OCR A-Level Computer Science Computer systems (01), June 2025: Question 2

21 marks · Medium difficulty · Extended Response

Compare processor architectures and evaluate thin-client versus thick-client approaches using virtual machines and virtual storage for remote database access.

Practise this question

Question

Question 2 consists of five parts. Part (a)(i) asks for two differences between Von Neumann and Harvard architecture (4 marks). Part (a)(ii) asks for two benefits of pipelining (2 marks). Part (a)(iii) asks for two differences between RISC and CISC architectures (4 marks). Part (a)(iv) asks to describe a benefit of a multicore CPU (2 marks). Part (b) presents a scenario involving remote employees accessing sensitive databases and asks to evaluate thin client versus thick client approaches, including virtual machines, virtual storage, and a justified conclusion (9 marks).
Question text

2 OCR Consultants is designing a computer for a large company.

(a) OCR Consultants is designing the architecture of the CPU.

(i) Compare two differences between the Von Neumann architecture and the Harvard architecture.

Difference 1 …

Difference 2 …

[4]

(ii) OCR Consultants is considering using pipelining.

State two benefits of using pipelining.

Benefit 1 …

Benefit 2 …

[2]

(iii) OCR Consultants can base the CPU around a RISC architecture or a CISC architecture.

Compare two differences between a RISC architecture and a CISC architecture.

Difference 1 …

Difference 2 …

[4]

(iv) Describe a benefit of making the CPU a multicore CPU.

… [2]

(b)* The employees of the large company work remotely from home and use large databases that

store a lot of sensitive data. They need to be able to use the databases to search and sort data

quickly.

OCR Consultants is considering two approaches:

• thin client approach – this is where the database is stored virtually on a remote server.

All of the data processing is also completed on a remote server on virtual machines. The

employees’ computers are only used to capture the data input and display the results from

the virtual machines

• thick client approach – this is where most of the processing is done on the employees’

computer using specialised software to access the remote database.

Discuss the benefits and drawbacks of each of these approaches.

You should include the following in your answer:

• what is meant by virtual machines and virtual storage

• the benefits and drawbacks of a thin client and thick client approach described in the

question

• a conclusion justifying which approach would be most suitable for this scenario. [9]

Mark scheme

Show the mark scheme Mark scheme for Question 2 in five sections. Part (a)(i) awards up to 4 marks for contrasting data/instructions in single vs separate memory, sequential vs concurrent fetching, and single vs separate buses. Part (a)(ii) awards 2 marks for benefits of pipelining (e.g., increased throughput, simultaneous FDE stages). Part (a)(iii) awards 4 marks comparing RISC and CISC (instruction set size, hardware complexity, clock cycles per instruction, pipelining ease). Part (a)(iv) awards 2 marks for multicore benefits (parallel execution, multiple threads at the same time). Part (b) is a 3-band level-of-response grid (up to 9 marks) evaluating definitions of virtual machines/storage, pros and cons of thin/thick clients, and a justified conclusion.

Question Answer Mark Guidance

2 (a) (i) 2 marks for each difference to max 4. 4 Must be a full comparison for 2 marks to be given.

● Von Neumann stores data and instructions in the Each comparison must be comparing the same

same memory unit feature.

● Harvard stores data and instructions in separate

memory units

● Von Neumann fetches data and instructions

sequentially/follows FDE cycle

● Harvard can fetch data and instructions at the same

time

● Von Neumann uses a single bus for both data and

instructions

● Harvard uses different buses for data and instructions

2 (a) (ii) 1 mark for each to max 2. 2 DNA ‘faster’/’more efficient’ on their own.

● Makes the CPU more efficient

● Increases the throughput of the CPU

● Allows parts of the FDE cycle to happen

simultaneously (so increases performance)

● Reduces latency/ idle time

2 (a) (iii) 2 marks for each difference to max 4. 11 4 Must be a full comparison for 2 marks to be given.

● CISC has a larger instruction set Each comparison must be comparing the same

● RISC has a smaller instruction set feature.

● CISC requires more complex hardware/ requires

cooling

● RISC requires less complex hardware/ requires little

cooling

● CISC takes multiple clock cycles to execute an

instruction

● RISC takes one clock cycle to execute an instruction

● CISC tends to use more energy

● RISC tends to use less energy

● CISC has more addressing modes

● RISC has fewer addressing modes

● CISC uses less RAM

● RISC uses more RAM

● CISC is difficult to pipeline

● RISC is easier to pipeline

2 (a) (iv) 1 mark for each to max 2. 12 2 Allow ‘FDE Cycles’ for MP1

● Multiple instructions/threads can be processed at

the same time

● Improves multi-tasking

● Enables parallel processing

● Increases the performance of the computer

2 (b) Mark Band 3 – High level (7-9 marks) 9 Answers may include, but are not limited to, some

of the points below:

The candidate demonstrates a thorough knowledge and

understanding of virtual machines. The material is generally AO1

accurate and detailed.

Virtual Machines:

The candidate is able to apply their knowledge and • A machine that exists as software

understanding directly and consistently to the context • Behaves as another program or piece of

provided. Evidence/examples will be explicitly relevant to the

hardware

explanation.

• Needs a real machine to run on

The candidate provides a thorough discussion which is well • For the user it will be indistinguishable

balanced. Evaluative comments are consistently relevant and

well-considered. Virtual Storage:

• A file will be visible to a virtual machine as

There is a well-developed line of reasoning which is clear and secondary storage

logically structured. The information presented is relevant and • Remote storage that is mounted locally to seem

substantiated. like local storage

• Stored on remote servers.

Mark Band 2 – Mid level (4-6 marks)

Thick/Thin client:

The candidate demonstrates reasonable knowledge and

understanding of virtual machines. The material is generally • Thick client involves all the processing being

accurate but at times underdeveloped. completed on the client. This can result in

minimal use for a server, possibly only storing

common data

The candidate is able to apply their knowledge and • Thin client means the client just deals with user

understanding directly to the context provided although one or input and output which are transmitted to a

two opportunities are missed. Evidence/examples are for the server. The server does all the complex

most part implicitly relevant to the explanation. processing

The candidate provides a sound discussion, the majority of AO2 Application

which is focused. Evaluative comments are for the most part

appropriate, although one or two opportunities for Thin client approach:

development are missed. ● Potentially safer to store the database/sensitive

data on virtual servers as the underlying

There is a line of reasoning presented with some structure. system is not vulnerable to security breaches.

The information presented is in the most part relevant and ● As the virtual machine is software, if it is

supported by some evidence. compromised it can be terminated and

restarted from a safe state

Mark Band 1 – Low Level (1-3 marks) ● More efficient use of hardware as

storage/processing resources are shared//not

The candidate demonstrates a basic knowledge of some

sat idle on a laptop not being fully utilised

aspects of virtual machines. The material is basic and

● Can dynamically shift resources

contains some inaccuracies.

o If a user is doing a processor intense

The candidate makes a limited attempt to apply acquired task more virtual cores can be allocated

knowledge and understanding to the context provided. to them // a user doing low intensity

tasks can have fewer

The candidate provides a limited discussion which is narrow o If a user needs larger storage

in focus. Judgments if made are weak and unsubstantiated. requirements, space can be dynamically

The information is basic and communicated in an allocated to them

unstructured way. ● There is limited offline functionality and users

may only be able to work if there is an internet

The information is supported by limited evidence and the connection.

relationship to the evidence may not be clear. ● It’s difficult to personalise for individual users.

● May need a lot of bandwidth to be able to

0 marks access the large databases and carryout the

No attempt to answer the question or response is not worthy data processing.

of credit.

Thick client approach:

● The computers have their own processing

power to run complex processing locally.

● Often has offline functionality as they can

continue to work without an internet connection

● Devices can be customised to meet the user’s

needs.

● Possibly less secure (than thin client) as the

security software on the devices needs to be

managed.

● May need more advanced devices so therefore

will increase the upfront costs.

AO3 Evaluation

● There are cost savings to using thin client

approaches.

● The ability to dynamically allocate resources

would benefit users who have unusually large

processing jobs as they will not cause

unacceptable wait times.

● As the users are mobile, they are reliant on

network access or mobile access.

o May not allow 3rd party access.

o Mobile data access may be slow or

even non-existent.

● If a user has a fully powered laptop in case of

connectivity issues it will remove the cost

saving advantages

● If there is slow network access the user could

be subject to unacceptable waits due to

latency issues.

How to answer it

CPU Architecture & Client-Server Virtualisation Guide

OCR A-Level Computer Science
Total: 21 Marks

📋 What this question tests

This question assesses fundamental system hardware and architecture principles from Component 1 (Computer Systems):

  • Structural and operational differences between Von Neumann and Harvard computer architectures.
  • Instruction pipelining and its measurable effects on processor performance.
  • Comparative features of RISC and CISC processor designs.
  • How multicore processors enable parallel instruction execution.
  • Extended Evaluation (9 marks): Virtualisation (VMs and virtual storage) combined with an evaluative comparison of Thin Client vs Thick Client implementations for remote access to large, sensitive corporate databases.
4 Marks

Part (a)(i): Von Neumann vs Harvard Architecture

Compare two differences between the Von Neumann architecture and the Harvard architecture.

✅ Model Answers (Choose Two Pairs)

Each valid difference requires addressing both architectures for the same operational feature (2 marks per difference):

  • Memory Structure: Von Neumann stores data and instructions in the same physical memory unit / address space, whereas Harvard stores data and instructions in separate memory units / address spaces.
  • Buses: Von Neumann shares a single common bus system (data & address bus) for both instructions and data, whereas Harvard has dedicated, separate buses for instructions and data.
  • Instruction Fetching: Von Neumann must fetch instructions and data sequentially (one after another) through the FDE cycle, whereas Harvard can fetch instructions and read/write data simultaneously / at the same time.

🧠 Exam Technique: "Direct Contrast"

  • The "Dual Statement" Rule: A comparison mark scheme gives 0 marks for writing only about one architecture. You must state Feature X in Von Neumann and Feature X in Harvard to earn the 2 marks per difference.
  • Use connective words like "whereas", "in contrast", or "while" to ensure your examiner sees both halves of the comparison immediately.

❌ Common Errors & Examiner Traps

  • Mismatched Comparisons: Saying "Von Neumann has one memory, whereas Harvard has separate buses." This fails to compare the same feature and scores 0.
  • Vague Speed Claims: Writing "Harvard is faster than Von Neumann" without explaining why (simultaneous instruction and data fetching). Speed alone is not a valid architectural difference.
2 Marks

Part (a)(ii): Benefits of Pipelining

State two benefits of using pipelining.

✅ Mark Scheme Answers (Any Two)

  • Increases the throughput of the CPU / number of instructions completed per unit of time.
  • Allows multiple stages of the Fetch-Decode-Execute (FDE) cycle to occur simultaneously / concurrently.
  • Reduces idle time / CPU component latency (e.g., the ALU is active while the decode unit works on the next instruction).
  • Makes CPU processing more efficient (must be qualified with throughput/concurrency).

💡 Key Knowledge: What is Pipelining?

In non-pipelined CPUs, execution hardware sits idle while fetching and decoding take place. Pipelining divides the instruction cycle into discrete stages. While Instruction 1 is Executing, Instruction 2 is being Decoded, and Instruction 3 is being Fetched.

❌ Mark Scheme Guidance Warning

DNA ("Do Not Accept"): The standalone words "faster" or "more efficient" without further explanation earn 0 marks. You must state what is faster or improved (e.g., higher instruction throughput or reduced processor idle time).

4 Marks

Part (a)(iii): RISC vs CISC Architecture

Compare two differences between a RISC architecture and a CISC architecture.

Feature CISC (Complex Instruction Set) RISC (Reduced Instruction Set)
Instruction Set Size Large, complex instruction set with many specialised instructions. Small, simplified instruction set with fewer, basic instructions.
Cycle Execution Instructions take multiple clock cycles to execute. Instructions execute in a single clock cycle.
Hardware & Cooling More complex internal circuitry; produces more heat, requiring active cooling. Simpler hardware design; requires less cooling / runs cooler.
Power Consumption Tends to consume more power/energy. Consumes less power/energy (ideal for mobile devices).
Addressing Modes Supports many complex addressing modes. Supports fewer / simple addressing modes.
RAM Usage & Code Size Compact machine code; uses less RAM. Longer program code; uses more RAM.
Pipelining Suitability Difficult to pipeline due to variable instruction lengths/cycles. Much easier to pipeline due to fixed-length single-cycle instructions.

🧠 Examiner Note for Full Marks

Just like part (a)(i), this is a 4-mark comparison question. You must explicitly pair the CISC characteristic with the corresponding RISC characteristic for the same feature to secure each 2-mark pair.

2 Marks

Part (a)(iv): Benefits of Multicore CPUs

Describe a benefit of making the CPU a multicore CPU.

✅ Scoring the 2 Marks

  • Mark 1 (Mechanism): Multiple instructions / threads / tasks / FDE cycles can be processed simultaneously / at the same time.
  • Mark 2 (Outcome): This enables parallel processing / improves multitasking performance / increases overall computer system throughput.

🧠 Note the Command Word: "Describe"

Because the command word is Describe (worth 2 marks), stating simply "It can do parallel processing" will only get 1 mark. You must describe how it works (independent cores simultaneously executing different instructions/threads) to unlock both marks.

9 Marks

Part (b)* Extended Evaluation: Thin Client vs Thick Client

Discuss the benefits and drawbacks of thin client and thick client approaches for remote employees searching and sorting large databases of sensitive data.
Must include: definition of virtual machines & virtual storage; benefits and drawbacks of each approach; justified conclusion.

📐 Mark Scheme Band Breakdown (Level of Response)

This is a synoptic Level-of-Response question assessed across three Assessment Objectives:

  • Level 3 (7–9 marks): Thorough knowledge of VMs and virtual storage (AO1). Clear, balanced application directly tied to remote workers and sensitive corporate database handling (AO2). Well-reasoned, fully justified conclusion choosing one architecture (AO3).
  • Level 2 (4–6 marks): Sound knowledge of virtualisation with some application to remote clients. Discussion has some balance, with an attempted conclusion.
  • Level 1 (1–3 marks): Basic knowledge; generic thin/thick points without meaningful application to the scenario. Unsupported conclusion.

AO1 Core Concepts

  • Virtual Machine (VM): A software-based emulation of a physical computer system that executes programs and behaves like physical hardware, running on host hardware via a hypervisor.
  • Virtual Storage: The pooling and abstraction of physical storage devices across a network (e.g. SAN / Cloud) to appear as a single, locally mounted storage drive to the client or VM.
  • Thin Client: Client device handles solely I/O (display, keyboard, mouse). All intensive computation, sorting, searching, and operating systems execute on the central server.
  • Thick Client: Client computer runs the OS and applications locally; all sorting, processing, and analysis occur using the client's own CPU/RAM.

AO2 Thin Client Analysis

Benefits:

  • Superior Security for Sensitive Data: Data never physically resides on remote laptops. If an employee's device is stolen or compromised, no sensitive customer data is leaked.
  • Dynamic Resource Scaling: The central server can dynamically spin up more vCPU/RAM resources to execute demanding search/sort queries across the massive database.
  • Centralised Integrity: VMs can be quickly reverted to a known clean snapshot if malware strikes a session.

Drawbacks:

  • Network Dependency: Requires constant, stable high-bandwidth internet. Workers with poor connections suffer latency or cannot work.
  • No Offline Functionality: Absolutely no searching or processing can take place without an active server link.

AO2 Thick Client Analysis

Benefits:

  • Local Performance: Client uses its own high-end multicore CPU/RAM to sort and search locally cached data without network lag.
  • Offline Resilience: Employees can continue reviewing or formatting database extracts while travelling without internet access.

Drawbacks:

  • Severe Security Vulnerability: Sensitive data stored or cached locally on portable home laptops is vulnerable to physical theft and endpoint malware.
  • Massive Data Transfers: Pulling gigabytes of raw database files over domestic broadband strains home bandwidth compared to streaming screen pixels.
  • Hardware Cost: Requires procuring expensive, high-spec laptops for every single remote employee.

AO3 Model Justified Conclusion

Top-band recommendation: The Thin Client approach is undeniably the most suitable choice for this company.

  • Primary Justification: The question highlights that the company manages sensitive data. In a remote work model, protecting against data breaches is paramount; thin clients ensure data never leaves the fortified data centre.
  • Processing Efficiency: Large database sorting and filtering operations are significantly faster when executed locally on high-performance clustered enterprise server hardware rather than disparate, low-power remote laptops.
  • Cost & Administration: Upgrading centralised server resources is substantially more cost-effective and manageable for IT consultants than continuously replacing client laptops across remote locations.

❌ What Distinguishes Top-Level Responses (Examiner Insight)

  • Do not write generic definitions: Weaker students talk about thin/thick clients in a vacuum. High-scoring candidates constantly tie points back to "remote home workers", "searching and sorting", and "sensitive database records".
  • Do not sit on the fence: You will not achieve Band 3 without an explicit, substantiated recommendation in your conclusion explaining why one method wins out over the other in this specific context.

Topics

1.1 The characteristics of contemporary processors, input, output and storage devices · 1.2 Software and software development · 1.3 Exchanging data · 1.1.1 Structure and function of the processor · 1.1.2 Types of processor · 1.1.3 Input, output and storage · 1.2.1 Systems Software · 1.3.3 Networks

Question and mark scheme from the OCR A-Level Computer Science examination, Computer systems (01), June 2025. QuestionVault is an independent revision resource; questions remain the copyright of the awarding body.