Most RFID hotel lock buying guides start with the handle, the finish, and the price per door. Start with the evidence, because the market is noisy at both ends.

Publicerad den: mar 06, 2026

How to Choose an RFID Hotel Lock

How to Choose an RFID Hotel Lock

Most RFID hotel lock buying guides start with the handle, the finish, and the price per door. That is the wrong order. Every hotel lock in this category is a card reader with a bolt attached, and the security of the whole installation is decided by a credential question most buyers never ask: what lives on the card, who can copy it, and how fast you can invalidate it when it walks out the front door.

Here is the short answer before the detail: choose an RFID hotel lock by working security-first — settle the card technology and the key-custody model of the encoding system, then verify the reader behavior and audit records that make those decisions real, and only then compare designs. This article is a technical resource for doing that: what the public cloning reports actually establish, what the datasheet line "RFID Mifare 1" does and does not commit a supplier to, the Mifare Classic-versus-CPU card trade-off stated honestly, and the question set that separates a credential-secure lock from a decorative one.

If you need the broader property-level decision — doors, software ecosystem, service model — that is covered in our companion piece on how to choose hotel smart locks. This one goes vertical on the RFID credential itself.

What the Public Cloning Reports Actually Establish

Start with the evidence, because the market is noisy at both ends. One camp says RFID hotel cards are uncrackable because "the chip is encrypted." The other sells anti-cloning upgrades by implying every hotel door can be copied by anyone with a USB reader. Neither is useful for a purchasing decision. You need to know precisely what has been publicly demonstrated, and what it changes.

The publicly documented facts, stated as industry reporting rather than as anything we tested in-house:

  • The cryptographic algorithm used by the Mifare Classic family of 13.56 MHz contactless cards was analyzed and its key-recovery weaknesses were described in open security research published more than fifteen years ago. This research is public record and was widely covered by technology and security media.
  • Since then, consumer hardware communities and product testers have repeatedly demonstrated, on camera and in write-ups, that commercially available reader-writer devices can duplicate the contents of susceptible contactless cards in minutes. Hotel access cards appear regularly in these demonstrations because they are common and often provisioned with poorly protected keys.
  • The exposed variable is not the brand of card. It is the key system: a Classic-family card running a hotel's own diversified keys behaves very differently from the same card type shipped with default factory keys reused across an installation.

The practical reading for a buyer: a cloned key card is not a theoretical threat at hotel doors — it is a documented, repeatedly demonstrated one, with a known root cause. The root cause is almost never the lock. It is the credential configuration. Which means the selection question shifts from "is RFID secure?" to "does this lock-and-encoder system let me run the credential securely, and can the supplier prove which model they mean when they say secure?"

Reading the Claims Without Hype

Claim you will hearPublic basisWhat it actually means for your purchase
"Mifare cards are clone-proof"Overstated. Open research documented Classic-family weaknesses; consumer-tool duplication demos are widely reportedAsk which card family and which key system, not which brand name. "Mifare" alone answers nothing
"Your hotel can be cloned with a €30 reader"Real demonstrations exist, but they presuppose weak or default key provisioning and physical possession of a sample cardA supplier who dismisses cloning and a supplier who monetizes panic both fail the first question. Look for the one who talks about key custody
"CPU cards cannot be cloned"Fair as an architecture statement: CPU card keys never leave the chip and each transaction authenticates mutually, per public smart-card industry documentationTrue, but only if the lock's reader platform supports CPU cards — verify this per model, in writing, before the card fleet decision

The Credential Tiers: ID Tags, Memory Cards, Classic Family, CPU Cards

"RFID hotel lock" collapses at least four genuinely different technologies into one phrase. A selection made without knowing which tier is being quoted is not a selection at all. Tiers below are ordered by what the reader can actually verify about the card.

Legacy ID-only tags. Low-frequency (125 kHz) cards that broadcast a serial number. The lock trusts any card presenting a known number. Public tester communities have demonstrated duplication of these for decades; consumer tag copiers exist for exactly this reason. Some older hotel estates still run on them. Treat them as a system to be migrated off, not a budget option.

Plain IC memory cards. High-frequency cards that store data in sectors, with no challenge-response between card and reader. The reader reads a block, the lock compares it to expectation. If someone can read a card once, someone can write that content to a blank. Many properties that believe they run "encrypted keycards" in fact run unauthenticated sectors with a vendor's default key.

Mifare Classic-family cards (the "Mifare 1" of datasheet lines). 13.56 MHz cards with sector-level keys and a stream cipher whose weaknesses are the substance of the open research described above. The critical point for buyers: this tier is only as strong as its provisioning. With hotel-owned, diversified sector keys, disciplined rotation, and reader-side controls, Classic-family fleets still defend hotel doors adequately for many low-to-mid-risk properties — this is why they remain common worldwide. With default keys, they are an open gate with a logo on it.

CPU cards. Contactless smart cards with an onboard operating system. Keys are generated inside the chip and never leave it; every transaction is a mutual authentication — the reader proves itself to the card before the card releases anything. Copying all card memory yields a duplicate of data that a legitimate reader will not accept, because it lacks the keys. This is the architecture the public cloning reports effectively cannot touch with desk tools, and it is what payment cards and national ID programs use. The trade-off: CPU cards cost more per piece, encoding systems and readers must support them, and not every hotel lock platform that reads Classic-family cards will read CPU cards.

TierWhat sits on the cardAuthentication modelExposure per public reportingRealistic hotel role
Legacy 125 kHz ID tagFixed serial numberNone — reader trusts the numberTrivially duplicated with consumer copiersMigration source, not a choice
Plain IC memory cardData sectors, possibly no key appliedOne-way readDuplicated once readOnly acceptable with per-property keys enforced
Mifare Classic familyKeyed sectors, stream-cipher authenticationMutual but cipher-limitedKey-recovery weaknesses publicly documented; depends entirely on key custodyWorkable for low-to-mid risk with diversified keys + rotation + audit
CPU cardApplications inside a chip OS; keys never exportedFull mutual authentication per transactionNo public desk-tool cloning of a properly provisioned CPU cardStandard for high-turnover, urban, and brand-standard properties

What "RFID Mifare 1" on the Datasheet Commits the Supplier To

Take a real hotel-lock spec block, the SLK-9936S published by this factory, and read only the credential-relevant lines. This is first-party product data, quoted as published; you can pull the same block from the hotel lock category and hold any supplier to the same exercise.

What a line like "RFID Mifare 1 @ 13.56 MHz" genuinely guarantees — and where the specification stops talking — determines what you must demand in writing.

Spec line (as published)What it commits the lock toWhat it deliberately leaves openThe written confirmation to demand
RFID Mifare 1 @ 13.56 MHzReader operates on the 13.56 MHz HF air interface and the Classic family it namesWhether the same reader platform can be configured or supplied to authenticate CPU cardsA per-model statement of supported card families, including CPU-card support yes/no
Read distance 0–5 cmIntentional short-range coupling: the card must be presented at the faceplateWhether range is fixed or field-configurable, and firmware tolerance to relay-style tricksDefault read range at delivery and whether range/parameter settings are changeable by the hotel
Read time <0.25 sTransaction speed at the door; guest experience, not securityNothing about authentication strength — a fast yes can be an unauthenticated yesThat full authentication completes within the quoted read time
Audit/event records up to 1,000Each lock buffers up to 1,000 events standalone, before centralized captureWhat each event contains (card identity? method? user class?) and how logs are pulledA sample exported event record, plus the record-depth rollover behavior
Mechanical emergency accessA physical fallback path when electronics failKey-blank control and whether the cylinder is bitted per property or shipped on a common keyThat emergency cylinders are keyed per project and blanks are not commercially available
Standalone or centralized database modesEncoding architecture can start offline and growWhich card functions (rotation, fleet invalidation) require the centralized modeA feature matrix of what standalone mode cannot do

A spec block is a promise with a defined perimeter. Read the perimeter, then make the supplier extend it in writing. If a supplier reacts badly to that request, that reaction is itself data.

Five Properties to Verify, in Order

Security-first selection is testable. These five properties, verified in this sequence, decide whether an RFID hotel lock system deserves your room count. Anything a supplier cannot answer concretely on all five should be treated as a failed model, not a paper question.

One: clone resistance of the credential as configured. Not the card family in the brochure — the card family with the keys your property will use. The verification artifact: a written statement of the key-generation model. A credible answer names who generates the master keys, where they are stored, and whether sector keys are diversified per property. The failing answer is any variant of "the cards come encrypted at the factory."

Two: key custody of the encoding system. The encoder, the operator cards, and the master key file are the real security layer of any RFID hotel installation. Ask which components hold cryptographic material, whether a former employee can re-encode without authorization, and whether the encoding software supports an operator hierarchy with distinct access to guest, staff, floor-master, emergency, and build-master card classes. The card-management discipline behind this is substantial enough that we cover it separately under RFID hotel lock card management.

Three: reader-side tolerance. The 0–5 cm read window on a spec like the 9936S is a security feature as much as an ergonomic one: a reader that insists on a close, deliberate presentation shrinks the attack surface for opportunistic reads in a corridor. Verify what happens on a failed authentication — silent rejection, event logging, and any anti-passback or repeated-failure signaling. A lock that logs nothing on rejection has removed itself from your detection chain.

Four: audit depth and retrievability. An up-to-1,000-event buffer per lock is a strong standalone figure, but audit value is three questions: what fields each event carries, when the buffer rolls over, and how events are pulled to a report. Post-incident, a record that says "door opened at 14:32" without a card identity or unlock class is close to worthless. We break the record structure down in audit trail and event records; here, the selection test is simpler — demand one real exported event sample before you compare designs.

Five: loss-response speed. The honest security metric of a hotel credential system is elapsed time between "guest reports card lost" and "that card no longer opens anything." In standalone systems, invalidation typically waits for the next re-encoding cycle or the card's time-window expiry; in centralized systems, it can be near-immediate at the doors on the network. Ask the supplier to walk you through the lost-card workflow end to end, for both modes, in your room count. Slow loss response is the most expensive vulnerability in hospitality, because hotel card loss is routine.

Compress the five into a shortlist scorecard — one row per model, filled in only from written supplier answers:

PropertyVerification artifactPASS looks likeFAIL looks like
1. Clone resistance as configuredWritten key-generation modelNamed custody, per-property diversified keys"Factory encrypted," no details
2. Encoder key custodyEncoder/operator-role descriptionRole hierarchy, auditable encoding logShared login, vendor-held master keys
3. Reader toleranceShipped read range + failed-auth behaviorClose-range default, rejections loggedConfigurable to long range, silent rejects
4. Audit depthOne real exported event recordCard identity + unlock class per eventTimestamps only, no failure records
5. Loss responseStep-by-step lost-card walk-through, both modesMinutes at door (centralized) or bounded expiry (standalone)"Replace the cylinder" energy

Mifare Classic versus CPU Card: the Actual Trade

The card-technology decision is where security-first selection either pays off or gets quietly abandoned at the purchasing desk. Stated plainly, with neither camp's rhetoric:

A properly provisioned Classic-family fleet — hotel-owned diversified keys, no reused factory keys, controlled card stock, routine encoder access review, audit logs actually looked at — defends a low-to-mid-risk property within the limits of publicly demonstrated attacks, provided the mitigations hold. The weaknesses the public research documented are cipher-level, but the exploits consumer testers demonstrate depend on bad provisioning. Good provisioning is work, and work degrades: staff turnover, shortcut re-encoding, shared operator logins.

A properly deployed CPU-card system moves that risk from operational discipline into chip architecture. Keys never leave the card, so a compromised back office cannot mass-produce valid credentials. The costs are per-card price on every re-order of fleet stock, encoder and software tiers that actually support the chip OS, and — the binding constraint — a lock reader platform that supports CPU cards in the first place. The 9936S datasheet names the Classic family; CPU-card compatibility across the 24-model hotel range is exactly the kind of per-model fact to verify in writing at quote stage rather than assume from a category page.

Property profileReasonable credential postureNon-negotiable mitigationsWhen to insist on CPU-capable locks
Select-service, low turnover, domestic guest mixClassic family, hotel-owned keysDiversified per-property keys; quarterly key rotation; operator card controlNot required if provisioning holds
Urban full-service, high card-loss rateCPU where reader platform supports itFast centralized invalidation; anti-cloning sector policy on any residual Classic stockDefault position
Brand/franchise with a group security standardWhatever the group standard mandates — get it before RFQGroup key-management framework compatibility; encoder from approved listIf the standard says CPU, the standard says CPU
Resort or extended stay, long card lifetimes, mixed staff accessSplit fleet: guest Classic-tier acceptable with rotation; staff and master tiers hardenedStrict card-class hierarchy; staff cards individually identifiable in logsAny shared-credential history is a reason
Any property with a prior cloning or unauthorized-entry incidentCPU-tier throughoutIndependent review of encoder custody before reinstallImmediate

The Question Set for the Supplier Call

Fourteen questions, grouped by the subsystem that actually determines your security. Send them before the call if you want a serious answer; ask them live if you want to see one. What follows is written to be pasted into an RFQ — the procurement-formal version of the same discipline sits in the hotel smart lock RFQ guide.

About the card:

  1. Which card families does this specific model's reader support — Classic only, or enhanced and CPU cards? Answer per model number, not per category.
  2. What does the lock do on a failed authentication attempt? Is the attempt logged?
  3. Is the quoted read range configurable, and what is the shipped default?

About the keys:

  1. Who generates the master keys for our property, and where are they stored after provisioning?
  2. Are any factory-default keys involved at any stage, at any door, for any card class?
  3. Can we re-key the entire credential estate without changing a single lock? What does that operation actually cost in card fleet and labor?
  4. Does the encoding software enforce operator roles, and can we audit who encoded which card class this month?

About the log:

  1. Send one real exported event record. What fields does it carry — card identity, unlock class, failure type?
  2. At what event count does the buffer roll, and does rollover overwrite the oldest record on a door nobody has pulled logs from in a quarter?

About the fleet:

  1. What is the lost-card workflow in standalone mode, step by step, and in centralized mode?
  2. How many encoder tiers does the system need (front desk, back office, engineering), and what secures each encoder?
  3. Can we mix card generations during a migration year without locking guests out?
  4. Do emergency and floor-master cards use the same key tier as guest cards, or a separate hardened application?
  5. What physical security exists on card stock — blank pre-encoded cards, personalization-on-demand, or pre-personalized per room?

A good supplier answers fourteen questions in writing within one reply cycle. A nervous supplier answers the marketing-friendly three and reframes the rest as "standard for the industry." Score the answers, not the tone.

The Encoder Is the Real Front Door: Key Custody and Card Classes

Every decision above executes — or decays — at the encoding desk in the back office. The lock on the door inherits whatever the encoder produced. Treat the encoding system as a first-class selection item, with the same scrutiny as the hardware.

Two custody models circulate in hotel RFID systems. In a vendor-keyed model, the supplier's card stock arrives pre-provisioned under keys the supplier controls. It is convenient and it is weak: your credential estate shares a secret space with every other property that supplier has stocked, and re-keying after an incident means replacing the fleet. In a hotel-owned model, the property generates or commissions a unique key set at project start, and the supplier's role ends at manufacturing compatible cards under it. Ask which model you are buying. It is rarely on the datasheet.

Then there is the card-class hierarchy — the reason a "stolen housekeeper badge" incident and a "lost guest card" incident belong in different risk registers. A defensible hotel system separates, at minimum: guest cards (room-scoped, time-scoped), staff cards (zone-scoped, individually named in logs), floor-master cards, property-master cards, and emergency/build-master cards, each in a distinct application or key tier, each encoded only by authorized operator roles, each appearing distinctly in the audit trail. If a supplier's answer to "how does the system distinguish an emergency card from a guest card in the log?" is silence, the system distinguishes nothing.

An Illustrative Audit: Consider a 144-Room City Hotel

This scenario is illustrative — a composite of the situation buyers in this category most often present, not a customer story.

Consider a 144-room city hotel with 70% transient occupancy, roughly two dozen guest cards reported lost per month, and a 9-year-old lock estate on a Classic-family system. A security incident review flags that two staff departures left operator logins active. The GM's instinct is to replace locks. The credential-first method says: audit the system before spending on hardware.

Step one, inventory: 144 door locks, three encoders (front desk, back office, engineering), one master key file of unknown provenance, blank card stock purchased on demand. Step two, the tests from the five-property list: authentication fails log nothing on the oldest lock batch; event records carry timestamps without card identity; emergency cards share the property-master application; the fleet runs on a single sector key set reused across all room sectors.

The finding: the exposure is provisioning, not silicon. Two defensible outcomes follow. Path A retains the lock estate and re-provisions — hotel-owned diversified keys, sector-key rotation policy, individually named staff credentials, log pull scheduled monthly; card fleet replaced once, encoders re-commissioned. Path B accepts that the group standard now requires CPU-card architecture and that the existing reader platform cannot be configured for it, which converts the project into a lock replacement across the RFID door lock range — and, note, requires confirming CPU support per model, in writing, before any sample order.

The arithmetic that decides it is not a price list; it is the structure. Path A costs one card fleet plus policy. Path B costs a lock estate plus the same card fleet plus the same policy — bought to make provisioning discipline enforceable by architecture instead of by staff. Both are legitimate answers for a 144-room property; the wrong answer is discovering the difference after the purchase order.

Specifications Worth Holding Every Supplier To

Beyond credentials, three lines from a published hotel-lock spec class matter to the security chain, and they are worth stating as the minimum the buyer should expect an equivalent of in any quote: the 4 × AA power architecture with battery cycles quoted above 15,000 operations, which keeps the door functional through the low-battery window without wired risk; door thickness coverage of 36–100 mm on the 9936S class, which determines whether the credential upgrade can happen on existing doors at all; and multilingual management software in either standalone or centralized database mode, because an encoding desk staffed by rotating teams is only as secure as the interface those teams can actually operate. The full block is published on the SLK-9936S product page; use it as the baseline your comparison table starts from, not as marketing.

If the credential decision goes the other way — a property that decides CPU cards, hardened encoding, and fast invalidation justify centralized mode — that architecture question is its own selection problem, and it should be settled with door-by-door network reality, not brand preference.

Making the Decision Concrete

The sequence, compressed: settle the credential tier (Classic-provisioned vs CPU) against your property profile before looking at finishes. Demand the five verifiable properties — clone-resistant configuration, key custody, reader tolerance, audit depth, loss-response speed — as written answers per model, using the fourteen-question set as the RFQ appendix. Treat any spec line that names "Mifare 1" as the start of the card conversation, never the end of it. Score suppliers on how they handle the awkward questions: key provenance, failed-authentication logging, emergency-card isolation.

Exact configurations, supported card families per model number, and software tier for a given room count are settled per project — this factory publishes no fixed price list, and the quote is where the credential answers become contractual. Start that conversation at request a quote, and before anything is ordered, put the shortlisted platform through a bench test on real cards with order samples: encode, clone-test against your own key configuration, pull the event log, and read the record yourself. The report from that afternoon is worth more than every brochure in the folder.

Din kundvagn
Din upplevelse på denna webbplats förbättras om du tillåter cookies. Cookiepolicy
Chat on WhatsApp