A pms integration portfolio case turns on how one lock platform behaves across several buildings and on who is allowed to reconcile the drift.

Опубликовано: июл 09, 2025

PMS Integration for a Multifamily or Office Portfolio

PMS Integration for a Multifamily or Office Portfolio

A property management system and an electronic lock system are two pieces of software that describe the same building in two vocabularies. One knows that unit 4B is occupied under a lease that started on the first, and that the resident's name has two l's in it. The other knows that a credential may open a particular door until a particular moment. The distance between those two statements is the work, and in a pms integration portfolio case it has to be governed across every address at once.

Direct answer. The one-line answer — pms integration portfolio case is the working record for pms integration portfolio case. pms integration portfolio case is worthless until pms-integration is attached to it as evidence. A pms integration portfolio case treats the handover as the unit of work. pms integration portfolio case decides the interface. pms integration portfolio case is operable on day one only if it does.

  • Map the interface — write down what hands over between lock, door and system at pms integration portfolio case
  • Test the handover — record the sequence that proves pms-integration works end to end
  • Keep the record — store the commissioning evidence with the file for unit of failure stops
  • For pms integration portfolio case, the order below is fixed and pms integration is the reference point.

Put one property system and one lock platform across several buildings — some residential, some office, some a mixed block with a serviced floor above them — and the job stops being an interface and becomes a constitution. The rules that now exist once and apply everywhere are that constitution, and they decide what happens in a building you have never visited when somebody changes a setting at half past ten on a Tuesday.

The fault that hurts is rarely the one the specification anticipated. It is the one designed once and inherited by eight addresses. A rule granting a contractor access between eight in the morning and six in the evening was written for a commercial floor; it now applies to a plant room in a residential block where nobody expected a contractor. A mapping that pairs unit numbers to doors was correct when it was written, still is for six buildings, and has been quietly wrong for the seventh since a staircase was renumbered during a refurbishment nobody told the access team about.

I have watched this argument three times, and it resolves the same way each time. The teams that get a portfolio right decide early which things must be uniform across the estate and which must stay local to a building, and write both decisions down with an owner attached. The teams that get it wrong discover during the first lease-end week that the instance in question is shared.

What follows is an account of what that constitution must contain, written for the systems integrator, the IT lead and the access-control engineer asked to make one property system and one lock platform agree across a set of buildings. It is a narrative because the order of the decisions matters more than their format: you cannot write the reconciliation intervals before you know which rules are estate-wide, and you cannot know that until you have measured the variance between the buildings.

The situations described are illustrative composites. No property, operator, building, client or contract named here exists, and nothing here states what any particular product does.

The unit of failure stops being the door

In a single building, an integration fault has an address. Someone reports that unit 214 will not open, or that the plant room door is being held open overnight. The address is enough to find the cause, and the cost of the fault is contained by the wall it happened behind.

Across a portfolio, most faults have no address at all. They have a scope, and the scope is the estate. That is the first structural difference, and it governs everything else, so the layers are worth drawing before the rules are.

flowchart TD
    A["Property system holds the business truth"] --> B["Connector holds the correspondence"]
    B --> C["Lock platform holds the device state"]
    C --> D["Door population holds the physical reality"]
    A -.-> E["A fault here reaches every building"]
    B -.-> F["A fault here reaches every building"]
    C -.-> G["A fault here reaches every building"]
    D -.-> H["A fault here stays at one door"]

Four layers sit between a lease event and a door opening. The property system holds the business truth: who is entitled to be inside, and until when. The connector holds the correspondence: the mapping from a unit, a suite or a floor to a physical door. The lock platform holds the device state: which authorities exist on which doors. The door population holds physical reality: the survey, the frame, the strike, the fail state, the approval.

They fail differently. The property layer fails rarely and wide, because every building depends on the same instance. The connector layer fails wide as well, and it is the layer most likely to hold a mapping that is correct in most buildings and wrong in one. The platform layer fails wide or narrow depending on whether the arrangement is centralised or standalone. The door layer fails narrow, always. The same decision, read from a different angle, is covered in PMS Integration Cost and Effort Drivers Without Quoting Prices. Specifications at this level come off the smart door locks.

The design goal that follows is not reliability in the abstract. It is narrower and more useful — keep the narrow failures narrow, and make the wide failures boring. A wide failure that produces a known, rehearsed, documented behaviour at each door is an operational event. A wide failure that produces different behaviour in different buildings, because each drifted on its own, is a project.

A configuration that is correct in four buildings and wrong

Here is the sentence that ends most portfolio arguments: the rule is the same everywhere. It is nearly always true and nearly always incomplete, because the buildings differ. One block has a single entrance and a concierge desk; another has two street doors and a service entrance off a lane. The office building has a floor demised to one occupier beside a floor of hot-desked teams. The plant building has a basement that is a fire compartment rather than a room. A rule cannot be identical across all of them, and the moment it is not, the difference has to be recorded somewhere other than in the mind of the person who noticed.

The structure that works is three layers of configuration, resolved in a fixed order.

graph LR
    Base["Estate base rule"] --> Building["Building override"]
    Building --> Door["Door exception"]
    Door --> Effective["The behaviour the door applies"]
    Ghost["An override made during commissioning and never written down"] -.-> Effective

The base layer holds what is true of the estate: credential format, the shape of an instruction, door naming, the event vocabulary. It changes rarely, it changes centrally, and a change to it is a change to every building at once. The building layer holds what is true of one address: door classes, fail states, approvals, the maintenance window, the local escalation route. The door layer holds the exceptions — a plant room locked during a fire-alarm test, a storage cage a tenant manages themselves, a loading bay on another schedule.

LayerWhat it holdsHow often it changesWho owns itBlast radius of a mistake
Estate baseCredential format, instruction shape, naming, event vocabularyRarelyThe integrator, with the platform ownerEvery door
BuildingDoor classes, fail states, approvals, windows, escalation routePer project phaseThat building's custodianOne building
DoorAn opening that departs from its building's patternOccasionallyThe custodian, on the integrator's adviceOne door

The failure mode this structure exists to prevent is the one where a building override is created during a commissioning problem, never written down, and then inherited as that building's real behaviour. An undocumented override is indistinguishable from a base rule, and it is permanent.

A second portfolio-specific hazard catches integrators who have only run a single site. A phased retrofit means the estate holds more than one generation of device over its life, and the interface contract has to hold across all of them, because a connector cannot be versioned per building without becoming two connectors. Whether a model implements a given interface is a per-SKU question: confirm on the dated product page for the specific SKU.

CONFIRM-PER-SUPPLIER — whether the lock platform in scope exposes an interface that can issue, update and withdraw an authority against a single door identifier, and whether that interface is the same across every SKU the estate will hold at once. Obtain the answer in writing, naming the SKU and the date of the page it came from.

Door schedules that refuse to become one schedule

Every correspondence in this integration starts with an identifier, and the identifiers are the trouble. The property system speaks in unit numbers, suite numbers and floor plates; the lock platform speaks in device identifiers and door identifiers. Between them sits a mapping built from a survey of a physical building, which is why per-building door schedule mapping cannot be an estate-wide artefact even when two buildings are identical twins.

The twin-building case is instructive precisely because it looks as though it should be a single artefact. Two residential blocks, identical in every respect, acquired years apart, refurbished years apart, and renumbered differently at some point in one of them. Their surveys are separate documents with separate dates, so the mapping is two documents that may differ.

That asymmetry has to be absorbed by change control: a renumbering in one building changes one mapping, no message crosses the interface that day, and yet the access behaviour of that building has changed.

Property of a building's mappingWhat it must carryWhy the estate cannot assume uniformity
IdentityBuilding reference, door identifier, unit reference, readable locationDevice identifiers are assigned by the platform; unit references come from the property system
ProvenanceSurvey date, survey method, the door schedule revision it was built fromTwo identical buildings have two surveys with two dates
VersionRevision number, approver, and the change that prompted itA renumbering alters access behaviour without a message
RetirementHow a superseded revision is withdrawn from serviceA stale mapping keeps granting on the old door

None of those properties is difficult, and all of them are easy to skip on the first building and impossible to add later across nine. The measurement work behind the door schedule is covered in the Smart Lock Door Measurement Guide, and the survey discipline in the Hotel Smart Lock Site Survey. The naming there is hotel-specific because that is where the published material sits; the method is the same corridor walk. What matters is what happens to the survey afterwards: it becomes a versioned mapping with an owner.

A lease event is not an arrival, and a floor is not a room

The hotel is the easy case, because a hotel generates a small set of crisp events: arrival, departure, room change, extension. Everything else gets harder once the estate holds a residential block and an office building, because those generate different events, on different clocks, with different legal weight.

A residential block generates a lease event: the lease starts, ends, renews, notice is served, the handover happens, and a possession claim may be running in parallel. The access system needs three of those — start, end, change — and tenant lifecycle access provisioning has to be accurate on the third. An office building generates a tenancy event: a floor is let, a fit-out runs, a company moves in, downsizes, a contractor needs access for six weeks. The property system may represent some of that and none of the rest, so office portfolio access control ends up leaning on a badge system the property system has never heard of.

Business eventResidential formOffice formWhat the lock side is asked to doThe dangerous direction
Entitlement startsLease commencementTenancy start or employee transferGrant an authority on the mapped doorRefusal at a door where somebody is standing with a key
Entitlement endsExpiry, surrender, notice to quitTenancy end, desk closure, leaverWithdraw the authorityAuthority surviving its cause
Temporary accessContractor booked for a repairFit-out contractor, auditor, engineerGrant for a bounded windowA temporary grant that becomes permanent
Change of spaceResident moves to another unitTenant moves to another floor or suiteMove the authority and close the old doorAuthority still open on the vacated door
Disputed or pendingPossession dispute, arrears holdDispute over a handover dateNothing, unless the operator decidesSoftware deciding a dispute it has no standing to decide

That last row is worth slowing down on. A property system will carry a contested lease, and a lock platform will happily accept an instruction derived from it. Nobody should be deciding, in software, whether a contested possession belongs to the landlord or the resident. The access system should follow the operator's decision, recorded with an author rather than inferred from a field.

The portfolio changes volume and timing as well. A hotel produces events clustered around check-in and check-out; a residential building produces them at nine in the morning on the first of the month, in batches. The estate's design has to survive a burst, and the burst is where a batch job written for a hotel becomes a liability.

What the property system sends across the estate

The instruction stream runs from the property system to the lock side through the connector. Each message must carry enough for the receiver to recognise the business object rather than an anonymous event: the business reference of the lease, tenancy or booking, the building it applies to, the mapped door, the person or role behind it, the validity window, the credential requested, the source operator, and a sequence marker so a retry is recognisable as a retry.

What the property system sends is not what it knows. It also holds the rent, the arrears position, the deposit, the tenancy notes, the maintenance flags and the personal details of the people involved. Almost none of that should cross into the access system, and the specification should say so plainly, because the alternative is a temptation to add one more field whenever a fix would be easier with it. The clean rule is that the property system sends entitlements and the lock side sends events.

Multi-building property system integration changes the arithmetic of that rule in one important way. One property system instance serves many buildings, and the connector sits between it and one lock platform that also serves many buildings, so every building's traffic crosses the same middleware. That makes the connector the single point through which all entitlement in the estate flows, and it makes a portfolio-wide simultaneous failure the estate's most likely one.

The local API versus cloud API question is therefore not a preference; it is a statement about who is on the hook when the link fails. If the property system calls a local interface on a server the operator controls, a network fault is the operator's incident and the integrator's diagnosis. If it calls a cloud interface belonging to a third party, that party's availability sits in the critical path of every door in the estate, and the escalation route crosses a boundary nobody in the room can sign for. The interface shape on the lock side is set out in the Hotel Smart Lock API Integration Guide, and the vendor questions are collected in Hotel Smart Lock API Requirements for Integrators.

Two portfolio-specific warnings belong here. An instruction that does not name its building can be applied to the wrong door when two buildings share a unit reference, and in a portfolio that reuse is common. And the thing a person holds is not the person: in an estate where the same contractor works in four buildings, a credential scoped to a role rather than to a building will be presented at the wrong one.

The interval in which two systems are allowed to disagree

Every integration has a degraded mode, and the degraded mode here has a dimension engineers routinely under-specify: time. The two sides are allowed to disagree for a period, and the length of that period is not one number. It varies by event class, it varies by building, and it has a direction, because one direction of error is survivable and the other is not.

sequenceDiagram
    participant PMS
    participant Connector
    participant Platform
    participant Door
    PMS->>Connector: grant authority for unit 4B
    Connector->>Platform: grant authority for unit 4B
    Platform->>Door: apply the authority
    Door-->>Platform: grant applied event
    Platform-->>PMS: acknowledgement carrying the business reference
    PMS->>Connector: withdraw authority for unit 4B
    Connector->>Platform: withdraw authority for unit 4B
    Note over Connector: the second instruction does not arrive
    Platform-->>PMS: no acknowledgement inside the window

If the second instruction does not arrive, four things have happened, not one. It may be late. It may be retried and delivered twice. It may be queued on the property system's side and land at an unpredictable point. Or it may be dropped. The response at the door differs in each case, so the four must be distinguishable from outside the building, and the only way to distinguish them is to observe the acknowledgement and the event stream — which means the acknowledgement has to carry the business reference, not merely a transport counter.

Event classThe interval that mattersDangerous directionWhat closes itWho closes it
Lease or tenancy commencementUntil the person arrives with their keyNo authority yetThe grant applies and is acknowledgedProperty system owner
Move-out and re-lettingUntil the new holder arrivesAuthority on the vacated doorThe withdrawal appliesLock platform owner
Exception sweepUntil the nightly comparison is readBoth sides disagree and nobody is toldThe reconciliation report is produced and signedIntegrator, with the custodian

A single target applied to all of these is worse than having none, because it hides the difference between an instruction that is late and an authority that has outlived its reason. Write the interval per class and the dangerous direction beside it.

There is a portfolio twist to that table which a single-building design does not have: the interval cannot be one number for the estate, because the buildings have different working patterns. A residential block's relevant moment is nine in the morning on the first of the month; a serviced floor in the same block runs on the hotel clock; an office building's relevant moment is the end of a fit-out. The interval is a property of the building's rhythm, not of the software.

Who picks up the pen when a building disagrees with itself

Reconciliation is not one party. In a single building there are four and the argument is local. Across a portfolio the same four remain, but fewer people can play each role, and the distance between the person who can change the data and the person who can authorise the change grows.

Each of them owns a different kind of truth. The property system vendor is answerable for emitting instructions and for who is entitled to be inside. The lock platform owner is answerable for applying an instruction to a door, withdrawing an authority, and reporting what followed. The integrator is answerable for the correspondence between the two — the mappings, the interval boundaries, the variant register, the health of the connector — and for nothing else. The operator holds the business truth: whether a resident has moved out, whether a floor is fit to let, whether an override is permitted, and who signs it.

SymptomWho looks firstWho changes the dataWho authorisesThe record that must exist
A door refuses a person who is entitledThe custodian, then the connector ownerProperty system ownerThe custodian's supervisorThe instruction and its acknowledgement
A door opens for a person who has leftSecurity, not the leasing officeLock platform ownerThe operator, on the lease or tenancy recordA manual withdrawal with reason and authoriser
The wrong door answersThe integratorIntegratorIntegrator, with the custodian informedA new mapping revision and a reissued authority

The line that matters is the fourth column. A design that says "reconcile discrepancies" without saying who may authorise a change has not drawn the boundary, and an undrawn boundary becomes a phone call that lasts a year and gets resolved differently in every building.

One structural consequence of the portfolio is that the operator role is usually split. Somebody holds the lease relationship and somebody holds the building, and when they disagree an access decision becomes a corporate decision. The design has to say who may stop a door from opening, because in an estate where one platform serves nine buildings, a hold applied in one building can be applied by a policy that reaches all of them. The wider division of responsibility is in Who Owns PMS Integration: Supplier, Integrator or Client.

One clock for the estate, or one clock per building

Events with two clocks cannot be reconciled. A portfolio has more than two. The property system server stamps the moment it decided, the connector the moment it forwarded, the platform server the moment it applied, the controller the moment the door reported, the device the moment the credential was presented. Each is a different fact, and only some answer the question somebody will later ask.

That question is usually: was this person inside the building when their entitlement was withdrawn? It needs the timestamp of the physical event, not of the record being written, and the difference can be minutes in a congested system. The specification should name the authoritative clock for the record, state a drift allowance, and say what happens when the allowance is exceeded. Outside the allowance the event is flagged, because a flagged event is a question and an unflagged one is a number nobody revisits.

The portfolio version of this problem is that the allowance has to be stated once and then survive nine buildings, nine maintenance regimes and nine network paths. A block whose controller synchronises from a local appliance on the building network is not running on the same clock as a block that synchronises from the estate's own server. The difference is small in absolute terms, and it is the kind of small difference that becomes an argument in a report about a single event.

Two records exist along the way, and they must not be confused with each other. The audit log holds an event with its time and its source. The event memory is the buffer a device keeps locally when the server is out of reach. The gap between them is where events are lost: a controller that stores four hundred events through a network fault and replays them afterwards has behaved correctly, and one that discards them has failed, and only a staged test separates the two cases.

erDiagram
    ESTATE ||--o{ BUILDING : "contains"
    BUILDING ||--o{ MAPPING : "maps"
    BUILDING ||--o{ DOOR : "contains"
    MAPPING ||--|| DOOR : "resolves to"
    DOOR ||--o{ AUTHORITY : "honours"
    AUTHORITY ||--o{ EVENT : "produces"
    ESTATE ||--o{ EVENT : "aggregates"
    ESTATE {
        string estate_ref
        string clock_policy
    }
    BUILDING {
        string building_ref
        string variant_set_ref
    }
    DOOR {
        string door_id
        string controller_id
        string fail_state
    }
    MAPPING {
        string mapping_ref
        string revision
        string survey_date
    }
    AUTHORITY {
        string authority_ref
        string granted_at
        string expires_at
    }
    EVENT {
        string event_ref
        string occurred_at
        string source
    }

An estate-wide record has one property a single-building one does not, and it is the property that makes people nervous: it can be queried across buildings. A report answering "how many times did this person present a credential in this estate last quarter" is a different artefact from one answering the same question for one building, and the difference is not technical. It is a difference in the number of people whose movements it describes. The single-building trail is treated in Smart Lock Access Logs for Property Managers.

When the link goes quiet in every building at once

Three failures, three different scopes, three different people to call.

The estate network between the platforms goes, and multi-tenant lock platform management becomes a single outage affecting every address. Every building's entitlement flow stops at the same moment, and the doors in every building change behaviour at the same moment. The specification must state how long a pending instruction is held, who is told, what each building's staff do in the meantime, and how the backlog is drained without applying stale entitlements. The platform server goes, and the connector goes with it: both systems may be perfectly healthy while the thing that ties them together is not, and a connector that fails silently for a night is worse than one that fails loudly, because the drift accumulates quietly across nine buildings and is discovered by a resident rather than by a monitor.

The controller goes. This one is local to a building, and it is the failure the door's own behaviour covers. Fail-safe and fail-secure describe the state a lock takes when power is removed, and the decision belongs to the door's role and to the approval for that building — not to the product, not to the integrator, and in a portfolio not to the estate's template. A block with a single stair core and a block with two street doors do not have the same egress profile, and an estate-wide default on fail state will be wrong somewhere.

FailureScopeWhat the door doesWhat the staff doWho is calledWhat closes it
Estate networkEvery buildingPer the arrangement, stated per door classFall back to the local procedureThe operator's network team, then the integratorThe backlog is drained
Platform serverEvery buildingAs above, with loss of event deliveryAs aboveThe platform ownerThe event buffer is delivered
ConnectorEvery buildingAs aboveAs aboveThe integratorThe instruction queue is replayed or reconciled
ControllerOne buildingAs approved for that buildingThe local procedure for that buildingThe building custodianThe controller is restored or overridden

The offline case is the design, not the exception, and it is written before the integration ships. The arrangement itself — whether the authoritative state lives on a server or on the device — decides what the answer in the second column is, and that comparison is set out in Centralized vs Standalone Hotel Locks. What the portfolio adds is that the fallback procedure has to exist once per building, in the local language, signed off by the person who works there. And where the door is asked to decide on its own, the lock reports the state; the system decides what the state means.

The escape side is not an integration feature

Everything above concerns the front of the door. The escape side is a different discipline with different approvers, and in a portfolio the difference compounds: the fire strategy is per building, the access system is per estate.

A request to exit is the indication an exit device sends when a door is worked from the escape side. It is not a release command: the door lets go on the escape side because of how the exit device is built, and the signal only tells the system that the operation happened. A door position contact reports open or closed; it reports a state, not a cause. A log line that concludes a door was forced has inferred more than the contact knows.

On the fire interface the governing rule is that release never becomes the lock's decision. What a fire signal does to an opening is settled by the fire strategy and the approval for that building, and in a portfolio that approval is per building and per opening, so the estate's access design has to express a different answer for each of them.

CONFIRM-WITH-LOCAL-FIRE-AUTHORITY — which body signs off the fire-alarm interface for each opening in each building of this estate, whether that interface is proved during fire-system commissioning instead of access-system commissioning, and what record the commissioning has to leave behind. It is asked per opening, and one estate-wide answer does not satisfy it.

CONFIRM-WITH-LOCAL-BUILDING-CONTROL-BODY — whether the release on the escape side, and the state each opening takes when power is lost, appear in the approved classification for that building, and which document is the record of that classification.

The reader-to-controller path carries its own portfolio variance, because a retrofit estate will have older wiring in some buildings and newer in others. The older convention is known as Wiegand: a data line, usually accompanied by a power line and a ground, with nothing in the name that promises encryption or a way of telling one device from another. Whatever the wiring, the reader reports and the controller holds the door's state and takes the decision.

Written answers before a single door is configured

Consumer ecosystems and short-range radio standards — Matter, Thread, Zigbee, Z-Wave, Apple Home Key, HomeKit, UWB — appear in market copy for products in this category, and not one of those names may be written into a portfolio design without verifying it for the specific SKU on the dated product page. Each SKU in scope needs the supplier's written statement naming that SKU and the interface in question before the row counts as an available path. Where a phone-held credential is offered as the answer to the unit-to-door correspondence, the same discipline applies: it is issued, personalised, delivered, returned and revoked by a different set of parties than a card.

The question to putA usable answerA reason to go back
Which interface the lock side exposesA named local API or cloud API for the platform in scope, and the failure behaviour of eachA claim that it works with most property systems
Implementation per SKUA written, dated statement naming the SKU and the interfaceA family-level statement covering every model
Event and audit exportField list, format, delivery method, a sample exportA list of features described in prose
Instruction identityThe business reference carried, and how a retry is recognised as a retryAn incrementing transport counter alone
Degraded behaviourThe stated behaviour of the door and the instruction queue when the link failsSilence, or a claim that it is never offline

Portfolio-wide PMS integration is decided in large part by the first row of that table, because the local-versus-cloud choice places the failure boundary in a different organisation and the degraded mode follows from it. The second row matters more in a portfolio, because the estate will hold several device generations and the connector has to speak to all of them. Answers of this kind belong in the requirements set, and the template is in PMS Integration Requirements Document Template.

Sequencing so that the first mistake is the cheapest one

The rollout order in a portfolio is not a convenience. It decides the cost of being wrong, and in a pms integration portfolio case it is the decision with the longest tail.

The common instinct is to start with the easiest building: the smallest block and the simplest door population. That is the instinct of a team that believes a pilot's purpose is to prove the design works. In a portfolio the pilot's real purpose is to find what the design missed, and the building that finds the most is the one with the most variance. A block with a concierge desk, a service entrance, a mixed tenancy model and a numbering scheme that does not match the property system's references will find the mistakes a simple building hides. If the pilot is the hard building, the easy buildings afterwards absorb the residual risk. If the pilot is the easy one, the programme fails on the hardest building at the end.

Rollout positionBuilding typeWhat it is meant to proveWhat it actually proves
PilotHighest variance in the estate: mixed tenancies, two entrance types, a non-matching numbering schemeThat the mapping, the variant register and the intervals survive a hard caseWhere the base rule was wrong
First waveBuildings sharing the pilot's door classesThat the design repeats without improvisationWhether the variant register is a process or a document
Steady stateThe remainder, plus every later additionThat the estate runs without the integrator in the roomWhether the procedures were written down

Two rules come out of that. In a multifamily lock platform rollout the pilot is the block with the mixed tenancies, not the one with the simplest door population, and it must run long enough to cross a lease-end cycle, because the failure that hurts is the one that shows up the first time a resident moves out while the tenant's desk is quiet.

A portfolio also has to decide how the estate is changed after it is live. Firmware and configuration changes applied to the platform propagate to every building that shares it. A change window that works for one building has to become nine windows, sequenced around occupancy, and a change applied during a lease-end week will be noticed by the resident whose door was involved.

The parts of a portfolio that scale badly are almost never the ones that scale: it is the configuration changed by hand in one building during commissioning, the spare part held in one building's cupboard, and the override code written on a note in one building's office. Provisioning a device into service with its identity, its door and its owner is a recorded operation, and decommissioning one is a separate recorded operation; in a portfolio both have to be, because the estate's record is the only place that knows which building a device belonged to. That discipline is covered in the Smart Lock System Integrator Procurement Guide.

The questions that should be on the table in week one

These are the questions an integrator should answer on paper before the first door is configured, and each one has a follow-up that exposes whether the answer was real.

  1. Which layer of configuration is estate-wide, which is per building, and which is per door — and who owns each layer?
  2. Where does the exception register live, and what happens to an entry when its building is re-surveyed?
  3. Which clock is authoritative for a door event, and what drift allowance applies before an event is flagged?
  4. For each event class — commencement, end, temporary access, change of space, exception sweep — what is the permitted interval, and which direction of error is dangerous?
  5. What does a door do when the connector is unavailable, and what does the person working in that building do in the same ten minutes?
  6. Which state does each door class take on loss of power in each building, and who approved that state for that building?
  7. When two systems disagree, which one is asked to change its data, and who signs that decision?
  8. What is recorded when a door is overridden by hand, and where does that record live?
  9. Can a report be produced for one building without producing one for the others, and who may run it across the estate?
  10. How many generations of device will the estate hold at once, and which building is the pilot?
  11. What is the escalation route at two in the morning, and does it work for a person covering four buildings?

The ninth question is the one most often deferred, because the estate-wide report is genuinely useful and genuinely dangerous. A log recording who presented a credential at which door at which time contains personal data about identified people, and the period it is kept, the people who may read it, and what is owed to the people in it are decided by the operator's own position in the jurisdiction where the building stands. The earlier questions are engineering. That one is not, and it belongs to the operator from the first workshop.

CONFIRM-WITH-LOCAL-DATA-PROTECTION-REVIEW — in this jurisdiction, whether records of who presented a credential at which door at which time attract a defined retention period, whether an estate-wide query across several buildings is treated differently from a single-building query, who inside the operator may read the record, and what notice is owed to residents and staff.

Where this essay stops, and what it hands

It is worth marking the edges of this essay so that a reader can tell what it does and does not cover.

Nothing here is a claim about what any product does. Every statement about a specific product's capability, interface, protocol implementation or conformity route is a question to be answered in writing, on a dated page, for the specific SKU. A portfolio design cannot be taken against an answer that has not arrived.

It answers nothing that belongs to the jurisdiction where the buildings stand. Retention periods, the treatment of door events as personal data, the fire interface and the escape classification are decided locally, and this article carries them as questions with an owner rather than as answers it invented.

It describes no estate, no operator, no client and no contract. The scenarios are composites, written so that a reader can recognise a situation from their own portfolio.

It does not replace the design work. The constitution described here has to be drafted against a real door schedule, a real exception register and a real escalation rota.

What can be carried forward is the shape of the argument. A portfolio integration is decided by three things: which decisions are estate-wide and which are local, what the interval between the two systems is allowed to be when they disagree, and who is holding the pen when they do. None of the three is a hardware question, and all three can be answered before anything is ordered.

The seam between a property system and a lock system is described more generally in the Hotel Smart Lock PMS Integration Guide, and the decisions that have to be settled before any of it is drawn are set out in PMS Integration for Smart Lock Projects: What Buyers Must Define First. The catalogue of failures a portfolio meets is in PMS Integration Failure Patterns and How to Recover, and the test cases that prove the behaviour are in PMS Integration: The Test Plan a Buyer Should Demand. None of them replaces the constitution you have to write for your own estate.

On the supply side, the ranges suited to a residential or commercial portfolio are described on the apartment solution page and the office building solution page, and samples can be raised through order samples so that the questions above are asked against hardware in hand. When the base rule, the building variants, the exception register and the intervals are written down, the commercial conversation belongs at request a quote.

Ваша корзина
Ваш опыт использования этого сайта будет улучшен, если вы разрешите использование файлов cookie Политика использования файлов cookie
Chat on WhatsApp