لخّصلي

Time sharing system
Time sharing (or multitasking) is a logical extension of multiprogramming. Several programs are executed by the CPU which switches between them but the switches are so frequent that users can interact with each program while it is running.

Time sharing systems were developed to provide interactive use of the computer system at a reasonable cost. An interactive or assisted computer system allows online communication between the user and the system. The user gives instructions to the OS or directly to a program and receives immediate responses.

A timesharing OS allows users to share the computer simultaneously. Since each action or command in a time-sharing system tends to be short. We only need a little CPU (quantum) time. As the system switches from one user to another very quickly, each user feels like they have their own computer. When in reality the computer is shared between several users. Time-sharing systems are characterized by:

sharing of CPU time in quantum,
allocation of the processor to a program during a quantum,
taking into account new programs,
very satisfactory degree of interactivity.

Figure 9 : CPU allocation in time-sharing system

I-4 Other operating modes
I-4-1 Personal computer systems
Personal computers appeared in the 1970s. They are microcomputers much smaller and less expensive than mainframes, their operating systems were neither multi-user nor multi-tasking; their goals have changed over time. Instead of maximizing CPU and peripheral usage, they opted for user convenience and responsiveness. Personal computer systems are:

interactive,
friendly,
without sharing the processor.
Example: MS-DOS
I-4-2 Real-time systems
A real-time system is used when there are compelling response time requirements for the operation of a processor or for the flow of data. It is used in the control of dedicated applications (control of a nuclear power plant, robotic assembly line, autopilot, etc.). The sensors bring data to the computer, which must analyze it and possibly adjust the controls to modify the sensor inputs.

There are two types of real-time systems:

Rigid real-time system ensures that critical tasks finish on time. This goal requires that all system delays be limited, from the retrieval of stored data until the end of a query time.
Soft real-time system is a less restrictive type where a critical real-time task takes priority over other tasks and maintains that priority until it completes. Example: Unix, MachI-4-3 Systèmes parallèles
I-4-3 Parallel systems
Most systems are monoprocessor systems, i.e. they have a single main CPU. Multiprocessor systems have more than one processor in close communication sharing the computer bus, clock, and sometimes memory and peripherals. These systems are called strongly coupled systems.
The reasons for creating such systems are as follows:
increase in processing capacity,
save the acquisition cost compared to several mono-systems,
increase reliability (harmonious degradation),
augmenter la fiabilité (dégradation harmonieuse).
Multiprocessor systems use two OS models:

symmetric multiprocessing: each processor executes an identical copy of the OS and these copies communicate with each other.
asymmetric multiprocessing: a master processor controls the system, the other systems receive instructions from the master. In a master-slave scheme.
Example: The Encore version of Unix for the Multimax computer is a symmetrical multi-processing system.

The Sun OS Version 4 OS is an asymmetric multiprocessing OS.
I-4-4 Distributed systems
In these systems the processors are loosely coupled, each processor has its own local memory and its clock. They communicate with each other through various communication lines or telephone lines. Processors may vary in size and functions.

The reasons for building these systems can be:
resource sharing,
acceleration of calculations (load distribution),
reliability and
communication (e-mail)communication (courrier électronique)