needs R2

P2 holds R2 needs R1

12.2 Detection Algorithms

 Resource allocation graph

 Wait-for graph

12.3 Avoidance

 Banker’s Algorithm
12.4 Recovery

 Process termination
 Resource preemption

13. Memory Management and Processes

(Large Expansion)
13.1 Paging

Allows non-contiguous memory allocation.

13.2 Segmentation

Divides memory into variable-size segments.

13.3 Demand Paging

Loads pages only when needed.

13.4 Thrashing

Too many page faults → CPU spends most time swapping.

14. Real-Time Processes

14.1 Types

 Hard real time: Medical systems, autopilot

 Soft real time: Streaming, gaming

14.2 Algorithms

 Rate Monotonic
 Earliest Deadline First
15. Case Study: Linux Process
Management (Large Expansion)
15.1 Process Model

 Everything is a file
 Uses fork() + exec()
 Strong IPC support

15.2 Linux Scheduler (CFS)

The Completely Fair Scheduler distributes CPU time "fairly" based on virtual runtime.

15.3 Tools for Process Management

 top
 htop
 ps
 /proc filesystem

16. Case Study: Windows Process

Management (Large Expansion)
16.1 Key Characteristics
 Thread-centric model
 Strong security architecture
 Process isolation

16.2 Windows Scheduling

 Priority-based preemptive scheduler

 32 priority levels

16.3 Process Objects

 Job objects
 Thread objects
 Handle tables
17. Virtualization and Process
Management
Virtual machines add another layer of scheduling:

 Hypervisor schedules virtual CPUs

 Guest OS schedules processes
 Host OS schedules threads

This creates multi-layer scheduling complexity.

18. Cloud Computing and Process

Management
Cloud systems run millions of isolated workloads.

Cloud scheduling considerations:

 Latency
 Fairness
 Load balancing
 Energy efficiency
 Fault tolerance

19. Future Directions

 AI-based schedulers
 Energy-aware scheduling
 Hybrid user-kernel threading
 Process containers (Docker)
 Serverless process execution
20. Conclusion
Process management remains a foundational discipline in operating systems. As
technology evolves—virtualization, cloud computing, AI—the complexity and importance
of efficient process control grow dramatically. A deep understanding of processes,
scheduling, synchronization, IPC, and deadlocks is crucial for students, developers, and
system designers.

21. References (Expanded - 10+

Sources)
1. Abraham Silberschatz, Peter Galvin, Greg Gagne – Operating System Concepts,
Wiley.
2. Andrew S. Tanenbaum, Herbert Bos – Modern Operating Systems, Pearson.
3. William Stallings – Operating Systems Internal Principles, Pearson.
4. Remzi Arpaci-Dusseau – Operating Systems: Three Easy Pieces.
5. Linux Kernel Documentation – [Link]
6. Microsoft Windows Internals – Mark Russinovich.
7. Maurice Herlihy – The Art of Multiprocessor Programming.
8. Intel CPU Architecture Manuals.
9. IBM Research Papers on Scheduling Algorithms.
10. ACM Digital Library – OS Process Management Papers.