Interprocess Communication (IPC)

• Interprocess Communication (IPC) allows processes to exchange data and information. Processes can be either independent or cooperating.
• Independent processes cannot affect or be affected by other processes in the system.
• Cooperating processes can affect or be affected by other processes in the system.
• Any process that shares data with other processes is a cooperating process.

Reasons for Cooperating Processes

• Information sharing enables shared access to files.
• Computation speedup can be achieved by breaking a task into subtasks that execute in parallel.
• Modularity involves dividing system functions into separate processes or threads.
• Convenience allows an individual user to work on multiple tasks at the same time.

IPC Models

• Shared memory
• Message passing

Shared Memory Systems

• Interprocess communication using shared memory requires processes to establish a region of shared memory
• The shared-memory region is in the address space of the process creating the shared-memory segment
• Other processes wishing to communicate must attach to the shared-memory segment's address space
• The operating system typically prevents one process from accessing another's memory
• Shared memory requires that processes agree to remove the restriction that the OS has put in place

Producer-Consumer Problem

• The producer-consumer problem can be addressed using shared memory to allow processes to run concurrently
• It uses a buffer of items filled by the producer and emptied by the consumer that resides in a shared memory region
• Producers create items while consumers use them

Buffers

• Unbounded buffer: Places no practical limit on the buffer's size which make the consumer wait for new items, but the producer can always produce new items
• Bounded Buffer - Has a fixed buffer size where the consumer waits if empty, and the producer waits if the buffer is full

Message-Passing Systems

• Message passing enables processes to communicate and synchronize without sharing the same address space
• This is useful in distributed environments where communicating processes reside on different computers connected by a network

Operations

• send (message)
• receive (message)

Message Size

• Fixed size The system-level implementation is straightforward but, programming can be difficult
• Variable size Requires complex system-level implementation but, simplifies programming

Communication Links

• It is a requirement between processes for communication
• It can be implemented by send() or receive() operations
• There can be direct or indirect communication
• There can be synchronous or asynchronous communication
• There can be automatic or explicit buffering

Naming

• Processes must have a way to refer to each other for communication
• They can use direct or indirect communication

Direct Communication

• Each process must explicitly name the recipient or sender

Properties of Direct Communication

• A link is automatically established between every pair of processes
• Processes need to know each other's identity
• A link is associated with exactly two processes
• There is only one link between each pair of processes
• The sender names the recipient, but the recipient isn't required to necessarily name the sender
• There is limited modularity of resulting process definitions because changing process identifier may require examining other process definitions

Indirect Communication

• Messages are sent to and received from mailboxes or ports

Properties of Indirect Communication

• A link is established between a pair of processes only if both have a shared mailbox
• A link can be associated with more than two processes
• There may be multiple links between each pair of communicating processes, each corresponding to a mailbox

Synchronization

• Communication relies on calls to send() and receive() primitives
• Message passing can be blocking or nonblocking

Message Passing Methods

• Blocking send, the sending process is blocked until the message is received
• Nonblocking send, the sending process sends the message then resumes operation
• Blocking receive, the receiver blocks until the message is available
• Nonblocking receive, the receiver retrieves either a valid message or a null value

Buffering

• Messages reside in a temporary queue irrespective of direct or indirect communication
• Zero capacity: maximum length of zero
• Bounded capacity: finite length n
• Unbounded capacity: potentially infinite length