Chapter 5 — Identity and access management: design alternatives
Lesson map
Lesson 05 published. Map, example, EX021.
flowchart TD
N0["Learning outcomes"]
N1["5.1 Three designs"]
N2["5.2 Criteria and weights"]
N3["Lab — Weighted scoring"]
N4["5.3 Reading the scores"]
N5["Synthesis"]
N0 --> N1 --> N2 --> N3 --> N4 --> N5
N5 --> EX["Worked example"]
EX --> Q["Exercise EX021"]
Logical drawing
flowchart LR
N0["Learning outcomes"]
N1["5.1 Three designs"]
N2["5.2 Criteria and weights"]
N3["Lab — Weighted scoring"]
N4["5.3 Reading the scores"]
N5["Synthesis"]
N0 --> N1 --> N2 --> N3 --> N4 --> N5 --> Q["EX021"]
[Learning outcomes]
|
v
[5.1 Three designs]
|
v
[5.2 Criteria and weights]
|
v
[Lab — Weighted scoring]
|
v
[5.3 Reading the scores]
|
v
[Synthesis]
|
v
[Worked example] --> [Exercise EX021]
Learning outcomes
You will (1) define realistic privileged access designs; (2) score them against agreed weighted criteria; (3) test the sensitivity of the ranking.
5.1 Three designs
Marta defines three designs. Design A keeps existing administrator accounts but adds a password vault for shared and emergency credentials and phishing-resistant keys. Design B replaces everyday-account administration with named administrator accounts, security keys and just-in-time elevation through the cloud directory's built-in features, with a vault only for emergency accounts. Design C buys a full privileged access management suite with session recording, credential rotation and approval workflows. Each design is described in enough detail to estimate cost and disruption. Costs are estimated over three years.
5.2 Criteria and weights
Criteria are agreed with Corran's IT manager and operations director before scoring: risk reduction weighted 35 percent, operability for a small team 25 percent, cost 20 percent and disruption 20 percent. Each design is scored from one to three on each criterion, with written evidence such as vendor documentation, a two-week trial of built-in elevation and staff estimates. Agreeing weights before scoring prevents the analysis being bent towards a favourite option after the numbers are known. The agreed weights are signed off in writing.
Lab — Weighted scoring
"""Weighted scoring of three privileged-access designs on agreed criteria (scores 1-3; stdlib only)."""
w = {"risk reduction": 0.35, "operability": 0.25, "cost": 0.20, "disruption": 0.20}
opts = {"A: vault + shared admin": {"risk reduction": 1, "operability": 3, "cost": 3, "disruption": 3},
"B: named admin + JIT": {"risk reduction": 3, "operability": 2, "cost": 2, "disruption": 2},
"C: full PAM suite": {"risk reduction": 3, "operability": 1, "cost": 1, "disruption": 1}}
for o, s in sorted(opts.items(), key=lambda kv: -sum(w[c]*kv[1][c] for c in w)):
print(f"{o:26} weighted={sum(w[c]*s[c] for c in w):.2f}")
Run with python3 lab_score.py. Verified output:
B: named admin + JIT weighted=2.35
A: vault + shared admin weighted=2.30
C: full PAM suite weighted=1.70
5.3 Reading the scores
Design B scores 2.35, Design A 2.30 and Design C 1.70. B leads, but only narrowly ahead of A, so Marta tests sensitivity. If risk reduction is weighted slightly less, A overtakes B; if risk reduction is weighted more, B's lead grows. Because Corran's insurer and board have stressed privileged access risk, the agreed weighting is defensible, but she reports the closeness openly. C ranks last because its operability and disruption scores are poor for a small team, despite strong risk reduction.
A lead of 0.05 is smaller than the swing from a modest change in weights, so the result is a narrow preference rather than a clear win.
Worked example
Weighted scores rank B (2.35) narrowly ahead of A (2.30) and well ahead of C (1.70); sensitivity analysis shows the B–A order depends on keeping risk reduction weighted highly.
Synthetic case. Designs, weights and scores are synthetic.
Errors and limits
Presenting a narrow scoring lead as decisive hides how sensitive the ranking is to the weights chosen.
Exercises
- EX021 — Design B content. What does Design B include? Rubric: 3 = named admin accounts, keys, JIT via built-in features, vault for emergencies;; 2 = JIT;; 1 = suite;; 0 = none.
- EX022 — Top score. Which design scores highest in the lab, and with what? Rubric: 3 = B, 2.35;; 2 = B;; 1 = C;; 0 = none.
- EX023 — Narrow margin. How large is B's lead over A? Rubric: 3 = 0.05;; 2 = narrow;; 1 = large;; 0 = none.
- EX024 — Sensitivity finding. What changes the B–A order? Rubric: 3 = weighting risk reduction lower;; 2 = weights;; 1 = vague;; 0 = none.
- EX025 — Suite weakness. Why does Design C rank last? Rubric: 3 = poor operability and disruption for small team;; 2 = operability;; 1 = security;; 0 = none.
Synthesis
Realistic alternatives, agreed weights and sensitivity analysis make design choices transparent.
Continue studying — chapter 05 of 12
← Chapter 04 (draft) · Course chapter index · Chapter 06 (draft) →