Operating System

                                                       OPERATING SYSTEM
INTRODUCTIONOperating systems are programs that manage computer hardware. Operating systems also providefoundation for application programs and acts as an intermediary between users withcomputer hardware.What does the Operating SystemA computer system can be divided into four parts, the hardware, operating system, programsapplications and the user (Figure 1.1).System and Application ProgramOperating SystemHardwareUsers UserFigure 1.1 Components of Computer SystemHardware parts, namely the Central Processing Unit (CPU), memory, input output (I / O devices) arebasic components in the process of calculation by the computer system. Application programs, such as wordprocessor, spreadsheed, compiler, and a web browser, a tool or how the infrastructurehardware will be used to solve problems that are owned by the user. Operating systemis the controlling and coordinating the use of hardware among various application programsthat exist.Viewpoint User (User)Setting the Operating System is based on the user's perspective, depending on the interfaceis used. In general, the Operating System on a computer system which only consists of Monitor, CPU,Keyboard, and Mouse is designed to let users (single user) to monopolize all the resourcesthere. Its main objective is to maximize the work is being done. Operating systemsattached to the computer system such as this, designed for ease of use with mostgreater emphasis on performance goals and a little on the utilization of resources (how many resourceshardware and software used for the common good).On mainframes or minicomputers, many users are connected to the system through terminalterminal.These users to exchange information and use system resources together.In such cases, the operating system used is designed for utilization of resources,maximize resources for CPU time, memory, I / O is used in an efficient and fair to eachthe user.There is also a model where the user uses worksation connected with other workstations,and server in a network. Each user has the resources on each workstationand also share resources with others such as networks, printers, and files. Systemoperating system on a circuit like this, built by combining resources and utilizationpersonal use of each user.Viewpoint SystemsFrom the standpoint of the system, the operating system is a program that most closely relate tothe hardware. In this connection, the operating system is considered as a resource allocator. A systemcomputers have a lot of resources needed to solve a problem, namely the CPUTime, memory, file strorage, I / O devices, and so on. Faced with so many conflictsdemand for use of resources, the operating system must determine howallocate these resources to a program and the user so that the computer system canrun efficiently and fairly.In addition, the operating system can also be emphasis on the needs of different kinds of control I / Odevices and programs that run user. In this position, the operating system is referred to as controlprogram (program controller). Control program to manage the execution of a program run by the userso there is no error and error use of computers, especially on the use of I / Odevices.Computer Systems OrganizationComputer System OperationModern general-purpose computers consisting of one or more cpu and some device controller thatconnected by bus (data path) as a provider of access to shared memory (Figure 1.2).Each device controller is connected to a particular function, such as disk drives, audio devices,and video displays. CPU and various kinds of device controller can be executed simultaneouslyand compete for the memory cycles. To ensure the order of accessing the sharedmemory, a memory controller is added as a means of synchronization.Figure 1.2 Modern computer systemsWhen a computer is turned on, for example, the system will require an initial program.Initial program or bootstrap program is usually simple and stored in Read Omly Memory(ROM) or Erasable Programmable Read-only Memory (EEPROM). The function of the bootstrap isperform the initialization of all aspects of systems, from CPU registers, device controllers, and memory.Bootstrap program must know how to load the operating system and execute the system.Based on these objectives, the bootstrap should be able to find and load him into the kernelthe operating system. Then the process begins executing the operating system its first and wait aevents (event) occurs.The emergence of an event is usually characterized by an interrupt from either hardware or software.Hardware can trigger an interrupt at any time by sending a signal to the CPU and usually throughthe system bus. While the software can perform an interrupt at any time via a special operation incall the system call.When the interrupt occurs, the CPU immediately stops what is being done and transferThe requested execution to the proper location. The location is referred to here usually have a start addresswhere the interrupt service routine is located.Figure 1.3 Interrupt process time line.Storage StructureThe computer program, when run should be at the main memory (main memory), which usuallydisesbut Random Access Memory (RAM), to be executed. Main memory is the only medialarge storage that can be directly accessed by the processor. Implementation of this memory is calledas Dynamic Random Access Memory (DRAM), which form a memory array in sizewords. Each of these words have each address. Interaction with memory occurs throughmechanism of load-store instruction to the specific address in memory. Load instructions move onewords from main memory into the CPU registers, while the store opposite.The series of instruction execution (instruction-execution cycle) begins taking instruction frommemory which is then placed in the instruction register. After the execution results can bereturned to the memory. For the record, memory only serves as a storage mediumwhich contains a series of addresses. It means he does not know how (by instrution counter, indexingor the other) or for what (instruction or data) circuit is formed.Ideally, it would be better if the software and data can reside in main memory forever.But unfortunately this can not be done because:1. Main memory is not large enough to be able to store the entire application and data.2. Mainmemory is a temporary storage medium that will lose all of its contents whencomputer is turned off.Due to above reasons, was made a second storage media (secondary storage) in addition to themain memory. Its main purpose is for storage of large amounts of datapermanent.The most common storage medium is a magnetic disk, which can be used to storeprograms and data. Most of the programs will be recorded in this place until the load into memory.Other examples of storage media is the cache memory, CD-ROMs, magnetic tapes, and so on.All storage media have the same basic functions, ie save the data tothen be used again. There are differences in speed, price, capacity, and longstorage (volatile). Figure 1.4 shows the composition of the storage medium based on speedand price.Figure 1.4 Composition of the storage media.In addition to price and speed, as already mentioned earlier, another factor is the long volatile.Volatile in question here is whether the data is lost when power is switched off or not. OnFigure 1.4 above, all the above elektronic disk storage media is volatile.I / O StructuresDepending on the device controller, there could be more than one device (device) is connected.For example, seven or more devices can be added to the small computer systems interface-(SCSI) controller. A device controller is set up some local buffer storage and special purposethe register. Device controller main task is to regulate the movement of data between devices in the controlwith local storage buffer. Generally each operating system has a device driver for each devicethe controller. Device driver then provides an interface for the operating system.When starting the operation of I / O, device drivers to load the registers related to the device controller. Devicecontroller then checks the contents of registers to determine what action needs to be done.Subsequently the data transfer from device to local buffer storage. Once the data transfer is completedevice controller tells the device driver via an interrupt that work has been completedcarried out. Device driver then returns control to the operating system, or providedata or a pointer if the requested operation before the reading.Figure 1.5 Employment of modern computer systemsThe above method works best if used for a small amount of data, but will generateoverhead when used for large data transfers, such as disk I / O. Completion of this problemis to use direct memory access (DMA). After penyestingan buffe, pointers, andcounter for I / O devices, device controllers directly transfer the entire block of data from local bufferstorage to memory without any intervention from the CPU anymore.Computer System ArchitectureSingle Processor SystemIn a single processor system, there is a CPU capable of executing a series of generalpurposeinstruction, including instruction from a user process. But in general,single processor is accompanied by a microprocessor that exist in each device. For example, ismicroprocessor that converts keystrokes on the keyboard into the code for latersent to the CPU (main processor). Microprocessor such as this, is a microprocessor that can not beon direct access by the CPU.Multiple Processor SystemMultiprocessor system is also called parallel systems or tightly coupled systems. This system has twoor more processors that share the bus and sometimes clock, memory, and other devices. Multiprocessor Systemshas three main advantages.The addition of throughput. With increasing processor, expected that more work can becompleted in a short time. Despite the fact that the ratio of the speed of the system with Nprocessor is not equal to N. When multiple processors working together, in part functionused so that all portions of the processor can run correctly. In addition, the use of sharedresources may also reduce the expected value of multiple processors.Economical in terms of numbers. Multiple processor system can cost less thansingle processor system with the same amount, because of multiple processor systems can shareshared resource, device, and power. If multiple programs use the same data, would be moreefficient if it is stored in one system with multiple processors than many single systemprocessor, karen does not require multiple copies of data.Increased reliability. If the function or command can be well distributed among severalprocessor, the failure of one processor will not be too influential on the performance of the system. Workthe failed processor can be replaced by other processors.Increased reability (increased resistance) of a computer system is critical in manythe application. The ability to continue to provide services (service) to the hardware stage is calledas graceful degradation. System that is able to work better than so-called degradation grecefulas a fault tolerant, where a system can suffer a failure at each of its components butstill capable of providing service. Fault tolerance requires a mechanism that allowserrors that occur can be detected, at diagnosis, and if possible improved.HP Nonstop System (formerly Tandem) using duplicate hardware and software forensure the system is still running. This system consists of many couples the CPU. Two processors ofa pair executing each instruction and then compare the results. If the resultsdifferent, then one CPU fails, and both the CPU is stopped. The process that wasexecuted and delivered to the couple's next CPU and instruction execution is repeated.There are two processing systems on multiprocessor systems. The first process is referred to asasymmetric multiprocessing in which each processor given a special job. A master processorcontrolling the specific duty to control the system, and other processors waiting for orders frommaster processor or particular tasks. The system uses a master-slave relationship.Master processor scheduling and allocating work to slave processors.Figure 1.6 Symmetric MultiprocessingThe most common system used is a symmetric multiprocessing (SMP), where all processorsworking on all parts of the operating system. One example is the Solaris SMP, the commercial versionof UNIX, which was developed by Sun Microsystems. A Solaris machine can be configured tousing a dozen processors are all running Solaris. The advantages of this system, many processescould walk together. N process can be well done if there are N CPUs, without experiencingsignificant loss in performance.Model for symmetric multiprocessing or asymmetric multiprocessing can be determined either byhardware or software. A special hardware can be used to distinguish meltipleprocessor or a software designed to create a slave master and the rest.More recent multiprocessor system called the Blade Server, where multiple processor boards,I / O boards, and networking boards placed on the same chassis. Different systems withclassics are multiprocessor each processor can boot (active) separately andrun the operating system itself.Exercise:1. What is the operating system?2. The emphasis of the Operating System from the viewpoint of the user (single user PC, mainframe, and clientserver).3. Operating system from the viewpoint of the system (resource allocator, program controller).4. Initial or bootstrap program.5. How the CPU executes the program from memory (instruction-execution cycle).6. The fundamental difference storage device.7. Operating mechanism of I / O (device drivers, device controllers, DMA)8. Excess Multiprocessor systems9. Fault tolerance10. Symmetric and assymentric multiprocessor system