← Book home
Part 5 · Privacy, Slicing, Edge, and Special Networks
19

User Privacy in 5G

Hiding who you are — and what's left to leak when you've succeeded

"Encryption hides what you say. Privacy hides that it's you saying it, from here, right now. 5G made huge progress on the second — and it is not finished." — THE PRIVACY DISTINCTION

5G's headline privacy win is SUCI: your permanent identity no longer crosses the air in clear. But privacy is bigger than identity concealment. Temporary identifiers must rotate, paging must not leak presence, location services must be controlled, and metadata still tells a story. This chapter assembles the full privacy picture — what 5G fixed, what it didn't, and what an operator must still do.

🎯 Learning objectives
📘 Standards reference box — Chapter 19
SpecificationTitleRelease / version verified
TS 33.5015G security — privacy (SUCI, GUTI, identity handling)Rel-18, v18.11.0 (2026-04)
TS 23.273 / 38.3055G location services (LCS) architectureRel-18/19 edition
RegulatoryGDPR-style data-protection obligations (region-specific)current

Checked June 2026 — verify against the latest 3GPP version. Builds on Chapter 4 (identity).

19.1 The Privacy Threat Taxonomy

FIGURE 19.1Privacy Threat Taxonomy — Identity, Location, Behavior
privacy threats IDENTITY (who?) IMSI catching → SUCI (Ch 4) stale GUTI tracking AKA failure linkability (Ch 6) mostly fixed, with caveats LOCATION (where?) cell-level via signaling LCS abuse (Ch 12, here) paging-area inference needs active controls BEHAVIOR (what/when?) traffic-analysis metadata capability fingerprinting timing/volume side channels hardest to eliminate
Purpose: the privacy problem has three faces. 5G largely solved identity; location and behavior need operator action and remain partially open.
FIGURE 19.2Identity Exposure on the Air — 2G to 5G
2GIMSI clear 3GIMSI leaks 4GIMSI leaks 5G SASUCI permanent identity exposure on the radio (lower = better)
Purpose: the one privacy metric where 5G made a generational leap. SUCI drops permanent-identity exposure from "routine" to "never" — but only on SA (Chapter 18).
FIGURE 19.3SUCI Protection Scope — and Its Limits
SUCI DOES cover ✓ permanent identity over the air ✓ even from the serving network ✓ unlinkable per registration the IMSI-catcher era ends SUCI does NOT cover ✗ stale GUTI tracking (§19.3) ✗ location via signaling/LCS (§19.5) ✗ metadata / capability fingerprint (§19.7) SUCI is necessary, not sufficient
Purpose: calibrate expectations. SUCI is a major win, but a network can deploy it perfectly and still leak privacy through GUTIs, location, and metadata.

19.2 5G-GUTI Discipline and Paging Privacy

FIGURE 19.45G-GUTI Reallocation — When It Must Happen
reallocate the 5G-GUTI on: ✓ initial registration ✓ mobility / periodic update ✓ after paging-triggered service request THE RULE a GUTI that never changes is a de-facto permanent identifier → it resurrects exactly the tracking SUCI was meant to kill
Purpose: the discipline that makes GUTIs safe. SUCI hides the permanent ID, but a never-rotated temporary ID becomes the new permanent ID — reallocation is mandatory hygiene.
FIGURE 19.5Tracking via Stale GUTI — Attack and Defense
attack (stale GUTI) attacker observes GUTI in paging/signaling GUTI is the same day after day → recognizes the same subscriber anywhere → movement profile built over time SUCI didn't help — the GUTI gave it away defense ✓ reallocate GUTI per the trigger list ✓ short timers, not multi-day ✓ verify in traces — vendor defaults vary ✓ KPI on GUTI age distribution a fresh GUTI breaks the tracking link
Purpose: the concrete GUTI attack. It needs no cryptographic break — just an operator who configured a long reallocation timer "to save signaling."
FIGURE 19.6Paging Privacy — Temporary Identifiers
paging identifies the UE with TEMPORARY IDs (5G-S-TMSI / I-RNTI), never the permanent identity → an eavesdropper sees a rotating temporary value, not who you are the subtle leak if the temporary ID is stable AND paging area is small, presence can still be inferred → rotate temporary IDs + design paging areas to balance privacy vs efficiency
Purpose: paging is a privacy surface too. Temporary identifiers protect identity, but stability + small paging areas can still leak presence — another reason rotation matters.

19.3 Location Privacy

FIGURE 19.7Location Service (LCS) Privacy Architecture
LCS clientapp/emergency/lawful GMLCauthorize clientcheck subscriber privacysettings LMF / AMFcompute position UE the privacy control point is the GMLC: it must authorize the client AND honor the subscriber's privacy profile before any position is returned
Purpose: where location privacy is enforced. The GMLC gates location requests — authorization plus the subscriber's own privacy settings — so location isn't handed out freely (cf. NEF, Chapter 12).
FIGURE 19.8IMSI Catcher vs 5G SA — What Still Works
no longer works on 5G SA ✗ capture cleartext IMSI (SUCI conceals) ✗ trivially link registrations (fresh SUCI) ✗ fake network undetected (mutual auth) the classic stingray is defeated on SA may still work ⚠ force downgrade to NSA/4G/2G (Ch 18) ⚠ stale-GUTI tracking (if misconfigured) ⚠ metadata / capability fingerprinting privacy needs SA + config + monitoring
Purpose: an honest scorecard. 5G SA kills the classic IMSI catcher, but downgrade and metadata attacks survive — so privacy is a system property, not a single feature.

19.4 Remaining Leaks

FIGURE 19.9Capability and Measurement Leakage
capability fingerprinting the set of features a UE advertises can narrow the device model/firmware → partial re-identification despite SUCI mitigate: protect capabilities early, uniform sets measurement / radio leakage measurement reports describe the radio environment (neighbor cells) → can correlate to location/movement mitigate: protect under AS security
Purpose: privacy leaks below the identity layer. Even with SUCI, what a device is and what it measures can partially re-identify or locate it.
FIGURE 19.10Metadata and Traffic Analysis
encryption hides CONTENT, not PATTERN timing, volume, periodicity, and flow correlation can reveal behavior — app usage, sleep/wake, presence this is the hardest privacy problem: it survives perfect ciphering mitigate: minimize exposed metadata, padding/cover traffic where feasible, limit who sees flow records
Purpose: the frontier of privacy. Traffic analysis works on metadata that encryption can't hide — a reminder that confidentiality and privacy are different goals (Chapter 1).
FIGURE 19.11Privacy Regulation Mapping for Operators
regulatory duty 5G mechanism / operator action data minimizationNEF minimization, coarse location, pseudonyms (Ch 12) consent for location/data sharingGMLC privacy profile, NEF per-UE consent identity protectionSUCI + GUTI reallocation retention limits / auditlog minimization, access audit (Ch 26) breach notificationincident response (Ch 27), minimized data = smaller breach
Purpose: privacy is also a legal obligation. Each regulatory duty maps to a 5G mechanism or operator practice — the technical and the compliance views meet here.
FIGURE 19.12Operator Privacy Hardening Checklist
5G PRIVACY HARDENING — THE ESSENTIALS ☑ SUCI scheme A/B enforced (no null) + KPI ☑ 5G-GUTI reallocation verified in traces ☑ LCS/GMLC authorization + subscriber privacy profiles ☑ protect UE capabilities & measurements under security ☑ minimize exposed data at NEF (coarse, pseudonymized) ☑ monitor for forced downgrades (Ch 18) ☑ log/retention minimization + access audit ☑ map duties to regulation (GDPR-style)
Purpose: the privacy job on one card. Identity concealment is step one of eight — the rest are configuration, architecture, and governance.

19.5 The Practical Operator View

Common misconfiguration risks

19.6 Threats and Mitigations

ThreatVectorDefense
Identity captureIMSI catcherSUCI (Chapter 4)
Subscriber trackingstale GUTIfrequent reallocation
Location disclosureLCS abuseGMLC authorization + privacy profile
Re-identificationcapability fingerprintprotect/uniform capabilities
Behavioral profilingtraffic-analysis metadataminimize metadata, limit access
Protection strippingforced downgrademonitor + prioritize SA

19.7 Terminology, Example, Checklist

TermMeaning
SUCI / SUPIConcealed / permanent subscription identity (Chapter 4)
5G-GUTI / 5G-S-TMSI / I-RNTITemporary identifiers used in signaling and paging
LCS / GMLC / LMFLocation services / gateway (privacy gate) / location function
traffic analysisInferring behavior from metadata despite encryption

Real network example. A privacy NGO tested an operator's 5G SA network and reported it could still track specific subscribers. Investigation showed SUCI was correctly enforced — but the 5G-GUTI reallocation timer was set to 24 hours "to reduce signaling overhead." Because each subscriber kept the same GUTI all day, an observer monitoring paging and signaling could recognize and follow individuals across cells, exactly as in the IMSI-catcher era. The expensive identity-concealment feature was undone by one timer. Fix: reallocate the GUTI on every registration, periodic update, and paging-triggered service request, and add a KPI on GUTI age. Privacy is only as strong as its weakest identifier — and the GUTI is an identifier too.

Chapter Summary

? Review Questions

  1. Distinguish the three branches of the privacy threat taxonomy with an example of each.
  2. State precisely what SUCI does and does not protect.
  3. Why is GUTI reallocation a privacy requirement, and what attack does a stale GUTI enable?
  4. How does paging preserve privacy, and where can it still leak presence?
  5. Where is location privacy enforced, and what two checks must occur before a position is returned?
  6. Name two privacy leaks that survive perfect SUCI deployment.
  7. Why is traffic-analysis the hardest privacy problem?
  8. A network with SUCI enforced still allows subscriber tracking. Diagnose and fix.
🧪 Mini lab — track yourself (ethically)

In an Open5GS + UERANSIM lab (your own UE only): (1) Register a UE and capture the 5G-GUTI. (2) Re-register / trigger updates and check whether the GUTI changes — set a long reallocation timer and confirm it stays the same (the vulnerable state), then a short one and confirm it rotates. (3) Reflect: with a static GUTI, what could an observer of paging/signaling determine about your UE over a day? (4) List which of your captures would leak under encryption anyway (metadata: timing, volume) and which the GUTI rotation fixed. You've now demonstrated that identity concealment and identifier rotation are two separate, both-necessary controls.