Authentication security on a smart lock is not a single feature a supplier either has or does not have. Five pillars carry the whole structure: something you

게시일: 11월 10, 2025

Smart Lock Authentication Security

Smart Lock Authentication Security

Authentication security on a smart lock is not a single feature a supplier either has or does not have. It is the discipline of deciding what credential the lock accepts, how it captures that credential, where it makes the admit-or-deny decision, what evidence it keeps, and what fallback path is left when the chosen credential fails or is contested. Five pillars carry the whole structure: something you know, something you have, something you are, something about the context, and the way these pillars are composed. Every standalone lock in the SmartLockFactory catalog — across the 77 Smart Door Locks, 37 Fingerprint, 24 Hotel, 18 Keypad, 10 Outdoor, 9 RFID, 6 Face Recognition, 6 Cabinet, 3 Video Camera and 3 Glass models — sits somewhere on each of those five pillars. The buyer's job is to read each pillar honestly, not to chase the supplier's headline.

This article is written as a decision-pillar explainer. Each pillar gets the same treatment: what the pillar claims, what a standalone lock can honestly deliver, where the pillar breaks down under attack or under operational failure, and what the buyer must verify on the product page before the pillar is accepted at face value. The five pillars are written to be lifted off the page and used as the working language of any authentication conversation between buyer and supplier. Distinct from Smart Lock Security Architecture (Smart Lock Security Architecture), which sets the architecture frame inside which these five pillars operate; this article tells the buyer how to read each pillar honestly. Distinct from Smart Lock Credential Security (Smart Lock Credential Security), which dissects the credential as a manufactured object; this article treats the credential as one input to a decision pillar. And distinct from Smart Lock Fingerprint Security (Smart Lock Fingerprint Security), which dissects the biometric pillar alone; the fingerprint pillar is the third pillar below, written here as one pillar among five.

A first-pass read of "smart lock authentication" tends to produce three lazy beliefs — that more credential methods mean more security, that any biometric is stronger than any PIN, and that networked / cloud-connected authentication is automatically safer than standalone / local. The five-pillar explainer below refuses each of those beliefs on its own terms, then walks the buyer through a way of reading supplier material that does not depend on any of them.

Why authentication security is a five-pillar question and not a checklist

The authentication conversation in access control has, for forty years, orbited the same three factors — something you know, something you have, something you are. Those three are necessary, but in a standalone lock they are not sufficient. A lock that lets the right person in at the wrong time has not authenticated; it has merely matched a credential. A lock that lets the right credential in from the wrong device, in the wrong sequence, with no record, has not authenticated; it has merely opened. The fourth pillar — context — forces the question of when and under what surrounding conditions a credential is honoured. The fifth pillar — composition — forces the question of how the pillars combine when one pillar is contested, spoofed, copied, lost or stolen.

For a standalone lock the question is sharper than for a networked system. A standalone lock cannot phone home to revoke a stolen credential, cannot refresh a key on a schedule pushed from the cloud, cannot run a behavioural anomaly model on a server. Every authentication decision must be local, must be immediate, and must survive a power event, a mechanical event, and an honest human forgetting, losing or lending their credential. The five pillars below are written for that constraint set — the standalone lock with limited local memory, finite battery, a finite number of audit records (the SLK-9936S product-page-documented upper bound is up to 1,000 audit/event records; the equivalent upper bound for any other model in the catalog is verified in writing on the product page), and a small set of credential methods chosen at design time.

The buyer who walks into a supplier call with only "do you have fingerprint and PIN and card and app" on their mental checklist will get a confident "yes" on all four — and will leave without having asked a single question that reveals whether the lock will hold up under attack or under operational reality. The buyer who walks in with the five pillars as their working language will ask, of each pillar, the three questions the supplier cannot bluff: what does this pillar do, what does it not do, and what evidence on the product page proves which one applies.

Pillar 1 — Something you know (the knowledge factor)

Purpose: bind access to a secret the legitimate user remembers and the attacker does not.

The knowledge factor in a smart lock is the passcode — sometimes called PIN, user code, password, or numeric code. On the SLK-8831 5-in-1 catalog line, the passcode is one of the five documented methods. On the SLK-9936S hotel line, the passcode is generally reserved for staff or management use rather than the guest credential (the guest credential is overwhelmingly the RFID card per the product-page-documented spec block). The honest reading of the knowledge pillar in standalone locks has four sub-points the buyer should not let the supplier gloss:

  1. Length matters more than the supplier's "6-digit" headline. A 6-digit PIN has a million possible values; a 4-digit PIN has ten thousand. A standalone lock with anti-peep / anti-trailing / wrong-code lockout features raises the attacker's cost per attempt but does not raise the keyspace. The product page is where length, anti-peep behaviour, lockout threshold and lockout duration are committed in writing; the supplier's brochure is where they are promised in words. Confirm both.

  2. Storage matters more than storage format. A passcode stored in plaintext on the lock is a different security object from a passcode stored as a salted hash. The product page rarely specifies which; the supplier rarely volunteers. The honest read: assume the worst, ask for the storage form in writing, and treat any answer that includes the word "encrypted on the card" as a non-answer (encryption-on-card describes the credential carrier, not the storage form inside the lock).

  3. Sharing kills the pillar. A knowledge factor that two people share is not the same factor. The buyer should plan a per-user PIN policy — staff PIN per role, contractor PIN per engagement, guest PIN per stay — and confirm that the chosen model's user-PIN count ceiling is documented. The SLK-9936S-class hotel product line is documented as standalone-or-centralized-database modes; whether the per-user PIN count is unlimited in centralized mode or capped per lock in standalone mode is a per-model confirmation.

  4. Forgetting is the operational failure. A forgotten PIN is not a security event; it is an operations event. The pillar produces no security if the legitimate user is locked out and the door is opened with the mechanical key — which is to say, the next pillar handles the failure of this one. Composition is handled in Pillar 5.

Pillar 1 — honest-reading checklist

Question to the supplierWhat an honest answer looks likeWhat a bluff answer sounds likecheck against the model specification
What is the supported passcode length?A specific minimum and maximum, with anti-peep behaviour named"Industry standard length"Yes — exact min/max digits
How is the passcode stored inside the lock?A named storage form (plain / hashed / salted-hashed) with the chip family"Securely stored"Yes — storage form
Is there a wrong-code lockout?A specific threshold and a specific lockout duration"Anti-tamper lockout"Yes — threshold and duration
How many distinct user PINs can the lock hold?A specific ceiling, separate for admin vs user PINs"Many users"Yes — ceiling per PIN class
What happens when the admin PIN is lost?A documented master-reset procedure that requires physical access and a written code"Contact support"Yes — master-reset path

A "yes" or "yes, of course" answer to any of these without the specific commitment in the cell above is not yet an answer. The pillar is honestly delivered only when the product page carries the commitment.

Pillar 2 — Something you have (the possession factor)

Purpose: bind access to a physical or digital token the legitimate user carries.

The possession pillar is the broadest of the five in a standalone smart lock. The catalog lists RFID / IC card, mechanical key, and the app/Bluetooth credential (TTLock / Tuya / Doodle as the three named app-platform lanes) as the documented possession methods; remote control is on a subset of models and must be confirmed per SKU. The honest read of the pillar has five sub-points:

  1. The mechanical key is the original possession factor. Every honest smart lock — including the SLK-9936S, which carries a 5-pin cylinder and a mechanical emergency access path per the product-page-documented block — keeps a mechanical key as a fallback. The mechanical key is also the lock's most attacked surface: the cylinder grade, the pick resistance, the key control policy (who can copy the key, where blanks come from) are the operational security of the pillar. The buyer should treat the mechanical key as a governed possession factor, not as a marketing afterthought.

  2. The RFID card is a small computer. An RFID credential at 13.56 MHz (the SLK-9936S product-page-documented band) is not a magnetic strip with a serial number; it is a contactless smart card with memory, optional cryptography, and a card-side decision about what data to release. The honest question is not "is it 13.56 MHz" — almost every modern hotel card is — but what data lives on the card, what is read, what is written, and whether the card itself can be cloned with off-the-shelf hardware. The supplier's brochure will say "secure RFID"; the product page is where the card format and the supported Mifare class are committed. Confirm both.

  3. The mobile credential is not the same credential as the card. A TTLock / Tuya / Doodle mobile credential is delivered over Bluetooth, signed by an app on the phone, presented to the lock as a transient packet. Its honesty profile is different from the card: it is harder to clone (the secret lives in the phone's secure element, not on a contactless card) but it depends on the phone's battery, the phone's BLE stack, and the user's app being logged in. The buyer should treat the mobile credential as a parallel pillar of the possession factor, with its own failure modes.

  4. Loss is the operational failure of every possession factor. A lost card, a lost phone, a lost key — every possession credential can be lost, and every possession factor must have a documented revocation path. For a standalone lock the revocation path is local: the next legitimate credential issued must invalidate the lost one. Whether the chosen model supports overwrite-on-next-tap, master-card-driven revocation, or both, is a per-model confirmation. The pillar produces no security if a lost credential remains valid until its natural expiry.

  5. Sharing kills the pillar the same way it killed Pillar 1. A borrowed card is not the bearer's credential. A shared phone unlock pattern is not a per-user credential. The honest read plans for per-user possession tokens and a revocation path per token, not for a single "house card."

Pillar 2 — honest-reading checklist

Question to the supplierWhat an honest answer looks likeWhat a bluff answer sounds likecheck against the model specification
What is the supported card format and Mifare class?Named format (e.g. Mifare Classic 1K, Mifare DESFire) with read range and read time"13.56 MHz RFID"Yes — exact format and read range
How is a lost card revoked on a standalone lock?A named path (overwrite-on-next-tap, master-card revocation, lock-reset) with the operational steps"Re-issue cards"Yes — revocation procedure
Is the mobile credential offered on this exact SKU?A named app-platform lane (TTLock / Tuya / Doodle) and the BLE generation"App control"Yes — exact app platform
What is the mechanical-key cylinder grade?A specific cylinder grade and a documented key-control policy"High-security cylinder"Yes — cylinder grade + key-control
How many distinct credentials can the lock hold?A specific ceiling, separate for cards vs PINs vs fingerprints"Large user capacity"Yes — ceiling per credential type

The pillar is honestly delivered when the product page carries the commitment to format, range, revocation and ceiling, not when the supplier asserts the credential is "secure."

Pillar 3 — Something you are (the inherence / biometric factor)

Purpose: bind access to a biological or behavioural characteristic of the legitimate user.

The biometric pillar in the SmartLockFactory catalog is documented on two credential classes: fingerprint (37 models in the Fingerprint category) and face recognition (6 models in the Face Recognition category, including a 3D-face-recognition line and the Future 1 / Q7 / 9L5-R references in the catalog notes). The 3D-face-recognition line and the small number of fingerprint-and-face combos are the only biometric methods documented in the catalog. Palm-vein, palm-print, finger-vein, iris, retina, voice-print, gait-recognition, and Apple Home Key / biometric-via-phone combinations are NOT in the catalog — none of those should be claimed or quoted for any SKU. The honest read of the biometric pillar has six sub-points the buyer should not let the supplier gloss:

  1. The biometric is a comparator, not a secret. A fingerprint template stored on the lock is a math object derived from the fingerprint, not the fingerprint itself; a face template is a vector of features, not a photo. The honest read asks: where is the template stored, who can read it, and what does the comparator accept as a match? A lock that stores templates in plaintext and accepts a match at a generous threshold is not the same security object as a lock that stores templates in a secure element and accepts a match at a tight threshold. The product page is where storage and threshold are committed. Confirm both.

  2. The False Acceptance Rate and False Rejection Rate are operational, not theoretical. FAR and FRR are the two numbers the supplier will quote if pressed. The honest read is that FAR is the number that matters for security (the chance an impostor's biometric is accepted) and FRR is the number that matters for operations (the chance the legitimate user is rejected). Both must be on the product page; both must be tested on the bench in the buyer's environment. A supplier that quotes FAR without FRR, or vice versa, has not finished the conversation.

  3. The biometric is not revocable. A leaked password can be changed. A leaked card can be reissued. A leaked fingerprint cannot. The pillar therefore demands template-storage discipline at a level the other pillars do not. The buyer should ask, in writing: what is the template storage form, can a stolen lock be opened with a stolen template, and is there a documented template-wipe procedure on a security event?

  4. Environmental failure is the operational weakness. A fingerprint reader fails on wet fingers, on worn fingers, on the elderly, on children, on manual-labour users. A face reader fails in backlight, in darkness without IR, with masks, with cosmetic change, with aging. The honest read plans for two or more non-biometric fallback methods (Pillar 5 below) and confirms that the biometric reader's environmental envelope — temperature, humidity, light, mask policy — is documented on the product page, not in the brochure.

  5. The biometric is the easiest pillar to fake on a brochure and the hardest to fake on a bench. Marketing renders show a clean thumb on a reader and a one-tap open; reality renders are fingerprint rejections on a contractor at 06:30 in January. The honest buyer demands a bench demo on the actual user population — not on the supplier's hand — before accepting the pillar. Smart Lock Fingerprint Security (Smart Lock Fingerprint Security) is the deep dive on this pillar; it is the anchor reference for any fingerprint-specific RFQ question.

  6. Multimodal biometric is not the same as multi-factor. A lock with both fingerprint and face on the same user is not two-factor; it is two instances of the same factor (inherence). A lock with fingerprint + PIN is two-factor. The honest read distinguishes the two — the second is meaningfully stronger than the first.

Pillar 3 — honest-reading checklist

Question to the supplierWhat an honest answer looks likeWhat a bluff answer sounds likecheck against the model specification
What is the FAR and FRR for the biometric reader?A specific pair of percentages, measured on a named test population"Industry-leading accuracy"Yes — FAR + FRR
Where is the biometric template stored?A named storage form (secure element / encrypted flash / hashed) with the chip family"Stored securely"Yes — storage form
How many templates can the lock hold?A specific ceiling, separate for admin vs user"Many users"Yes — ceiling
What is the environmental envelope?Temperature, humidity, light, and mask policy stated as ranges"Works in all conditions"Yes — environmental ranges
What is the documented template-wipe procedure on a security event?A named path (master-card, admin PIN, physical reset) with the steps"Contact support"Yes — wipe procedure
Is the biometric multimodal or multi-factor?Explicit: same factor (two inherence) vs two factors (inherence + knowledge / inherence + possession)"Multiple biometrics"Yes — composition (see Pillar 5)

The pillar is honestly delivered when the product page carries the FAR/FRR pair, the storage form, the ceiling, and the environmental envelope — and when the bench demo replicates the real user population, not the supplier's hand.

Pillar 4 — Something about the context (the context factor)

Purpose: bind access to the time, location, sequence, or surrounding state under which the credential is presented.

The context pillar is the one the three-factor textbook rarely covers, and it is the one the standalone smart lock most often has without ever naming. A standalone lock that admits a guest card between 14:00 on the check-in day and 12:00 on the check-out day, refuses the same card at other times, and tracks the sequence in which cards are presented — that lock has a context pillar, even if no marketing page has ever said so. The honest read of the pillar has five sub-points:

  1. Time windows are the most common context attribute. A guest PIN valid for the stay window, a staff PIN valid only for the shift window, a contractor PIN valid only for the engagement window — each is a context factor layered onto a knowledge or possession factor. The honest question is where the window is enforced (lock-side, app-side, hotel PC software-side) and what the clock's accuracy dependency is (does the lock keep time on a battery-backed RTC, and what happens to the window if the RTC drifts during a long power event).

  2. Sequence is a context attribute that earns its keep. A standalone hotel lock that requires the guest card to be presented at the front-desk encoder before it will open the guest door has a sequence attribute — the lock will not honour a card that has not been written to by the desk. A residential lock that requires an "unlock from inside" turn before the PIN is changed has a sequence attribute. The honest question is whether sequence is enforced and what the failure mode is if the sequence is broken.

  3. Location is harder to do honestly on a standalone lock. A networked lock can read GPS from a phone and refuse an unlock outside the property. A standalone lock generally cannot — there is no continuous location signal at the door. The honest read treats location as a possession-context attribute (the lock reads the credential, the credential carries a property-id binding, the lock refuses mismatched bindings) rather than as a true location pillar. For hotel locks the SLK-9936S-class line is documented as standalone-or-centralized-database modes; the room-binding is enforced locally in standalone mode and through the database in centralized mode.

  4. Occupancy and dual-presentation are rare but real. A lock that requires a manager card plus a guest card for a late check-out, a lock that requires two staff cards simultaneously for a vault door, a lock that refuses an unlock when the room is occupied (interior handle set) — each is a context attribute that goes beyond the credential. The honest read asks for the documented behaviour of the lock under each of these conditions, on the product page, in writing.

  5. Context is silent in the brochure. The honest buyer will not find "context factor" on a supplier's marketing page; they will find it in the lock's per-model operating manual, in the hotel PC software's per-door policy, and in the staff-card hierarchy document. The pillar therefore lives in artefacts, not in promises.

Pillar 4 — honest-reading checklist

Question to the supplierWhat an honest answer looks likeWhat a bluff answer sounds likecheck against the model specification
What time-window policies does the lock enforce locally?A named set (shift window, stay window, contractor window) with the enforcement plane named"Time-based access"Yes — window types + plane
Where does the lock get its clock?A named RTC source, battery-backed, with drift tolerance"Built-in clock"Yes — RTC source + battery
What sequence is required before a credential is honoured?A named sequence (e.g. desk-encode-then-present) with the failure mode"Smart workflow"Yes — sequence + failure mode
Is room / property binding enforced?A named binding mechanism (room-id-on-card, lock-id-in-database)"Property-specific"Yes — binding mechanism
What dual-presentation / occupancy policies are supported?A named list with the lock behaviour under each"Advanced features"Yes — supported policies

The pillar is honestly delivered when the product page or the operating manual names the window types, the clock source, the sequence and the binding mechanism — and when the buyer's bench rehearsal exercises the failure mode of each.

Pillar 5 — Composition (the way the pillars combine or fall back)

Purpose: name how the four pillars above are combined for routine use, what falls back when one pillar is contested, and what is left when everything else fails.

Composition is the pillar that converts four credential types into one authentication decision. The honest read has six sub-points:

  1. Combination is not stacking. A lock with five credential methods is not five-factor; it is a one-factor lock with five input options. A lock that requires PIN + card on the same door is two-factor (knowledge + possession). A lock that requires card + fingerprint on the same door is two-factor (possession + inherence). A lock that requires fingerprint + PIN is two-factor (inherence + knowledge). The composition pillar demands explicit per-door policy: which doors are single-factor, which are two-factor, which require an admin override to switch.

  2. Fallback is the operational consequence of composition. A biometric reader that fails on a wet finger must fall back to a non-biometric method. The honest question is what the fallback chain is — PIN, card, mechanical key, app, master credential — and whether the fallback is logged. A fallback that does not write to the audit trail is an unaudited access. For the SLK-9936S, the product-page-documented spec block lists mechanical emergency access as a documented fallback; for siblings in the hotel line the equivalent fallback path is a per-model confirmation.

  3. Mechanical override is the last-resort possession factor. Every honest standalone lock — including the SLK-9936S with its documented mechanical emergency access — keeps a mechanical key. The mechanical key is the override path for the entire composition. The honest read treats the cylinder grade and the key-control policy as part of the composition pillar, not as a separate issue.

  4. Audit-trail depth bounds the composition. A standalone lock can only audit what it has memory to store. The SLK-9936S product-page-documented upper bound is up to 1,000 audit/event records; the equivalent for any other model in the catalog is a per-model confirmation. A lock that hits the audit-rollover boundary must have a documented rollover policy — overwrite-oldest, stop-recording, or fail-secure. The composition pillar is not honestly delivered if the audit trail silently drops events on a busy door.

  5. Decision plane must be local for the lock to be standalone. The honest composition has its decision made at the door, on credentials and policy the lock holds locally. A standalone lock that requires a phone ping to admit a credential is not a standalone lock; it is a networked lock with the network hidden. The buyer should ask, in writing: does the door open without network, without phone, without cloud? The honest composition answers yes to all three; a networked-dependent composition answers no to at least one, and the answer must be on the product page.

  6. Revocation is the closing of the composition. A composition that admits a credential but cannot revoke it is a composition that bleeds access over time. The honest read asks: what is the revocation latency — how long after a credential is marked invalid does the lock honour it — and what is the revocation plane — local overwrite, master-card, central database. A standalone lock's revocation latency is the time between the credential being marked invalid locally and the lock's next read of that invalidation; for SLK-9936S-class standalone mode the latency is bounded by the next card-present event.

Walk one presentation through the composition pillar and the standalone lock behaves like a small state machine — local decision, throttle, fallback chain, and the mechanical override that closes the loop:

stateDiagram-v2
  [*] --> Presented
  Presented --> LocalDecision: credential reaches the door
  LocalDecision --> Granted: admitted on local policy and audit row written
  LocalDecision --> Denied: expired window · unenrolled credential · mismatched binding
  Denied --> Lockout: wrong-attempt threshold crossed
  Lockout --> Fallback: lockout clears or staff path is invoked
  Fallback --> LocalDecision: next method in the chain · PIN · card · app
  Fallback --> MechanicalOverride: every electronic path unavailable
  MechanicalOverride --> [*]: the key is the whole product that day
  Granted --> AuditTrail: open event recorded within ceiling and rollover rule
  AuditTrail --> [*]

Pillar 5 — honest-reading checklist

Question to the supplierWhat an honest answer looks likeWhat a bluff answer sounds likecheck against the model specification
Which doors require two-factor authentication, and which are single-factor?A per-door policy, in writing"Multi-factor capable"Yes — per-door policy
What is the documented fallback chain per door?A named order (e.g. fingerprint → PIN → card → mechanical key) with audit logging"Multiple ways to open"Yes — fallback chain
What is the audit-trail depth and rollover policy?A specific ceiling (record count) and a named rollover rule"Large audit log"Yes — ceiling + rollover
Does the door open without network, without phone, without cloud?Explicit yes/no per dependency, per door"Works anywhere"Yes — dependency matrix
What is the revocation latency and plane?A specific latency bound and a named plane (local / master / central)"Instant revocation"Yes — latency + plane

The composition is honestly delivered when the product page carries the per-door policy, the fallback chain, the audit ceiling and rollover, the dependency matrix, and the revocation latency — and when the bench rehearsal exercises the failure of one pillar and confirms the others hold.

A pillar-by-pillar honest-reading scorecard

The buyer who walks into the supplier call with the table below as the worksheet can finish the conversation in one meeting. Each row is one pillar; each column is the honesty test the buyer must apply; each cell carries one of three markings — Documented (the product page commits in writing), Confirm (the supplier has answered but the product page has not yet been read), or Open (no answer yet). A pillar that ends the meeting with too many Open cells is not a pillar the buyer can accept.

PillarHonest claimTest the buyer appliesMarking
Pillar 1 — Knowledge"PIN/passcode supported"Length range, storage form, lockout threshold and duration, ceiling per PIN class, master-reset pathConfirm / Documented / Open
Pillar 2 — Possession"Card / mobile / mechanical key supported"Card format and read range, mobile app platform, mechanical cylinder grade, ceiling per credential type, revocation path per lossConfirm / Documented / Open
Pillar 3 — Inherence"Fingerprint / face supported"FAR + FRR pair, template storage form, ceiling, environmental envelope, template-wipe procedure, multimodal-vs-multi-factor distinctionConfirm / Documented / Open
Pillar 4 — Context"Time / sequence / binding supported"Window types and enforcement plane, RTC source, sequence requirement, room/property binding mechanism, dual-presentation policyConfirm / Documented / Open
Pillar 5 — Composition"Multi-method / multi-factor / fallback supported"Per-door factor policy, fallback chain with audit logging, audit ceiling and rollover, dependency matrix, revocation latency and planeConfirm / Documented / Open

The scorecard is the working artefact the buyer carries to every supplier call and to every bench rehearsal. A pillar-by-pillar honest-reading scorecard with no Open cells on the Open row is the precondition for accepting the lock's authentication security as deliverable, not promised.

What authentication security does not require (the honest out-of-catalog list)

A standalone smart lock's authentication security is bounded by what the lock can do locally, on credentials it holds, with the memory and power it has. Some authentication modalities that appear in adjacent markets are explicitly NOT in the SmartLockFactory catalog and the buyer should refuse any quotation that asserts them for any SKU:

  • Palm-vein, finger-vein, palm-print, vein-recognition — not in catalog. No quotation should assert them for any SKU. If a supplier's brochure contains the word "vein" against a model, the brochure is wrong; ask for the SKU number and check the model’s published specification.
  • Iris, retina, sclera — not in catalog.
  • Voice-print, voice-recognition at the door — not in catalog.
  • Gait-recognition, behavioural-biometric continuous authentication — not in catalog.
  • Apple Home Key, Matter, Zigbee, Z-Wave, Thread, UWB — not in catalog. None of these credential carriers should be quoted for any SKU.
  • NFC credit-card-form-factor credentials beyond the documented RFID / IC card line — not in catalog beyond what is documented.

The honest out-of-catalog list is not a refusal of those modalities in principle. It is a refusal to accept a brochure that names them for a model that does not carry them. The buyer's standing instruction to the supplier: every credential modality in any quotation must be named against an SKU and verified on the product page. No name on the product page means no modality in the order.

Consider a 60-unit residential block…

A 60-unit residential block — one main entrance, two stair entrances, one amenity door (gym / co-work), one bike room, one plant room — replaces a mechanical master-key system with a standalone smart lock system. The buyer adopts the five-pillar explainer as the working language of the procurement conversation.

Pillar 1 (knowledge) in practice. Each unit has a 6-to-8-digit resident PIN, distinct from a contractor PIN and a management PIN. The buyer confirms on the product page the lock supports the 6-to-8 range, has a wrong-code lockout at a specific threshold and a specific lockout duration, and stores PINs as hashed values rather than plaintext. The pillar is documented.

Pillar 2 (possession) in practice. Each unit has a resident RFID card and a mobile credential through one of the three documented app-platform lanes (TTLock / Tuya / Doodle, named per the chosen model). The mechanical key remains as a fallback. The buyer confirms the card format, the read range, the mobile-credential app platform, the cylinder grade and the revocation path for a lost card. The pillar is documented.

Pillar 3 (inherence) in practice. The main entrance and the stair entrances carry a fingerprint reader. The buyer confirms on the product page the FAR/FRR pair, the template storage form, the ceiling per lock, the environmental envelope (winter cold, summer humidity, mask policy is not applicable at a residential entrance) and the template-wipe procedure. The pillar is documented.

Pillar 4 (context) in practice. Contractor PINs carry a date window; gym access carries a 06:00–23:00 window; the plant room carries an always-on policy restricted to a single management credential. The buyer confirms the window types, the RTC source, the sequence requirement (none at a residential entrance; the desk is the resident, not a front desk) and the binding mechanism. The pillar is documented.

Pillar 5 (composition) in practice. Unit doors are single-factor (PIN + card; either is sufficient for the resident). Main and stair entrances are two-factor (card + fingerprint). The plant room is single-factor with a master-credential override. The fallback chain is mechanical key for unit doors, mechanical key plus management card for main entrances. The audit ceiling and rollover policy are documented on the product page. The dependency matrix is "yes" on standalone operation without network, without phone, without cloud. The revocation plane is local overwrite-on-next-tap. The composition is documented.

The outcome. The five-pillar honest-reading scorecard leaves the meeting with no Open cells. The buyer accepts the lock's authentication security as deliverable. The composition is honest because every pillar is honest. The supplier's brochure is no longer the source of truth; the product page, the bench rehearsal and the per-door policy are.

What the supplier's brochure hides (and how to read past it)

The honest five-pillar read is not the brochure's read. Three patterns recur on authentication security marketing material and the buyer should know them in advance:

  1. "Multi-factor" as a label, not as a configuration. A brochure that lists fingerprint, PIN, card and app as four methods does not describe a multi-factor lock; it describes a single-factor lock with four input options. The honest read asks which combinations are enforced per door (Pillar 5 above).

  2. "Encrypted" as a label, not as a description. A brochure that says "end-to-end encryption" without naming the cipher, the key length, the key-rotation policy and the storage form has not yet made a security claim. The honest read asks for the cipher and the storage form, in writing, on the product page.

  3. "AI" or "smart" as a label, not as a description. A brochure that says "AI-powered biometric" without naming the model, the training data, the comparator threshold and the template-wipe procedure has not yet made a security claim. The honest read asks for the FAR/FRR pair and the storage form, in writing, on the product page.

The buyer who reads past these three patterns, and who walks into the supplier call with the five-pillar honest-reading scorecard above, leaves the conversation with a deliverable authentication security — not a promised one.

How the five pillars connect to the rest of the security cluster

The five pillars above are the authentication layer of the lock's overall security posture. Three sibling articles in the C09 Security/Testing cluster carry the other layers:

  • Smart Lock Security Architecture — Smart Lock Security Architecture sets the architecture frame inside which the five pillars operate. The buyer should read Smart Lock Security Architecture first to place the pillars in the architecture (door hardware / credential / reader / decision plane / audit / fallback) before reading this article.
  • Smart Lock Credential Security — Smart Lock Credential Security dissects the credential as a manufactured object — the card, the mobile credential, the PIN, the biometric template — at the level of physical and logical security. The buyer should read Smart Lock Credential Security alongside Pillar 2 (possession) and Pillar 3 (inherence) of this article.
  • Smart Lock Fingerprint Security — Smart Lock Fingerprint Security is the deep dive on the biometric pillar. The fingerprint pillar in this article is the third pillar; that deep dive is the reference the buyer cites in any fingerprint-specific RFQ question, including FAR/FRR, template storage, and the bench demo discipline.

Together, Smart Lock Security Architecture places the pillars, this article explains the pillars honestly, Smart Lock Credential Security dissects the credentials, and Smart Lock Fingerprint Security deep-dives the biometric pillar. The four are a reading set; the buyer who treats them as a set leaves the supplier call with a deliverable authentication security.

A short buyer checklist, fit for the RFQ packet

The five pillars above reduce to ten lines the buyer pastes into the RFQ packet. The supplier that answers each line in writing, on the product page, has met the honest-reading test. The supplier that answers in adjectives has not.

  1. Pillar 1: name the supported passcode length range, the storage form, the wrong-code lockout threshold and duration, the ceiling per PIN class, and the master-reset path.
  2. Pillar 2: name the supported card format and Mifare class with read range, the mobile-credential app platform, the mechanical-key cylinder grade, the ceiling per credential type, and the revocation path for a lost credential.
  3. Pillar 3: name the FAR and FRR pair for each biometric reader on the chosen SKU, the template storage form, the ceiling per lock, the environmental envelope, the template-wipe procedure, and the multimodal-vs-multi-factor composition per door.
  4. Pillar 4: name the time-window types and the enforcement plane, the RTC source and drift tolerance, the sequence requirement per door, the room / property binding mechanism, and any dual-presentation policy.
  5. Pillar 5: name the per-door factor policy, the documented fallback chain per door with audit logging, the audit ceiling and rollover rule, the dependency matrix (network / phone / cloud), and the revocation latency and plane.

The RFQ packet that carries these ten lines, and the supplier's written answers to each, is the artefact the buyer's organization carries to the order meeting. A packet with adjectives in place of answers is the artefact the buyer walks away from.

Closing — the five pillars are a working language, not a checklist

The five pillars of authentication security — knowledge, possession, inherence, context, composition — are the working language the buyer and supplier use together to decide whether a lock's authentication is deliverable or merely promised. The pillars do not replace the product page; they tell the buyer what to demand the product page commits to. They do not replace the bench rehearsal; they tell the buyer what to exercise on the bench. They do not replace the supplier; they tell the buyer when to ask the supplier for evidence and when to accept the supplier's word.

The standalone lock that delivers all five pillars honestly — knowledge with a committed length and lockout, possession with a committed card format and revocation, inherence with a committed FAR/FRR pair and storage form, context with a committed window and clock, composition with a committed per-door policy and audit ceiling — is a lock whose authentication security the buyer can sign for. The lock that delivers three pillars honestly and is silent on two is a lock whose authentication security the buyer cannot sign for, no matter what the brochure claims.

For the supplier conversation that turns this article into an order, the next step is to bring the five-pillar honest-reading scorecard and the ten-line RFQ packet to the chosen SKU's product page — and to /order-samples for the bench rehearsal that proves the product page. For the conversation that turns the order into a contract, the next step is /request-a-quote, with the RFQ packet attached and the honest out-of-catalog list as the boundary the supplier must respect. The supplier that answers every line in writing is the supplier the buyer signs with. The supplier that answers in adjectives is the supplier the buyer walks away from — five pillars, no bluff, no brochure.

장바구니
쿠키를 허용하면 이 사이트에서의 이용 경험이 개선됩니다. 쿠키 정책
Chat on WhatsApp