COS3721 Operating Systems Architecture Original Study Mastery

COS3721 Operating Systems Architecture Original Study Mastery COS3721 — Operating Systems and Architecture Original Study Mastery Workbook — 2026 UNISA NQF Level 7 module — independently authored revision resource UNISA describes COS3721 as a 12-credit, NQF Level 7 year module covering operating-system functionality including CPU scheduling, process coordination and concurrency, deadlocks, memory management, protection/security, and distributed systems. This workbook is an original study resource. It does not reproduce the linked paid Stuvia notes, assignments, past-exam questions, or answer keys. Area High-yield rule Scheduling Turnaround = completion − arrival; waiting = turnaround − CPU service; response = first service − arrival. Concurrency Protect shared state with appropriate synchronization; distinguish race conditions, deadlock, starvation, and priority inversion. Deadlock Four necessary conditions: mutual exclusion, hold-and-wait, no preemption, circular wait. Paging Virtual page → physical frame; TLB caches recent translations; page faults require OS handling. Replacement FIFO uses arrival age; LRU uses recent use; working-set ideas exploit locality. Files Directories provide names; metadata structures describe objects; allocation and journaling affect performance/recovery. I/O Interrupts reduce busy waiting; DMA reduces CPU involvement in bulk transfers; buffering absorbs rate mismatch. Security Authentication = identity; authorization = permitted actions; least privilege = minimum necessary access. Distributed systems Expect communication delay, partial failures, replication/consistency trade-offs, and imperfect clocks. COS3721 Operating Systems & Architecture — Original Study Mastery 2026 Page 2 1. Operating-System Foundations 1. OS role — Which statement best describes an operating system? A. A program used only to edit files B. Software that manages hardware/resources and provides services to applications C. A compiler only D. A physical memory module Answer: B. Software that manages hardware/resources and provides services to applications Rationale: An OS manages resources such as CPU time, memory, storage, and devices while providing abstractions and services to programs. 2. System calls — Why are system calls important? A. They provide a controlled interface through which programs request OS services B. They replace all hardware C. They increase RAM physically D. They eliminate process states Answer: A. They provide a controlled interface through which programs request OS services Rationale: System calls provide the controlled interface between user programs and privileged operating-system services. 3. Dual mode — What is the main purpose of user mode and kernel mode? A. Prevent all interrupts B. Separate application privileges from privileged OS operations C. Make every program a kernel D. Remove memory protection Answer: B. Separate application privileges from privileged OS operations Rationale: Dual-mode operation limits privileged instructions and sensitive resources to trusted kernel code. 4. Interrupts — An interrupt is useful because it can: A. Notify the CPU that an event needs attention B. Permanently stop the CPU C. Delete the scheduler D. Disable all devices Answer: A. Notify the CPU that an event needs attention Rationale: Interrupts let hardware or software events request processor attention without continuous polling. 5. Multiprogramming — The principal goal of multiprogramming is to: A. Keep the CPU productive when another job waits B. Run only one process C. Disable I/O D. Eliminate memory management Answer: A. Keep the CPU productive when another job waits Rationale: Multiprogramming overlaps computation and waiting to improve resource utilization. COS3721 Operating Systems & Architecture — Original Study Mastery 2026 Page 3 6. Protection — Protection in an OS primarily concerns: A. Controlling access to resources B. Increasing clock frequency C. Formatting displays D. Replacing scheduling Answer: A. Controlling access to resources Rationale: Protection mechanisms control which subjects may access which resources and under what conditions. 7. Microkernel — A microkernel design generally aims to: A. Put every service in one monolithic kernel B. Keep a small privileged core and move more services outside it C. Eliminate IPC D. Eliminate process isolation Answer: B. Keep a small privileged core and move more services outside it Rationale: Microkernels minimize kernel-resident functionality and commonly rely on IPC for communication between components. 8. Process definition — A process is best described as: A. A program in execution together with its execution context/resources B. A source-code comment C. A disk sector D. A CPU register only Answer: A. A program in execution together with its execution context/resources Rationale: A process includes executable code plus execution state and associated resources. 2. Processes and Process Coordination 9. PCB — Which information is commonly stored in a process control block? A. Process state, program counter, registers, scheduling and accounting information B. Only source code C. Only file names D. Only network addresses Answer: A. Process state, program counter, registers, scheduling and accounting information Rationale: The PCB stores information needed to manage and resume a process. 10. Blocked state — A running process requesting I/O will typically move to: A. Ready B. Waiting/blocked C. New forever D. Terminated immediately Answer: B. Waiting/blocked Rationale: A process waiting for I/O cannot continue execution and therefore becomes blocked. COS3721 Operating Systems & Architecture — Original Study Mastery 2026 Page 4 11. Context switch — A context switch occurs when the OS: A. Saves one execution context and restores another B. Changes disk capacity C. Recompiles the OS D. Deletes the PCB Answer: A. Saves one execution context and restores another Rationale: The OS saves the current process state and loads another saved execution context. 12. Shared memory IPC — Which IPC method can provide communication through a shared memory region? A. Shared memory B. Only DNS C. Only a compiler D. Only a printer Answer: A. Shared memory Rationale: Shared-memory IPC maps a common region into cooperating processes; synchronization is then required. 13. Message passing — Message passing is particularly useful when processes: A. Need structured communication without sharing an address space B. Must share every variable C. Cannot block D. Have no OS Answer: A. Need structured communication without sharing an address space Rationale: Message passing provides explicit communication operations and can cross address-space boundaries. 14. Orphan — An orphan process is a process whose: A. Parent terminates before it B. Child terminates first C. PCB is always deleted D. CPU is permanently disabled Answer: A. Parent terminates before it Rationale: An orphan is reparented according to the OS's process-management model. 15. Zombie — A zombie process is typically one that: A. Has terminated but still has an entry holding exit status until its parent collects it B. Is running without a PCB C. Has never started D. Owns all system memory Answer: A. Has terminated but still has an entry holding exit status until its parent collects it Rationale: A terminated child can remain as a process-table entry until the parent performs the appropriate wait operation. COS3721 Operating Systems & Architecture — Original Study Mastery 2026 Page 5 16. Process creation — A parent process creating a child normally results in: A. A new process with its own execution context, while selected resources may be inherited/shared by design B. A new CPU C. Automatic deadlock D. Deletion of the parent Answer: A. A new process with its own execution context, while selected resources may be inherited/shared by design Rationale: Process creation creates a separate execution entity; exact inheritance semantics depend on the OS. 3. Threads and Concurrency 17. Thread — A thread is best viewed as: A. A unit of execution within a process B. A separate hard disk C. A page table only D. A compiler token Answer: A. A unit of execution within a process Rationale: Threads can share a process address space while maintaining separate execution state. 18. User threads — One common advantage of user-level threads is: A. Low management overhead in some designs B. Guaranteed multicore parallelism in every model C. No scheduling D. No synchronization needs Answer: A. Low management overhead in some designs Rationale: Thread libraries can perform some operations without kernel transitions, depending on the model. 19. Kernel threads — Kernel-supported threads can be advantageous because the kernel can: A. Schedule threads directly across processors B. Eliminate context switches C. Guarantee no blocking D. Remove address spaces Answer: A. Schedule threads directly across processors Rationale: Kernel visibility allows individual threads to be scheduled and potentially run in parallel. 20. Race condition — A race condition occurs when: A. The result depends on timing/interleaving of concurrent operations B. CPU frequency is too low C. A file is compressed D. A process never executes Answer: A. The result depends on timing/interleaving of concurrent operations Rationale: Unsynchronized concurrent accesses can produce different outcomes depending on execution order. COS3721 Operating Systems & Architecture — Original Study Mastery 2026 Page 6 21. Concurrency — Concurrency means: A. Multiple tasks make progress during overlapping time intervals B. Only one instruction exists C. Every task executes simultaneously on one CPU D. No task can block Answer: A. Multiple tasks make progress during overlapping time intervals Rationale: Concurrency is overlapping progress; simultaneous execution requires multiple execution resources. 22. Thread pool — A thread pool is useful because it: A. Reuses a bounded set of worker threads for submitted tasks B. Creates an unlimited thread for every request C. Eliminates queues D. Disables synchronization Answer: A. Reuses a bounded set of worker threads for submitted tasks Rationale: Pools reduce creation overhead and can limit concurrency. 23. Many-to-one — In a many-to-one user-thread model: A. Many user threads map to one kernel thread B. One user thread maps to many kernels C. Every thread is a process D. No user threads exist Answer: A. Many user threads map to one kernel thread Rationale: Many user threads are multiplexed over one kernel execution entity, limiting parallelism. 24. Priority inversion — Priority inversion occurs when: A. A higher-priority task is delayed by a lower-priority task holding a needed resource B. All priorities are equal C. A page is replaced D. A file is locked Answer: A. A higher-priority task is delayed by a lower-priority task holding a needed resource Rationale: Priority inversion is a scheduling/synchronization problem in which lower-priority work can indirectly delay higher-priority work. 4. CPU Scheduling COS3721 Operating Systems & Architecture — Original Study Mastery 2026 Page 7 25. FCFS — First-come, first-served scheduling is characterized by: A. Serving processes in arrival order B. Always selecting the shortest burst C. Always preempting D. Ignoring arrival time Answer: A. Serving processes in arrival order Rationale: FCFS uses arrival order and is usually non-preemptive. 26. SJF — Shortest-job-first tends to minimize: A. Average waiting time when burst lengths are known accurately B. Number of processes C. Memory size D. Disk capacity Answer: A. Average waiting time when burst lengths are known accurately Rationale: Under classic assumptions, SJF minimizes average waiting time. 27. Round robin — Round-robin scheduling is primarily associated with: A. A time quantum and cyclic allocation of CPU time B. No preemption C. Only batch systems D. Ignoring ready processes Answer: A. A time quantum and cyclic allocation of CPU time Rationale: Round robin gives each ready process a time slice. 28. Priority — A major issue with priority scheduling is: A. Starvation of low-priority processes B. No process can ever run C. It cannot be preemptive D. It requires no priorities Answer: A. Starvation of low-priority processes Rationale: Low-priority processes can wait indefinitely unless fairness mechanisms such as aging are used. 29. Aging — Aging is used to: A. Gradually increase the priority of waiting processes B. Reduce RAM C. Increase disk latency D. Disable interrupts Answer: A. Gradually increase the priority of waiting processes Rationale: Aging improves fairness by preventing indefinite starvation. COS3721 Operating Systems & Architecture — Original Study Mastery 2026 Page 8 30. Turnaround — Turnaround time is commonly: A. Completion time minus arrival time B. Burst time minus waiting time C. Response time plus RAM D. Arrival time minus completion time Answer: A. Completion time minus arrival time Rationale: Turnaround measures total elapsed time from arrival/submission to completion. 31. Response — Response time measures: A. Time from arrival/request until first service or r