Cambridge A Level Information Technology (ICT): Hardware and Software — Revision Notes (9626)

Related: Topic 2 study guide

Condensed, exam-focused notes for Topic 2 of Cambridge AS & A Level Information Technology (9626), 2025-2027 series.

2.1 Mainframes and supercomputers

	Mainframe	Supercomputer
Built for	High-volume transaction throughput	Raw processing speed
Example use	Census, transaction processing	Weather forecasting, quantum mechanics
Metrics	RAS and fault tolerance apply to both classes; MIPS is conventionally quoted for mainframes	RAS and fault tolerance apply to both classes; FLOPS is conventionally quoted for supercomputers

• Both RAS (reliability/availability/serviceability) and fault tolerance are characteristics of both classes, not a discriminator between them — the syllabus gives one shared characteristics list for mainframes and supercomputers together.
• RAS: reliability (how rarely it fails), availability (how much of the time it’s usable, accounting for planned maintenance), serviceability (how quickly/easily it’s repaired).
• Fault tolerance: ability to keep running, often in a degraded state, when a component fails.
• MIPS and FLOPS are both performance metrics usable for either class; MIPS (instructions per second) is conventionally quoted for a mainframe’s transaction workload, and FLOPS (floating-point operations per second) for a supercomputer’s computation-heavy workload — a matter of typical workload, not a rule.
• Keep the use case attached to each term — mainframe/supercomputer confusion is a common error.

2.2 System software

• Types: compilers, interpreters, linkers, device drivers, operating systems, utilities.
• Compiler: translates and stores an entire program before execution. Interpreter: translates and executes line by line (why development-stage code is often run through an interpreter first).
• Running on different systems: the syllabus asks how a high-level program is translated to run on different computer systems using interpreters (the same source code runs anywhere a suitable interpreter exists) and cross compilers (a compiler on one type of computer produces machine code for a different type).

2.3 Utility software

• Named types: anti-virus, back-up, data compression, disk defragmentation, formatting (low-level, partitioning, high-level), file copying/deleting, file management systems, disk management systems.
• System software manages/translates for the computer itself; utility software performs a specific maintenance task for the user. Mixing up the two categories weakens an otherwise correct answer.

2.4 Custom-written vs off-the-shelf software

• Off-the-shelf: lower cost, available immediately, well tested by a large user base; may include unneeded features, can’t be tailored exactly.
• Custom-written: matches exact requirements, adaptable as needs evolve; costs more, takes time, depends on ongoing developer support.
• Also: proprietary vs open-source software.

2.5 User interfaces

• Types: command line, graphical, dialogue, gesture-based.
• Match each interface type to a suitable real-world use (e.g. command line for a systems administrator, graphical for a general consumer device), not just a definition.

Exam technique for this topic

Nearly every sub-topic in this topic ends in “advantages and disadvantages,” so be ready to weigh options, not just define them — build revision around comparison tables (mainframe vs supercomputer, interpreter vs compiler, proprietary vs open-source, GUI vs command line) rather than isolated definitions. When a question describes a scenario and asks you to justify a software or hardware choice, weigh both options against the scenario’s specific details (budget, timescale, how unusual the workflow is) rather than asserting one option is simply “better” — a generic “it’s more advanced” answer earns little credit compared with a scenario-linked justification.

Worked example: mainframe or supercomputer?

A national statistics agency needs to process millions of census transactions reliably, ensuring no data is lost even during scheduled maintenance windows. A mainframe suits this scenario because census processing is one of its named uses and it is built for high-volume transaction throughput, exactly matching a census’s need to process enormous numbers of individual records accurately and continuously; it also offers the RAS — high reliability, near-continuous availability, and quick serviceability — that the no-data-loss requirement demands (RAS is a characteristic of both classes, so it supports the choice rather than deciding it). A supercomputer, by contrast, would be the wrong choice here despite its greater raw processing power, because the census task is not primarily computation-heavy in the way a climate simulation or quantum mechanics calculation is — it needs sustained high-volume transaction throughput, not maximum floating-point operations per second (FLOPS). This kind of matched, scenario-specific justification — not simply “mainframes are for business, supercomputers are for science” — is what a strong answer on this sub-topic looks like.

Worked example: choosing an interface type

A public information kiosk in a museum needs to be usable by visitors of all ages and technical abilities with no training. A graphical user interface with large touch-friendly icons suits this scenario because it requires no prior knowledge of commands and gives immediate visual feedback. A command line interface would be entirely unsuitable here — it demands memorised syntax that a casual, untrained visitor cannot be expected to know — but would suit a systems administrator managing many servers, where speed, precision and scriptability outweigh ease-of-learning concerns. Being able to generate this kind of matched reasoning for any interface type and any described user group, rather than reciting a fixed list of “GUI is easy, CLI is powerful,” is what distinguishes strong answers.

Self-test

1. What distinguishes a mainframe from a supercomputer, in terms of what each is built for?
2. Define reliability, availability and serviceability (RAS) as three separate concepts.
3. What is the key difference between a compiler and an interpreter?
4. Give one advantage and one disadvantage of custom-written software.
5. Which interface type would suit a systems administrator, and why?

Answers: 1. Mainframe: high-volume transaction throughput (census, transaction processing). Supercomputer: raw processing speed on computation-heavy problems. 2. Reliability = how rarely it fails; availability = how much of the time it’s usable; serviceability = how quickly/easily it’s repaired. 3. A compiler translates and stores the whole program before execution; an interpreter translates and executes line by line. 4. Advantage: matches exact requirements/adaptable. Disadvantage: higher cost/longer development time. 5. Command line — allows precise, scriptable control suited to a technically skilled user.

Why this topic rewards comparison over recall

Topic 2’s content is unusually dense with named categories (mainframe, supercomputer, system software, utility software, custom-written, off-the-shelf, four interface types), and it can be tempting to revise each as an isolated definition to memorise. Because almost every sub-topic asks for advantages and disadvantages, a different approach works better: for every category, hold in mind at least one alternative it could be confused with, and one concrete scenario where choosing between the two matters. This comparison-first revision style — rather than a long list of separate definitions — is what turns recognition of a term into the ability to justify a choice, which is what nearly every question in this topic actually tests.