Paper 3: Cloud architecture and capacity
- 2026-11-21
Paper 3: Cloud architecture and capacity
Chapter practice with ten items, answers and a one-line note for each, about 25 minutes.
Paper 3 Ten items on this chapter with answers for immediate marking.
01 Paper details (check answers afterwards)
| Item | Detail |
|---|---|
| Questions | 10 |
| Suggested time | 25 |
| Difficulty mix | Foundation 4 · Intermediate 2 · Challenge 2 · Error redo 2 |
| Chapter | Chapter: Cloud computing and virtualisation |
| Level | Challenge |
02 Questions (10 items)
- FoundationTwo services in series each achieve 99.9%. What is the combined availability and yearly downtime?
- FoundationAt 99.95% availability over 8,760 hours, what is the allowed downtime?
- FoundationA hundred VMs average 15% utilisation. At a 75% target, how many hosts are needed?
- Foundation1.5 TB of data is stored in three replicas. How much space is really used?
- IntermediateA host has 64 cores and 512 GB RAM, each VM needs 4 vCPU and 8 GB, with 20% headroom. How many VMs?
- IntermediateOn-demand is $0.20 per hour at 730 hours a month; a one-year reservation costs $1,200. What is the saving?
- ChallengeUsing 1,024 GB per TB at $0.09 per GB, what does 1 TB of egress cost?
- ChallengeA host has 24 GB RAM but runs four VMs of 8 GB each. What happens?
- Error redoAuto-scaling goes from 0 to 10 machines, each taking 90 seconds to boot. How long until ready?
- Error redoA bank keeps core banking on-premise and puts its web front end in a public cloud. What is this called?
03 Answers and one-line notes
| # | Answer | One-line note |
|---|---|---|
| 1 | 99.8001%, about 17.51 hours a year. | 0.999×0.999=0.998001; chaining lowers availability below every single link. |
| 2 | 4.38 hours. | 8760×0.0005=4.38; the gap is 0.05%, not 5%. |
| 3 | Twenty. | 100×0.15/0.75=20, total demand divided by target utilisation. |
| 4 | 4.5 TB. | 1.5×3=4.5; replication triples the footprint and cost. |
| 5 | Twelve, because CPU binds first (64×0.8/4=12.8); RAM alone would allow 51. | Capacity is limited by the tightest resource, not by total memory. |
| 6 | $552.00, since on-demand totals $1,752. | 0.20x730x12=1752 and 1752-1200=552. |
| 7 | $92.16. | 1024×0.09=92.16; convert TB with 1,024. |
| 8 | Memory overcommit: 32 GB requested against 24 GB, so swapping hurts performance. | Overcommit boots fine but collapses under simultaneous load. |
| 9 | About 90 seconds in parallel, or 900 seconds if booted one by one. | Parallel and serial differ tenfold, so plan for start-up latency. |
| 10 | A hybrid cloud. | Hybrid placement balances compliance against elasticity. |
04 What to do after marking
- After marking, redo only the wrong items and write down which step went wrong (formula, unit, or misread number).
- Repeat the same item two days later; move to the next paper only when both attempts are correct.
- Parents only need the answer and note columns; there is no need to explain every item.
Want the whole chapter? See all four difficulty papers on the chapter page.Tutor Registration →
Questions and answers are written originally by the TutorZone team and are not taken from any school, publisher or examination authority paper. Every value is checked by computer.



