Chapter 42 — Security architecture and engineering: applied research

Chapter 42 — Security architecture and engineering: applied research

Lesson map

Lesson 42 published. Map, example, EX124.

flowchart TD
  N0["Learning outcomes"]
  N1["42.1 A question from practice"]
  N2["42.2 Responsible data use"]
  N3["Synthesis"]
  N0 --> N1 --> N2 --> N3
  N3 --> EX["Worked example"]
  EX --> Q["Exercise EX124"]

Logical drawing

flowchart LR
  N0["Learning outcomes"]
  N1["42.1 A question from practice"]
  N2["42.2 Responsible data use"]
  N3["Synthesis"]
  N0 --> N1 --> N2 --> N3 --> Q["EX124"]

[Learning outcomes]
    |
    v
[42.1 A question from practice]
    |
    v
[42.2 Responsible data use]
    |
    v
[Synthesis]
    |
    v
[Worked example] --> [Exercise EX124]

Learning outcomes

You will (1) formulate an applied research question from engineering practice; (2) design a small study using Corvia data responsibly; (3) report findings with limits.

42.1 A question from practice

Sofia's applied research project asks whether short-lived device certificates, introduced in chapter 31, reduced the window during which retired validators could still authenticate. The question comes from practice and matters to Corvia. She defines the outcome as days between retirement and the last successful authentication, compares devices retired before and after the change, and lists rivals such as a simultaneous improvement in asset records. Herley and van Oorschot (2017) warn against security claims that cannot be tested; her question names exactly what would show no effect.

42.2 Responsible data use

The study uses Corvia's operational logs, which contain device identifiers but no passenger data. Sofia agrees with Daniel and the data protection officer which fields she may use, stores extracts in an access-controlled project area and deletes them after the study. She pseudonymises device identifiers in all outputs. Even technical data can be sensitive: a list of devices that kept authenticating after retirement could guide an attacker to weak spots. Her report therefore presents aggregates and distributions, not device lists, and she shares the detailed list only with the operations owner.

42.3 Findings with limits

Devices retired after the change stopped authenticating within a median of two days, compared with eleven days before. The asset-record improvement happened at the same time, so Sofia cannot separate the two causes fully; she notes that devices whose asset records were already accurate before the change also improved, which weakly supports the certificate effect. She reports the finding with this limitation and recommends continued monitoring. The project teaches her that even modest engineering questions need careful design to produce trustworthy answers.

Worked example

Sofia studies authentication after retirement before and after short-lived certificates, uses logs under agreed limits and reports a fall from eleven to two days with a confound noted.

Synthetic case. Logs, devices and results are synthetic; Herley and van Oorschot (2017) is real.

Errors and limits

Before-after designs cannot exclude simultaneous changes. Small device counts make medians unstable.

Exercises

• EX124 — Outcome definition. What outcome did Sofia measure? Rubric: 3 = days between retirement and last successful authentication;; 2 = defined;; 1 = vague;; 0 = none.
• EX125 — Device list caution. Why does the report avoid listing individual devices? Rubric: 3 = such a list could guide attackers to weak spots;; 2 = attacker map;; 1 = no reason;; 0 = none.
• EX126 — Confound. What confound limits the conclusion? Rubric: 3 = asset-record improvements happened at the same time;; 2 = asset records;; 1 = vague;; 0 = none.

Synthesis

Applied research turns engineering changes into tested claims when questions, data handling and limits are explicit.

Continue studying — chapter 42 of 50

← Chapter 41 (draft) · Course chapter index · Chapter 43 (draft) →