2 Plan
Yes—this is still realistic, especially because you solved the material before. You are reactivating knowledge, not learning the entire course from zero.
The crucial change is: do not redo all ten sheets sequentially. They contain research, programming, and practical tasks that are useful pedagogically but inefficient for a written exam. The sheets collapse into about seven examinable clusters.
Priority hierarchy
Priority 1: direct reference-exam matches
These should become automatic:
fork(), process trees and process states- Race conditions, semaphores and synchronization
- Deadlocks, resource-allocation graphs and Banker’s algorithm
- Paging, page-table calculations and address translation
- FAT and i-nodes
- Protocol layers, encapsulation and addressing
- Subnetting, prefixes and forwarding tables
- Dijkstra and Bellman–Ford
Priority 2: likely supporting material
- Page replacement: FIFO, LRU and Clock
- TLB and reference bits
- CPU scheduling
- IP fragmentation
- Disk scheduling and RAID
- Memory fragmentation and allocation algorithms
Priority 3: postpone unless time remains
- Long programming exercises
- Internet research and API-detail questions
- Traceroute, Whois and Wireshark experiments
- Historical hardware details
- Learning-methodology questions
- Reimplementing FAT, Banker, LRU or producer–consumer in code
For programming exercises, understand the underlying algorithm and execute it by hand. Do not spend hours coding it again.
Ten-day core plan
Assume approximately 4–6 focused hours per day. If you have 11–13 days, use the extra days for consolidation rather than expanding the syllabus.
| Day | Focus | Essential exercises and outcomes |
|---|---|---|
| 1 | Diagnostic + processes | Attempt the 2016 exam under time pressure, even if incomplete. Then review Sheet 2 A2(a), A3. Know fork() return values, process trees and the five process states. |
| 2 | Synchronization | Sheet 3 A1, A2 and A4(a–b). Trace a race condition instruction by instruction; repair it with semaphores; understand wait, signal, mutexes and condition variables. |
| 3 | Paging fundamentals | Sheet 4 A5, A7; Sheet 5 A2, A6. Split addresses into page number/offset, construct page-table entries, and calculate normal/inverted page-table sizes. |
| 4 | Advanced paging and virtual memory | Sheet 5 A8; Sheet 6 A1, A4. Multilevel tables, TLB, FIFO/LRU/Clock and reference-bit approximation. |
| 5 | Deadlocks and scheduling | Sheet 8 A2 and A4 first; A6 if time remains. Know the four deadlock conditions, resource graphs, Banker matrices and hand execution. |
| 6 | File systems | Sheet 7 A4 and A5; review the FAT mechanism from A2 without implementing it. Practise i-node maximum-file-size and pathname traversal calculations. |
| 7 | Network stack | Sheet 9 A4 and A9. Reconstruct encapsulation, endpoints and addressing at each layer; then practise IP fragmentation. |
| 8 | Subnetting and forwarding | Sheet 10 A2, A3 and A4. Work entirely from scratch: masks, address counts, network/host IDs, prefixes and longest-prefix matching. |
| 9 | Routing algorithms | Sheet 10 A5 and A6(c); 2016 exam Question 5. Complete Dijkstra tables, derive next hops from predecessor pointers, and perform Bellman–Ford updates. |
| 10 | Timed mixed mock + repair | Assemble unused exercise parts into a 90-minute mock. Mark it immediately. Redo every incorrect question from a blank page without consulting the solution. |
If you have 11–13 days
Use the additional days as follows:
- Day 11: Repair your weakest two topics.
- Day 12: Second full timed mock covering all major clusters.
- Day 13: Light closed-book recall, formulas and common mistakes. No major new topics.
If you have only ten days, the core plan already covers the important material. Do not sacrifice mock-exam practice to squeeze in every supplementary topic.
How to structure each study day
A good five-hour day would look like this:
- 30 minutes: closed-book retrieval
- Write everything remembered from yesterday.
- Reproduce algorithms, definitions and calculation procedures.
- 60–75 minutes: targeted theory recovery
- Use the summary only to repair what you could not recall.
- Avoid reading an entire chapter passively.
- Two 75-minute problem blocks
- Solve selected exercises without notes.
- Work in a handwritten, exam-style format.
- Impose time limits.
- 30 minutes: correction and error log
- Record each substantive mistake.
- Classify its cause: forgotten fact, wrong method, arithmetic error or misreading.
- 20–30 minutes: delayed repetition
- Reattempt one problem from two or three days earlier.
If you have only three hours, keep the same structure but shorten the theory section—not the problem-solving section.
Build an error log, not another long summary
Use five columns:
| Topic | Mistake | Cause | Correct rule/method | Retest |
|---|---|---|---|---|
| Subnetting | Used \(2^h\) as usable-host count | Forgot network/broadcast addresses | Normally \(2^h-2\), unless stated otherwise | Tomorrow |
| Dijkstra | Updated an already-finalized node | Algorithm rule forgotten | Nodes in \(N'\) are final | In two days |
Your final revision should be driven primarily by this list.
Learn reusable solution templates
For each major exam type, have a standard procedure:
fork(): draw one branch per call; label parent and child return values.- Race condition: expand operations into load–modify–store and show the interleaving.
- Banker: calculate available resources, find a satisfiable process, release its allocation, repeat.
- Paging: derive offset bits from page size, then split the logical address.
- I-node: calculate pointers per indirect block before calculating capacity.
- Subnetting: determine required host bits, then prefix length, block size and aligned boundaries.
- Longest-prefix matching: list all matching prefixes and choose the longest.
- Dijkstra: maintain \(N'\), \(D(v)\) and \(p(v)\) systematically.
- Bellman–Ford: minimize link cost plus the neighbour’s advertised distance.
The most important rule
Do not measure progress by “pages read” or “sheets reviewed.” Measure it by:
Can I solve this problem from a blank page, correctly and within exam time?
After about five days, some topics will already feel substantially more familiar. That is the advantage of having solved the sheets before. A focused ten-day recovery is much more plausible than the current breadth of the material makes it appear.