data loss
Enable an agent to recognise suspected data loss, establish what information is missing or unusable, judge recoverability and consequences, and choose authorised preservation and recovery actions.
Bundle → Layer → Finding → Questions Filled
6 bundles · 11 layers · 18 findings · 28 questions
Loss definition and baseline Establishes what information was expected to persist and the boundary within which its loss is assessed.
Expected information
Identifies the information and version that should still exist.
Preservation expectation
Record the evidence that particular information existed or was committed for preservation, and who defines that obligation.
- Which records, objects, versions, or acknowledged transactions were expected to survive? definition
- What receipt, inventory, transaction log, agreement, or other evidence establishes that expectation? provenance
Assessment boundaries
Separates loss of a copy, loss of access, and loss of information.
Loss versus inaccessibility
State whose loss definition applies and whether the assessment concerns one copy, operational availability, or all known recoverable information.
- Does the applicable definition treat recoverable deletion or temporary inability to decrypt as data loss, and who uses that definition? definition
- What evidence distinguishes missing information from an intact copy hidden by permissions, indexing, connectivity, or service failure? boundary
Missing information and extent Describes exactly what is absent or unusable and how much of the expected information is affected.
Loss manifestation
Distinguishes deletion, overwriting, corruption, and loss of interpretability.
Unusable content
Identify whether content is physically absent, replaced by another version, or present but unusable because required information is missing.
- Is the expected content deleted, overwritten, truncated, corrupted, or inaccessible because a required key or interpretive dependency is missing? definition
- Which comparison, integrity check, decoding attempt, or application-level check demonstrates the affected information? measurement
Extent and uncertainty
Measures affected information against an explicit denominator.
Bounded loss inventory
Record confirmed missing units, potentially affected units, and intact units separately, including gaps in the baseline itself.
- How many records, objects, versions, or transactions are confirmed lost, and relative to which expected total? measurement
- Which portions cannot yet be assessed because inventories, logs, or comparison copies are incomplete? boundary
Loss timeline and propagation Reconstructs when information ceased to be preserved and how the loss reached other copies.
Event reconstruction
Separates the last known good state, the suspected loss event, and discovery.
Loss event evidence
Record the time bounds and supported causal explanations without treating discovery time as loss time.
- When was the affected information last verified intact, and when was its absence or damage first verified? measurement
- Which evidence supports deletion, failed persistence, media failure, destructive modification, or another cause? provenance
Copy propagation
Tracks whether destructive changes reached replicas, synchronised folders, snapshots, and backups.
Surviving copy boundary
Map known copies to their version and integrity status, including whether they share the suspected failure mechanism.
- Which replicas, snapshots, archives, exports, or recipient-held copies predate the suspected loss and remain independently usable? provenance
- Which authorised pause, isolation, or retention action would prevent further propagation or expiry of a potentially intact copy? action
Recovery and validation Assesses restoration or reconstruction and determines whether recovered information meets the required state.
Recovery candidates
Evaluates candidate copies, logs, and reconstruction inputs.
Achievable recovered state
Record the latest demonstrably recoverable state, its prerequisites, and the information each recovery option would leave missing.
- Which candidate can restore the required information, and are its keys, logs, formats, and dependent datasets available? action
- What time interval and specific changes would remain missing after using that candidate? measurement
Recovered information verification
Checks completeness, correctness, and consistency after restoration.
Verified recovery outcome
Assess recovery against the preservation baseline, including transaction consistency and relationships that file-level checks may miss.
- Which baseline comparisons and application-level checks demonstrate that recovered content is complete and usable? measurement
- Which records, versions, relationships, or transactions remain missing or uncertain after validation? boundary
Consequences and loss disposition Connects missing information to dependent activities and records decisions about further recovery or accepted residual loss.
Information-dependent consequences
Identifies decisions, operations, and evidence that require the missing information.
Affected use and priority
Link missing information to specific blocked or unreliable uses, distinguishing observed consequences from possible ones.
- Which transactions, decisions, services, or obligations cannot be completed or trusted because this information is missing? boundary
- Which information should be recovered first based on its dependent uses, deadlines, and available substitutes? action
Residual loss decision
Records the remaining uncertainty and authority for ending recovery efforts.
Documented loss disposition
State what remains lost, what recovery avenues were assessed, and the basis for continuing, suspending, or closing recovery.
- What evidence and search limits support the conclusion that no further recovery is presently achievable? provenance
- Who may accept the residual information gap, and what new evidence would trigger renewed recovery work? action
Evidence and external alignment What the world already says about this thing, gathered so the model can be checked against it.
Reported evidence
Findings from the breadth pass, kept separate from the structural claims.
Check these first
Recalled without web access and unsourced; every item is a lead to verify.
- Which of these check these first hold for the sense of data loss this model covers, and on what evidence? provenance
Kinds and varieties
Recalled without web access and unsourced; every item is a lead to verify.
- Which of these kinds and varieties hold for the sense of data loss this model covers, and on what evidence? provenance
Standards and regulation
Recalled without web access and unsourced; every item is a lead to verify.
- Which of these standards and regulation hold for the sense of data loss this model covers, and on what evidence? provenance
Real-world use
Recalled without web access and unsourced; every item is a lead to verify.
- Which of these real-world use hold for the sense of data loss this model covers, and on what evidence? provenance
Typical measurements
Recalled without web access and unsourced; every item is a lead to verify.
- Which of these typical measurements hold for the sense of data loss this model covers, and on what evidence? provenance
Failure modes and hazards
Recalled without web access and unsourced; every item is a lead to verify.
- Which of these failure modes and hazards hold for the sense of data loss this model covers, and on what evidence? provenance
Regional variation
Recalled without web access and unsourced; every item is a lead to verify.
- Which of these regional variation hold for the sense of data loss this model covers, and on what evidence? provenance
Neighbouring kinds and how to tell them apart
Recalled without web access and unsourced; every item is a lead to verify.
- Which of these neighbouring kinds and how to tell them apart hold for the sense of data loss this model covers, and on what evidence? provenance
Classifiers Filled
- Family
- Thing Registry
- Category
- Cross-cutting context
- Entry kind
- thing
- Plane
- XCT
- Domain
- XCT.QLT
What it is Filled
Data loss is the unintended destruction, disappearance, or corruption of recorded data such that the required information cannot be retrieved in usable form from the affected system, whether or not it remains recoverable elsewhere.
In scope
- The information expected to survive and the evidence establishing that expectation
- Missing, destroyed, overwritten, or corrupted information and the affected copies or versions
- The loss interval, causal mechanism, and propagation across replicas or backups
- Recovery options, dependencies, achievable recovery points, and validation
- Consequences of missing information and decisions about residual loss
Out of scope
- Unauthorised disclosure or exfiltration without loss of the original information
- Service outages in which preserved data remains intact and usable after access returns
- General backup architecture and routine storage administration
- Routine authorised disposal that satisfies the applicable preservation obligations
- Complete cybersecurity incident management beyond its data loss aspects
- General data quality defects that do not remove or make expected information unrecoverable
Why it exists Filled
Enable an agent to recognise suspected data loss, establish what information is missing or unusable, judge recoverability and consequences, and choose authorised preservation and recovery actions.
Distinguishing features Filled
- An instance requires identifiable expected information and a preservation baseline; an empty dataset alone does not establish loss.
- Failure to retrieve data establishes suspected loss, but intact data behind a failed service or permission barrier may indicate an access outage instead.
- Disclosure alone changes who can access information; data loss changes whether the expected information remains available in a usable form within the stated boundary.
- Deletion of one replica establishes copy-level loss, while intact recoverable replicas may prevent information-level loss.
- Corruption counts as loss to the extent that expected information cannot be correctly interpreted or reconstructed; changed bytes alone do not establish that extent.
What robots and AI may and may not do Filled
Must not
- Delete, overwrite or corrupt data it is not authorized to change.
- Hide or downplay a data loss from the data owner.
- Claim data is recovered without verifying it.
- Run bulk changes on important data without a backup.
Only with a human decision
- Notifying people or authorities when personal data is lost.
- Choosing to restore over current data, which may erase newer changes.
May
- Detect and report data loss on systems it is authorized to monitor.
- Restore data from backups when the owner asks and the restore is tested.
- Keep backups and verify they can be restored.
Moral aspects Filled
- Loss of records can harm people's health, money and rights.
- Hiding losses breaks trust and the law where personal data is involved.
Who is affected
- Data owners
- People whose data is lost
- Organizations relying on the data
Owners Filled
Steward
The data owner or controller; the system operator answers for safeguards.
Links to other meta-models Filled
parent
- Q124805656 - registry parent class
- Q3030513 - registry parent class
What else AI and robots need to interact with it Filled
Identity and identifiers required Filled
- Vercy registry: vr.tr.data-loss
- Wikidata: Q891179 (https://www.wikidata.org/wiki/Q891179)
Direct properties not applicable Not applicable
- Preservation baseline: Expected dataset, version, retention obligation, acknowledged write, or other evidenced preservation commitment
- Assessment boundary: Individual copy, storage system, organisation-held copies, or all known recoverable copies
- Loss manifestation: Missing records, deleted objects, overwritten versions, truncation, corruption, lost decryption capability, or undetermined
- Affected extent: Records, objects, files, bytes, transactions, or percentage of an explicitly defined baseline
- Loss interval: Earliest and latest possible loss timestamps, with timezone and uncertainty
- Recoverability: Unknown, recovery candidate identified, recovery demonstrated, partially recoverable, or no recovery demonstrated within stated search limits
- Recovery gap: Duration and identified changes between the latest validated recoverable state and the required state
- Evidence confidence: Suspected, corroborated, confirmed for specified information, or disproved
- Lost information dependencies: Affected decisions, transactions, services, obligations, and downstream datasets
Plane XCT: no invented physical properties.
Recognition not applicable Not applicable
Not applicable
Data loss is an information-system condition without an intrinsic visual appearance, and visible errors alone cannot establish it.
Capabilities and actions required Filled
- Compare observed data against an evidenced preservation baseline to identify suspected gaps.
- Trace affected versions and copies to distinguish local loss from wider information loss.
- Recommend authorised containment that limits overwriting and propagation while preserving recovery evidence.
- Assess recovery candidates for completeness, integrity, dependencies, and achievable recovery point.
- Validate restored information against the baseline and record residual gaps.
- Escalate unresolved or consequential loss with explicit evidence, uncertainty, and decision ownership.
Hazards and failure modes required Filled
- Untested backups can fail when needed.
- Silent corruption can spread into backups before it is noticed.
Standards and interfaces required Filled
- Regulation (EU) 2016/679, the General Data Protection Regulation, European Union.
- ISO/IEC 27001 Information security management systems, International Organization for Standardization and International Electrotechnical Commission.
Context of use required Filled
- Data loss is encountered when stored, transmitted or processed information becomes missing or unusable.
- The concept is used in incident response, backup planning and reliability assessment to distinguish damaged information from temporary access failures.
- Recovery work evaluates the affected information, remaining copies and consequences for dependent services or records.
Sources Filled
- Wikidata item Q891179: data loss - identity and sense of the item
- Wikipedia: Data loss - general description of the item
Open questions
- Which definitions used in storage engineering, security practice, and business continuity should this registry entry represent, particularly for recoverable deletion and temporary unavailability?
- What evidence threshold should justify describing information as permanently lost rather than unrecovered within a bounded search?
- How should the model distinguish never-captured information from loss of acknowledged or contractually expected writes?
- When should missing encryption keys, schemas, or software dependencies count as data loss rather than a separate access or interpretability failure?
- How should reconstructed or re-entered information be judged equivalent to an original when exact content, provenance, or evidential value cannot be restored?
Machine files
Provenance
thing registry research (pass 2) · unreviewed
Built from: models/things/publications/thing-q891179/spec.json