V5 Ultimate
Ultimate
PricingResourcesCompany
English (original)العربيةDeutschΕλληνικάEspañolFrançaisItalianoעברית日本語한국어NederlandsPolskiPortuguêsSvenska
Talk to us Log in Start free trial
Food & Beverage
  • Bakery & Confectionery
  • Food Processing
  • Ingredients & Dry Mixes
  • Sausage & Meat Products
  • Produce Packing
  • Pet Food & Pet Supplements
  • Dairy & Chilled Foods
  • Beverages & Liquid Manufacturing
  • Infant Formula
Pharma & Life Sciences
  • Radiopharmaceuticals
  • Pharmaceutical
  • Medical Devices
  • Veterinary Pharmaceuticals
  • Blood & Tissue
  • Diagnostics & IVD Kits
  • Reference Materials & Standards
  • OTC Topical Creams & Gels
Supplements & Personal Care
  • Dietary Supplements
  • Cosmetics & Personal Care
  • Cannabis & Hemp Manufacturing
  • Fragrances & Perfumery
  • Soaps & Personal Hygiene
  • Botanical Extracts & Essential Oils
Chemicals & Materials
  • Plastics & Resins
  • Agricultural Chemicals
  • Specialty & Industrial Chemicals
  • Printing Inks, Pigments & Construction Chemicals
  • Lubricants & Metalworking Fluids
  • Coatings, Adhesives & Sealants
Other Manufacturing
  • Consumer Products
  • Electronics & Electrical Assembly
  • Textiles & Nonwoven
  • Contract Manufacturing & Co-Packing
  • Packaging & Food-Contact Materials
  • Regulated Warehousing & Distribution
Explore all industries
  • Warehouse & TraceabilityReceive, store, pick and trace
  • Production & PlanningPlan jobs and guide production
  • Quality & LaboratoryTest, investigate and review
  • Documents & TrainingControl procedures and qualifications
  • Equipment & PerformanceMaintain assets and measure losses
  • Orders & IntegrationsConnect customers, suppliers and systems
  • Outputs & InsightsReports, KPIs and label design
Why V5? Product & Features
Why V5?Product & Features
  • Warehouse & TraceabilityReceive, store, pick and trace
  • Production & PlanningPlan jobs and guide production
  • Quality & LaboratoryTest, investigate and review
  • Documents & TrainingControl procedures and qualifications
  • Equipment & PerformanceMaintain assets and measure losses
  • Orders & IntegrationsConnect customers, suppliers and systems
  • Outputs & InsightsReports, KPIs and label design
Explore all industries
    • Bakery & Confectionery
    • Food Processing
    • Ingredients & Dry Mixes
    • Sausage & Meat Products
    • Produce Packing
    • Pet Food & Pet Supplements
    • Dairy & Chilled Foods
    • Beverages & Liquid Manufacturing
    • Infant Formula
    • Radiopharmaceuticals
    • Pharmaceutical
    • Medical Devices
    • Veterinary Pharmaceuticals
    • Blood & Tissue
    • Diagnostics & IVD Kits
    • Reference Materials & Standards
    • OTC Topical Creams & Gels
    • Dietary Supplements
    • Cosmetics & Personal Care
    • Cannabis & Hemp Manufacturing
    • Fragrances & Perfumery
    • Soaps & Personal Hygiene
    • Botanical Extracts & Essential Oils
    • Plastics & Resins
    • Agricultural Chemicals
    • Specialty & Industrial Chemicals
    • Printing Inks, Pigments & Construction Chemicals
    • Lubricants & Metalworking Fluids
    • Coatings, Adhesives & Sealants
    • Consumer Products
    • Electronics & Electrical Assembly
    • Textiles & Nonwoven
    • Contract Manufacturing & Co-Packing
    • Packaging & Food-Contact Materials
    • Regulated Warehousing & Distribution
PricingResourcesCompany
Start free trial Log in Talk to us
HomeGlossaryMES and SCADA Integration
Systems & integration · The complete guide

MES SCADA Integration: Meaning, Evidence and Controls

In short

Connect production context, supervisory observations and controlled requests without confusing message delivery with equipment execution.

4,178 words · ~19 min read
MES SCADA integration context: engineer at supervisory screens with process trends and a tablet, production line beyond
Decisions depend on the product, jurisdiction and approved procedure.
On this page
  1. 01What MES SCADA integration means—and what it does not
  2. 02Choose data flows and assign authority before connecting systems
  3. 03Define the data contract, including quality and missingness
  4. 04Bind observations to the correct order and phase instance
  5. 05Handle out-of-order events, replay and genealogy
  6. 06Separate command acknowledgement from physical execution
  7. 07Route alarms and exceptions without losing the original evidence
  8. 08Map US regulatory obligations to the intended record
  9. 09Protect the interface without obscuring local authority
  10. 10Test failure behaviour, not just successful data transfer
  11. 11Go-live checks that prevent recurring interface failures
On this page · 11 sections
  1. 1What MES SCADA integration means—and what it does not
  2. 2Choose data flows and assign authority before connecting systems
  3. 3Define the data contract, including quality and missingness
  4. 4Bind observations to the correct order and phase instance
  5. 5Handle out-of-order events, replay and genealogy
  6. 6Separate command acknowledgement from physical execution
  7. 7Route alarms and exceptions without losing the original evidence
  8. 8Map US regulatory obligations to the intended record
  9. 9Protect the interface without obscuring local authority
  10. 10Test failure behaviour, not just successful data transfer
  11. 11Go-live checks that prevent recurring interface failures
AI · Explain it for MY operation

How does MES and SCADA Integration apply to your shop floor?

Pick your industry and scale — Ask V5 rewrites the definition in your context, gives a worked example, and shows what V5 does on day one.

Your scale

01What MES SCADA integration means—and what it does not

MES SCADA integration is the governed exchange of production context, supervisory observations and authorised operational requests between a manufacturing execution system and a supervisory control and data acquisition system. An integration may be observation-only; bidirectional command exchange is a separately designed capability, not an inherent consequence of connecting the systems.

TopicPrincipal concernBoundary with MES SCADA integration
MES SCADA integrationContextual exchange and evidence across systemsDefines what exchanged information means and how it is handled.
PLC tag mappingSignal addresses, types and scalingIdentifies signals; does not establish their production-record meaning.
Machine data acquisitionCollection of equipment observationsCollects observations; does not necessarily govern contextual exchange.
Permissive conditionsLocal conditions for allowing an actionDetermine local eligibility, rather than cross-system record handling.
Clock driftAgreement and divergence between time sourcesAddresses timing alignment, not the complete exchange contract.

These neighbouring subjects contribute to an interface but do not replace its data contract: the shared definition of message content, meaning and handling. Their engineering boundaries vary between sites; naming an interface does not settle ownership.

The human-machine interface (HMI) display is also distinct from the underlying supervisory interface. A visible value or operator indication does not, by itself, establish the production record’s identity, context or delivery history. Interface assurance concerns the exchanged data and transaction behaviour, not merely whether a screen looks correct.

02Choose data flows and assign authority before connecting systems

Choose the exchange pattern by the meaning it preserves and its failure behaviour, not by connector availability. The following are engineering design choices, not universal system capabilities.

PatternIdentity, quality and context originsDelivery checkpoint and failure boundary
Event subscriptionProducer assigns event identity and source quality; producer or a designated binding service attaches production context.Consumer persists its checkpoint after durable receipt. Recovery depends on retained events.
PollingCollector assigns sample identity; endpoint supplies quality; context binding uses an explicit order/phase reference.Collector persists the completed poll boundary. Snapshots cannot establish unobserved transitions.
Historian-mediatedHistorian preserves source identity, quality and context, or documents transformations; extraction assigns separate delivery identity.Extractor persists a retrieval cursor. Recovery depends on retained detail, not merely available trends.

Use historian-mediated exchange for observations only when necessary semantics survive storage and retrieval; do not assume it provides a command interface.

Keep observation and request paths separate. In one negotiated site model, MES supplies the production order and requested operation; SCADA reports supervisory state and local execution context; the equipment-control design determines whether and how an action occurs. Agree this allocation against the installed architecture.

  • Automation owns observation schemas and acknowledgement meanings; manufacturing systems owns request schemas and production-context binding.
  • Operations owns the exception queue and coordinates recovery decisions with automation and manufacturing systems.
  • Quality owns record-impact review and disposition; security owns interface access policy and security-incident escalation.

Document the workflow as exchange → durable receipt → reconciliation → review. Receipt establishes persistence, not agreement between systems. Assign recovery authority at each boundary rather than implying a distributed transaction across every participant.

From supervisory exchange to reviewable evidence
  1. 1
    Agree OwnershipDefine contract, authority and recovery responsibilities.
  2. 2
    Bind ContextResolve order, phase instance and equipment assignment.
  3. 3
    Validate exchangeCheck identity, schema, context and quality.
  4. 4
    Persist ReceiptRecord evidence before acknowledging agreed durable receipt.
  5. 5
    Reconcile deliveryResolve gaps, duplicates and reordered events.
  6. 6
    Confirm outcomesCorrelate command acceptance with local execution evidence.
  7. 7
    Review dispositionAssigned humans resolve exceptions and approve records.
Message handling, execution evidence and human disposition remain distinct.

03Define the data contract, including quality and missingness

Treat the contract as an engineering agreement, not merely a list of tags. For every field, specify direction, data type, unit representation or non-applicability, nullability, identifier uniqueness scope, schema compatibility rules and accountable owner. The following field names are illustrative, not a standard schema. Rejection means exclusion from operational use while retaining the original message for investigation.

Contract elementExample field namesMeaning agreed by both systemsInvalid or missing handling
Message identityevent_id, producer_id, source_epoch, sequence_numberEvent uniqueness, originating producer, restart boundary and sequence scope.Quarantine ambiguous identities; flag conflicting duplicates.
Schemaschema_version, message_typePayload structure, purpose and supported compatibility.Quarantine unsupported versions or types.
Production contextorder_id, batch_id, operation_instance_id, phase_instance_id, equipment_id, context_versionApplicable production instances, equipment and context revision.Hold unresolved attribution; distinguish non-applicable fields.
Measurementsignal_id, value, unitSignal identity, typed quantity and unit representation.Retain invalid input; block incompatible conversions.
Qualityquality_raw, quality_normalized, quality_mapping_versionOriginal status, interpreted category and mapping revision.Classify unmapped flags as unknown.
Timesource_timestamp, received_timestampObservation time versus receipt time; agreed timestamp representation.Flag missing or invalid timestamps separately.
Provenancesource_reference, payload_digestOrigin reference and digest scope and algorithm.Flag unresolved origin or digest mismatch.
Request correlationrequest_id, idempotency_key, correlation_idRequest identity, repeat-detection scope and exchange linkage.Hold exchanges with unresolved required linkage.
Deliverystream_id, durable_sequence, gap_statusStream identity, durable checkpoint semantics and known discontinuities.Mark gaps; keep completeness unresolved.

Define each quantity as an instantaneous reading, cumulative counter or incremental contribution before permitting aggregation. Repeated counter readings are not separate contributions. Document conversion rules, including scale, offset and rounding where applicable; unlike units or quantity types are not interchangeable.

Illustrative normalised quality categories are good (passes agreed source-quality checks), questionable (requires qualification), bad (fails those checks) and unknown (status absent or unmapped). Preserve the source flag and mapping version. Keep missing data distinct from a reported poor-quality value: transport success and a non-null value establish neither measurement fitness nor suitability for aggregation. Retain rejected observations without silently substituting zero or the last accepted value.

Model the message envelope, production context, observation, command exchange, delivery checkpoint and review decision separately, linked by identifiers. The envelope describes the message; context identifies production; observations hold source facts; exchanges track requests; checkpoints track delivery; decisions record interpretation or disposition. Production context tagging supports attribution without embedding transport state in the observation. Preserve the original observation when interpretation or disposition changes. As a data-integrity design practice, declare retention and retrieval needs according to intended record use, rather than assigning a universal retention period.

An illustrative MES–SCADA evidence model

Message Envelope

Identifies origin, sequence and message timing.

  • event_id
  • producer_id
  • source_epoch
  • sequence_number
  • schema_version
  • source_timestamp
  • received_timestamp
  • payload_digest

Production Context

Binds evidence to a specific execution.

  • order_id
  • batch_id
  • operation_instance_id
  • phase_instance_id
  • equipment_id
  • context_version
  • binding_status

Observation

Preserves value and source-quality interpretation.

  • observation_id
  • event_id
  • signal_id
  • value
  • unit
  • quality_raw
  • quality_normalized
  • quality_mapping_version

Command Exchange

Separates request admission from execution evidence.

  • request_id
  • idempotency_key
  • command_type
  • context_version
  • requested_by
  • acceptance_status
  • execution_status
  • execution_event_id

Delivery Checkpoint

Tracks durable progress and unresolved gaps.

  • stream_id
  • source_epoch
  • durable_sequence
  • replay_from_sequence
  • gap_status
  • checkpoint_revision

Review decision

Records accountable disposition without replacing observations.

  • review_id
  • event_id
  • reason_code
  • disposition
  • reviewer_id
  • decided_at
  • supersedes_review_id
Separate source facts, delivery state, production context and later decisions.

04Bind observations to the correct order and phase instance

ISA-95, Enterprise-Control System Integration, addresses enterprise and control-system integration; ISA-88, Batch Control, addresses batch control. The order–operation–phase-instance model below is a practical design example, not a reproduction of either standard.

Give each order, operation execution and phase execution a stable instance identifier. A phase name describes an activity; a phase-instance identifier distinguishes a particular execution, including repetitions within an order. Equipment identity remains stable while its assignment to an order changes. Keep that assignment separate from equipment identity in production context tagging.

Bind each observation to context_version: an identifier for a preserved context mapping linking equipment with the relevant order, operation and phase instances. Record start, transition, completion and reassignment events in the process event log to explain how context changed without overwriting earlier assignments.

Where the source cannot carry production identifiers, use a controlled mapping. Record its provenance: source identity, mapping version, assignment evidence and responsible system or person. If those records cannot establish attribution, retain an unresolved state rather than silently selecting the currently active order.

Worked case assumptions: a categorical phase-state event has no physical unit and carries its original context_version. It is generated under order_previous and phase_previous, then buffered while the equipment is reassigned to order_current. Arrival context does not replace generation context.

Original contextArrival contextResulting disposition
context_version resolves to order_previous / phase_previousEquipment assigned to order_currentBind to the original order and phase instance.
Original context_version retained; instance mapping cannot be establishedEquipment assigned to order_currentPreserve event as unresolved; do not attribute it to order_current.

05Handle out-of-order events, replay and genealogy

Treat ordering and recovery as explicit interface-contract decisions. The following are proposed engineering controls, not regulatory requirements. Keep these coordinates distinct in the process event log:

CoordinateInterpretation
Source timeTime assigned by the originating source; identify what occurrence it timestamps.
Receipt timeTime recorded by the receiver when that delivery arrives.
Source sequenceOrdering value within the contract-defined stream, subject to its restart semantics.
Source epochIdentifier distinguishing source sequence lifetimes, including resets defined by the contract.

Preserve each timestamp’s declared representation and origin, even when storing a normalized value alongside it. Do not infer global causal order from timestamps alone. Sequence-based ordering applies only within the defined stream and epoch semantics. Clock agreement is an input to interpretation, not proof that separately generated events have a causal relationship.

Replay means requesting delivery again from a durable checkpoint that records recovery progress. For a checkpoint representing contiguous completion, leave it before any unresolved gap, even if later events have arrived. Detect duplicates by stable event identity, not timestamp or payload similarity. An identical redelivery adds delivery history but no accepted contribution; the same identity with a different payload goes to conflict review rather than silently replacing the accepted event. Retain both delivery history and the accepted event set. Replaying observations does not replay operational commands.

Worked case: delayed mass increments

Assume stable identities, the same source epoch and production context, and accepted incremental masses in kg. The source records event_alpha with q_alpha kg, followed by event_beta with q_beta kg. Delivery order is event_beta, event_alpha, then a replay of event_alpha. The accepted contribution is (q_alpha + q_beta) kg, not (q_alpha + q_beta + q_alpha) kg. A genuinely repeated physical operation produces a new event identity.

Carry lineage from each derived quantity to its accepted event identities and the production-context revisions used. If an event or context remains unresolved, expose that dependency in the derived record rather than presenting complete genealogy.

06Separate command acknowledgement from physical execution

As engineering practice, distinguish transport acknowledgement from application acceptance: delivery does not establish admission, and neither establishes physical execution. Use explicit command states rather than a single success flag.

StateEvidence availableWhat is not established
receivedMessage reached the defined endpointApplication acceptance or equipment action
acceptedRequest passed the declared admission checksExecution has started
executingLocal execution evidence reports progressSuccessful completion
completedDefined local completion evidence is correlatedBroader production success or quality disposition
rejected or failedReason and local status are recordedAbsence of partial equipment action
outcome unknownEvidence is insufficientEither execution or non-execution

Specify request identity, intended equipment and phase instance, context revision, authenticated authority, expected local state and downstream deduplication behaviour. Define whether a repeated identity returns the existing result or can initiate another action; retry safety depends on the actual command implementation.

Assign transition ownership: the endpoint records receipt; the supervisory admission handler records acceptance or rejection; local control supplies execution, completion or failure evidence. The command-ledger service records outcome unknown when evidence is insufficient and closes requests only against the contract’s correlated terminal evidence.

A lost acknowledgement leads to reconciliation against the command ledger and local execution evidence. Blind resend is not the default: first establish the existing request’s outcome and whether downstream deduplication remains effective.

Worked case: assume acceptance means admission to the supervisory request queue. This categorical-state example has no physical unit. A start_phase request is accepted, but local control subsequently declines execution. Preserve accepted and not_executed as separate facts, correlate the local reason with that request, and leave production progress in the batch execution history unchanged until supported by execution evidence.

Safety functions remain in the appropriate local equipment and safety-system design; a cloud MES is not the safety interlock. Local control evaluates the conditions assigned to it independently of the business-system connection. Keep this boundary explicit when designing interlock logic.

07Route alarms and exceptions without losing the original evidence

As engineering practice, keep equipment alarms, operator acknowledgements, interface delivery exceptions, invalid observations and manufacturing review decisions distinct. An alarm acknowledgement records a response to an alarm; it establishes neither process recovery nor a quality decision. Select alarm information for its production-record purpose rather than copying every HMI display event indiscriminately. Coordinate that selection with alarm flood prevention, without treating display activity as production evidence.

Route exceptions through the controls defined above, with an accountable owner and explicit closing evidence:

ExceptionImmediate handlingResolution ownerClosing evidence
Incompatible schemaQuarantine; retain payloadInterface ownerApproved mapping; reprocessing result
Questionable observation qualityFlag; withhold accepted useProcess/quality reviewerDocumented usability decision
Missing production contextHold attributionProduction ownerReviewed order/phase association
Conflicting event identityIsolate conflictInterface ownerIdentity reconciliation
Interrupted deliveryBuffer; expose queue stateOT/interface ownerReconciled delivery or gap record
Command outcome unknownReconcile before retryControls/operations ownerConfirmed outcome or unresolved-state handover

For controlled reprocessing after a mapping or context correction, retain the original payload, identify the interpretation version, and record the reason and reviewer. Associate the resulting decision with its predecessor in the batch execution history. Label the result as a corrected interpretation, not a newly received equipment observation. Closing an interface ticket does not close the associated manufacturing disposition.

For prolonged disconnection, use the declared buffering capacity and observable queue state to trigger escalation to the assigned owner. Track remaining capacity and unresolved delivery work. When storage or source-replay capability is exhausted, record irrecoverable gaps explicitly and route their production impact for review. Recovery remains bounded by the designed storage, source behaviour and operating procedures; an interface alone cannot promise lossless recovery.

08Map US regulatory obligations to the intended record

Start with the intended record, not the tag list. As an engineering scope-assessment method, identify which supervisory observations become production records, how those records are used, and which system maintains them. A tag inventory is not itself a Part 11 applicability decision; check the current text of Part 11 for scope. Assess supplements, food, devices, chemicals and other markets separately rather than extending the US drug example below.

The supported obligations below are binding within the stated scope. The final column contains illustrative engineering evidence—not regulatory wording, prescribed technical solutions or proof of compliance.

Jurisdiction and product or record scopeProvisionSupported obligationIllustrative interface evidence
US drug products21 CFR 211.68(b)Controls must restrict changes in master production and control records or other records to authorised personnel.Authorised-change records identifying the person and affected record.
US drug products21 CFR 211.68(b)Computer-system input and output of formulas, records or data must be checked for accuracy.Input/output reconciliation comparing source observations with stored production records.
US drug products21 CFR 211.68(b)The degree and frequency of input/output verification must be based on system complexity and reliability.Verification rationale linked to interface complexity and reliability.
US drug-product electronic records subject to Part 11 in closed systems21 CFR 11.10(a)Requires validation for accuracy, reliability, consistent intended performance and ability to discern invalid or altered records.Intended-use test results covering record transfer and alteration detection.
US drug-product electronic records subject to Part 11 in closed systems21 CFR 11.10(b)Requires accurate and complete copies in human-readable and electronic form suitable for agency inspection, review and copying.Retrievable record copies checked against the maintained record.
US drug-product electronic records subject to Part 11 in closed systems21 CFR 11.10(h)Requires device checks, as appropriate, to determine validity of the source of data input or operational instruction.Source-identity checks with accepted and rejected input examples.

For practical data integrity review, connect each applicable obligation to a named record, responsible owner and retrievable evidence. This keeps the assessment focused on record use rather than assuming that successful message delivery demonstrates adequate control.

Separately, US NIST SP 800-82 Rev. 3 provides OT-security guidance addressing performance, reliability and safety. Keep that guidance distinct from the binding US drug-record obligations and the engineering examples above.

09Protect the interface without obscuring local authority

For manufacturing OT, US NIST SP 800-82 Rev. 3 provides security guidance addressing OT performance, reliability and safety considerations. It is guidance, not a product-specific manufacturing regulation.

The following checklist is engineering practice, not a statement of NIST requirements. Apply it to the MES–supervisory interface using the local-authority boundary and data-contract definitions established above:

  • Separate business and supervisory network zones.
  • Permit only the defined exchange paths; identify the initiating system and permitted direction for each path.
  • Distinguish observation access from command authority rather than granting both through a shared access role.
  • Restrict service identities to their assigned functions.
  • Record configuration ownership, including who approves changes to exchange paths and access rights.

Evaluate these controls against the installed topology and operating needs. A network diagram describes intended separation; it is not proof of safe operation. Review whether actual routes, permissions and maintenance arrangements preserve the agreed boundary during normal operation and recovery.

As engineering practice, treat remote maintenance as a separately authorised access path, with accountable sessions and controlled changes. Assign ownership for time sources and document how timestamp origin or uncertainty is represented in the contract. Choose site-appropriate controls without inventing a universal timing tolerance or reopening the local-authority design through maintenance access.

10Test failure behaviour, not just successful data transfer

As engineering practice, build an intended-use test matrix linking each contract behaviour to a stimulus, expected stored evidence, operator indication and responsible reviewer. Separate observation-only from command-capable paths. Apply the regulatory mapping above only where product and record scope is established. For each scenario below, specify the contract’s disposition, its visible status and the reviewer responsible for accepting the result.

StimulusExpected evidenceUnacceptable outcome
Duplicated observationIdentity retained; duplicate disposition recordedDouble counting
Reordered deliveryContract ordering applied; arrival history retainedArrival-based attribution
Source restart with a new epochEpoch preserved; continuity disposition recordedIdentity collision
Incompatible schemaVersion captured; rejection or quarantine recordedSilent coercion
Quality transitionOriginal quality and mapped disposition retainedInvalid evidence treated as valid
Missing contextUnresolved attribution retained for reviewGuessed order
Interrupted acknowledgementCommand identity and unresolved state retainedAssumed execution
Unauthorised requestDenial recorded under access rulesRequest executed
Interrupted replayCheckpoint and unresolved gaps reconciledSilent omission

Observation case: assume stable event identities and a missing event between persisted neighbours. Compare the source event log, interface persistence, MES gap status and reviewable reconciliation record. Interrupt persistence and restart replay: an advanced checkpoint is not evidence that the gap was filled.

Command case: assume a retained command identity and a lost acknowledgement after local execution. Compare local execution evidence, interface command persistence, MES pending status and the reviewable recovery record. Retry across the acknowledgement boundary: the retry cannot establish another physical execution; reconcile against local evidence.

Include controlled schema, quality-map and context-mapping changes, rollback and restoration in review. Capture deployed versions and the rationale for coverage. Use site-approved acceptance criteria, not invented thresholds; document mismatches in a CSV deviation report and retain their resolution.

11Go-live checks that prevent recurring interface failures

Use this deployment checklist to confirm ownership and evidence for the controls covered above:

  • Contract: named production, automation and MES owners approve the deployed data contract.
  • Context and quality: unresolved-context and poor-quality queues have operating procedures, assigned reviewers and escalation routes.
  • Recovery: replay and restoration tests have closing evidence reconciling recovered events, duplicates and outstanding gaps.
  • Commands: operators can see acknowledgement, execution and unresolved command states.
  • Security: network boundaries and remote-access administration have named owners.
  • Versions: change control covers both schema versions and interpretation versions, including effects on existing records.

In this operating model, assigned humans review exceptions, approve records and make applicable release decisions. A successful exchange, a completed command or a cleared interface queue is not itself a product-release decision.

Frequently asked questions

Q.What should a supplier demonstrate before a historian-fed MES interface is accepted?+

Ask the supplier to trace representative events from SCADA through the historian into MES, preserving source identity, engineering units, quality status, event time, receipt time and order/phase context. Demonstrations should include compressed or interpolated values, missing context, disconnection, duplicate delivery and out-of-order replay. Acceptance evidence should explain every missing, rejected or quarantined event, not merely show matching trends. Define these challenges in the historian interface specification and demonstrate recovery through operational qualification appropriate to the interface’s intended use.

Q.Who resolves an order or phase assignment when MES and SCADA disagree?+

Use a documented authority matrix rather than assuming either system always wins. A practical arrangement assigns production responsibility for confirming the executed operation, automation responsibility for interpreting equipment events, and quality responsibility for assessing record impact where relevant. Quarantine disputed assignments instead of attaching events to whichever order arrived last. Preserve both assertions and the resolution rationale. The production context model should identify who owns order assignment, while the phase definition should distinguish an intended phase from an observed execution.

Q.How can a production record remain reviewable without reproducing every HMI alarm?+

Define which alarm events explain process execution, interventions or exceptions, then retain their identifiers, relevant transitions, acknowledgements and links to supporting history. Keep display-only notifications separate from production evidence under a documented selection rule. A concise record can summarize an alarm episode without concealing its underlying sequence. Test that reviewers can retrieve the supporting history after archival or system migration. Exception-based review works best when the process event log preserves traceable detail rather than relying on screenshots of an HMI.

Q.What should an operator do when a command has been sent but its outcome remains unknown?+

Treat the command as unresolved, not failed. Follow the local procedure to establish equipment state and check the command identifier, acknowledgement stage and execution evidence before considering another attempt. Receipt or acceptance by SCADA does not establish physical completion. Avoid blind retries, especially where repetition could repeat a material transfer or equipment action. Escalate an unresolvable outcome and preserve the command history. Keep safety decisions within the appropriate local design; a cloud MES should not serve as the equipment’s safety interlock.

Q.How should a late context correction be handled after a production record has been approved?+

Use a controlled amendment linked to the approved record, recording the original assignment, corrected assignment, reason, supporting events and responsible reviewers. Assess whether calculations, genealogy, exception conclusions or disposition decisions depend on that assignment. Do not silently regenerate an approved report under the corrected context. For US human-drug records subject to Part 11 in closed systems, 21 CFR 11.10(e) requires changes not to obscure previously recorded information. An audit trail review workflow can support assessment of the amendment and affected records.

Q.How should US manufacturers approach Part 11 scope assessment for SCADA-derived electronic records?+

Assess the record and its intended use, rather than using tag connectivity as the scope decision. Document what is retained, what supports manufacturing decisions, and how electronic and paper components relate; assess product-specific applicability separately. For US human-drug electronic records subject to Part 11 in closed systems, 21 CFR 11.10(b) requires the ability to generate accurate and complete human-readable and electronic copies. A hybrid record assessment should identify which components together form the reviewable record, including necessary SCADA-derived context.

Q.Which interface changes should trigger renewed testing and controlled deployment?+

Review changes that alter meaning, authority or recovery behavior: tag semantics, units, quality-code translation, order/phase binding, timestamp handling, historian compression, replay rules, command states, permissions and schema versions. A transport change also deserves assessment if it changes delivery ordering or duplicate behavior. Select regression tests from the affected risks, including failure and recovery paths, rather than rerunning only successful transfers. Use change control to document deployment and rollback decisions, with a traceability matrix connecting affected expectations to test evidence.

Q.What should an MES outage procedure say about local operation, evidence buffering and subsequent reconciliation?+

Specify which operations may continue locally, who authorizes that mode, and what conditions call for a controlled stop under the equipment’s local procedure. Identify where evidence is buffered, how buffer exhaustion is detected, and how operators record context that MES cannot provide. On recovery, reconcile events against actual execution using stable identifiers and duplicate controls; do not replay historical commands as new instructions. Keep unresolved gaps visible in the batch execution history, and assign responsibility for assessing completeness before normal record approval resumes.

Primary sources

  • NIST SP 800-82 Rev. 3
  • 21 CFR 11.10
  • 21 CFR 211.68
Software that covers MES and SCADA Integration
V5 Ultimate (21 CFR Part 11)
V5 Ultimate gives FDA-regulated manufacturers bound e-signatures, an audit trail, authority checks and printed-copy controls. An…
V5 Ultimate MES Software
V5 takes planned work orders to a kiosk at the line. Operators follow the configured steps, record weights and checks, and sign.…
V5 Ultimate Manufacturing Execution System
Work orders, electronic batch records, OEE, deviations and calibration on the same platform as your quality records, with ERP…
V5 Ultimate eBR Software
Electronic batch records with bound e-signatures, an audit trail and deviation workflow alongside execution, replacing paper…
Talk to us about MES and SCADA Integration

Want to see how MES and SCADA Integration could fit into your own records and workflows? Explore the related V5 pages or talk to our team about what applies to your operation.

Start free
Back to glossary
Where this term comes up
PharmaceuticalMedical DevicesFood ProcessingCosmeticsChemicalsRadiopharma
Inside V5
  • → eBMR / eDHR — the batch record fills in as the work is done.
  • → 21 CFR Part 11 — signatures and audit history your team can validate.
Example reports
V5Ultimate

Sample report

Electronic Batch Manufacturing Record (eBMR)

Plan & make

Electronic Batch Manufacturing Record (eBMR)View on screen or print
V5Ultimate

Sample report

Electronic Device History Record (eDHR)

Plan & make

Electronic Device History Record (eDHR)View on screen or print
V5Ultimate

Sample report

Compliance Readiness Assessment

Quality system

Compliance Readiness AssessmentView on screen or print
V5Ultimate

Sample report

Validation Support Pack

Quality system

Validation Support PackView on screen or print
→ All example reports
Regulatory anchors
  • ISA-95
  • ISA-88
  • 21 CFR Part 11
  • EU GMP Annex 11
Related terms
  • → ISA-88
  • → Machine Data Acquisition
  • → EBR
  • → Audit Trail
  • → GAMP 5
  • → 21 CFR Part 11

Next Step

Try V5 with your own records, or ask a question first. Ask V5 opens with an editable question; nothing is sent until you choose to.

Start your free trial Browse All Features

Showing the site for the United States. Local spelling, dates and prices are applied automatically; change the language at the top of the page.

V5 Ultimate
Ultimate

Warehouse, quality and manufacturing software for regulated operations.

ProductIndustriesPricingResourcesSecurity & TrustCompanyLegal Centre

© S.G. Systems, LLC