Threads Vs Processes

You might come across situations where you need to decide whether to use a thread or a process to achieve a certain task. Here I am sharing my experience on selecting what's best depending on your requirements

First, let's what is a process?

• A process is an executing instance of an application/program
• Process control block holds information about the process.
• Process priority, process id, process state, CPU register, program counter, stack pointer, the status of opened files, scheduling algorithms, etc. 
• A process can create other processes which are known as Child Processes.
• The process takes more time to terminate and it is isolated means it does not share the memory with any other process.

Then what is a thread?

• A thread is a path of execution within a process. 
• A process can contain multiple threads
• A thread takes less time to terminate as compared to a process

Let's see how does a process work

• Process creation using fork system call
• fork() creates a replica of the parent process 
• Page Table and Kernal Stack are duplicated, new PCB is created
• Child process state is set to NEW -> READY
• Process is moved to the ready queue of the Kernal
• Copy On Write (COW) is used by the child process
• The first process of Unix system is known as the Super parent.
• Location: /sbin/init
• Created by the kernel during boot
• Typically starts several scripts present in /etc/init.d
• Process termination
• Voluntary termination: exit(0)
• Involuntary termination: kill(pid, signal) 
• Pid: process Id
• Signal: asynchronous message send by the OS/another process (eg: SIGTERM, SIGQUIT)
• Process communication(IPC)
• Shared memory: shmget, shmat, shmdt 
• Message passing:
• Shared memory is created in kernel
• Slow
• Pipes
• Signals

How does a thread work?

• Threads are very inexpensive to create and destroy, and they are inexpensive to represent
• When a new thread is created it shares its code section, data section and operating system resources like open files with other threads.
• But it allocates its own stack, register set and a program counter.
• With so little context, it is much faster to switch between threads
• Creating a thread, switching between threads and synchronization between threads can all be done without the intervention of the kernel

Processes are good when you have a task, where the task

• Has sufficient work to do for a long period of time continuously
• Does not require much communication between tasks
• Has less context switching
• Can independently work from other tasks
• Does not require to be controlled

Threads are good when you have a task, where the task

• Has less heavy work to do
• Has a small life span
• Requires communication between other tasks
• Requires context switching
• Depends on other tasks
• Requires synchronization
• Depends on shared data/resources
• Requires control and coordination

You may consider the following factors when selecting between threads and processes

• Type of work (IO-bound, computation, etc.)
• Lifetime (Short vs Long)
• Scheduling
• Resource Sharing
• Data Sharing
• Communication
• Thread/Process security and safety
• Control and Synchronization
• Environment and technology stack