Bottom-Up, Top-Down, Follow the Flow — and Choosing a Technique
Where you are: Module 6 — Isolation Techniques · Lesson 6.3 · Based on Chapter 6 (§6.4, §6.7) of Dr. Tahseen Al Doori's Troubleshooting Essentials.
What you'll be able to do
- Apply the bottom-up approach, starting from power and cables, and say when it is the best choice.
- Apply the top-down approach, starting from the application, and say when it is the best choice.
- Follow the flow of data or a signal from source to destination, checking each stage.
- Select the most suitable isolation technique for a given situation using the book's table, and combine techniques where useful.
Watch: Bottom-Up, Top-Down, Follow the Flow — and Choosing a Technique
The video shows the two directions of testing, then traces a fictional print job from an accounts PC to a printer until the stage where it stops. It closes with the book's "choosing a technique" table. Questions to keep in mind:
- When does starting at power and cables waste time?
- What do you check at each stage when following the flow?
- Which technique does the book recommend when there is no obvious starting point?
Following the path
Chapter 6's fourth technique, following the path, gives a logical order of testing. The book describes three approaches.
Bottom-up
Start at the physical level (power, cables, hardware) and work upwards to software and settings.
The book recommends bottom-up when hardware or connectivity is suspected, and its table makes it the default when there is no obvious starting point: begin with power and cables. The logic is that everything above depends on the physical level, so a physical fault produces symptoms at every level above it. Bottom-up also echoes Chapter 3's advice to consider "working from the physical layer upwards", and it is the natural order for the network layers you will meet in Module 9.
Top-down
Start at the application the user sees and work downwards.
The book recommends top-down when only one application misbehaves. If every other program works, the power, cables and network are almost certainly fine, and starting at the bottom would waste time testing things that are clearly working. Begin with the application: does it fail with every file or just one, for every user or just one? Then examine its settings and add-ons before moving down.
Follow the flow
Trace data or a signal from its source to its destination, checking each stage. The book's example is a print job: keyboard, computer, network, print server, printer. At each stage, ask: has the data arrived here correctly? The first stage where the answer is no is where the fault lies (between that stage and the one before it).
Follow the flow is a natural partner of half-split: once you have listed the stages, you can test the middle stage first rather than walking from one end.
Worked example: following a print job
(Fictional scenario.) The accounts PC sends invoices to the office printer, and nothing comes out. No error appears on the PC, and the quick-win checks have changed nothing. The path is PC → network switch → print server → printer.
| Time | Step | Action and result |
|---|---|---|
| 14:00 | Test | Stage 1, the PC: its own print queue is empty, so the job has left the computer. |
| 14:03 | Test | Stage 2, the print server: its queue shows the invoice job has arrived. The PC, switch and network up to the server are working. |
| 14:05 | Test | Stage 3, server to printer: the job is stuck in the server's queue, and the printer shows as offline. The fault lies between the print server and the printer. |
| 14:08 | Test | The printer prints a test page from its own menu, so the printer itself works (a split described in Chapter 11). |
| 14:10 | Theory | The printer's network cable or switch port is faulty. |
| 14:12 | Implement | Substitution: a known-good cable is fitted; the printer comes online. |
| 14:14 | Verify | The waiting invoices print; the accounts user confirms. |
| 14:20 | Document | The faulty cable is labelled and removed from use; the ticket records cause and fix. |
Three techniques worked together: following the flow found the stage, a test page split the printer from the link, and substitution proved the cause.
Choosing a technique
The book provides a table that matches common situations to the most suitable technique:
| Situation | Most suitable technique |
|---|---|
| A long chain of components or stages | Half-split |
| Spare parts or a second device are available | Substitution |
| An identical working system exists | Comparison |
| The problem began after a known change | Rollback |
| Many add-ons, applications or accessories are installed | Remove and restore |
| There is no obvious starting point | Bottom-up from power and cables |
| Only one application is affected | Top-down |
Read the table as guidance, not as a rigid rule. Real situations often match more than one row. A fault that began after an update, on a laptop with spares available, matches both rollback and substitution; the book's review question on a driver update suggests starting with the change, because rollback tests the most likely cause directly. When several rows apply, start with the technique that tests your most probable theory most cheaply and safely, and remember the principle change one thing at a time.
Chapter 6's key takeaway sums up the whole module: isolation turns a large problem into a small one. Halve the system, swap in known-good parts, compare with a working system, strip back to the minimum or undo the last change, and the cause will be cornered.
Try it
Goal: practise choosing and sequencing techniques on paper.
For each scenario, write (a) the most suitable technique from the table, (b) one sentence explaining why, and (c) your first two concrete tests.
- A spreadsheet program crashes whenever one user opens it; every other program on their PC works.
- A desktop PC is completely silent after an office move; there is no obvious starting point.
- A kiosk computer, one of twenty identical kiosks, shows the wrong time zone.
- A web browser became very slow after the user installed several extensions.
- A signal passes through twelve connectors and cables between a control room and a display wall, and the display shows nothing.
You have finished when all five scenarios have a technique, a justification and two tests, and you have checked your choices against the table.
Knowledge check
When does the book recommend the top-down approach?
A desktop PC fails to start after being moved, and there is no obvious starting point. According to the book, where should you begin?
A print job leaves the PC and reaches the print server, but never reaches the printer. Using follow the flow, where does the fault lie?
A laptop's browser became slow after many extensions were installed. Which technique fits best?
A fault affects a chain of 30 cable segments and connectors. Which technique does the book's table suggest?
Answers explained
- B. Top-down is for a single misbehaving application. Hardware or connectivity suspicions and the no-starting-point case call for bottom-up; spare parts suggest substitution.
- D. With no obvious starting point, the table says bottom-up from power and cables, which also suits a device that has just been moved. Top-down needs a working system; comparison and rollback do not fit a move.
- A. The data arrived correctly at the print server but not at the printer, so the fault lies between those two stages. The earlier stages were shown to work.
- C. Many add-ons installed is the table's case for remove and restore: disable them, then restore one at a time. Half-split could speed up restoring, but the situation matches remove and restore.
- B. A long chain of components or stages is the table's case for half-split, which needs at most 5 tests for 30 items. The others suit single applications, recent changes or many add-ons.
Key takeaways
- Bottom-up starts at power and cables; use it when hardware or connectivity is suspected or there is no obvious starting point.
- Top-down starts at the application; use it when only one application is affected.
- Follow the flow traces data from source to destination; the first stage where the data is missing locates the fault.
- Use the book's table to match situation to technique, and combine techniques when the problem calls for it.
Sources
- Dr. Tahseen Al Doori — Troubleshooting Essentials, University Edition (2026), §6.4, §6.7, Chapter 6 Key Takeaway and Review Questions (with cross-references to §3.2, §11.1 and the Glossary)
