What Is Data Integrity?

Data integrity is the property that data has not been altered, corrupted, or destroyed in an unauthorized or unintended manner from the point of its creation through transmission and storage. The concept encompasses both the correctness of stored values and the assurance that those values remain unchanged unless a deliberate, authorized modification has been made. The NIST Cybersecurity Resource Center defines data integrity as "the property that data has not been altered in an unauthorized manner," a definition that applies across data at rest, in transit, and in processing. The discipline draws from cryptography, database engineering, and systems assurance, and sits alongside confidentiality and availability as one of the three foundational properties of information security.

Integrity applies at multiple levels. At the storage level, it concerns whether bits on disk or in memory match what was written. At the semantic level, it concerns whether values remain internally consistent and conform to their defined business rules. A database might satisfy storage-level integrity (no bit flips) while violating semantic integrity (a foreign key referencing a deleted record). Complete data integrity programs address both levels simultaneously.

Integrity Mechanisms and Quality Controls

Technical controls that enforce data integrity include cryptographic hash functions, checksums, and digital signatures. A hash function such as SHA-256 produces a fixed-length digest from an input data block; any modification to the block, however small, produces a different digest, making unauthorized changes detectable. Database systems use transaction mechanisms, specifically the atomicity, consistency, isolation, and durability (ACID) properties, to ensure that multi-step operations either complete fully or roll back entirely, preventing partial writes from leaving data in an inconsistent state. Input validation rules applied at data entry points reject values that violate defined constraints, such as dates outside a valid range or codes absent from a reference table. NIST Special Publication 1800-25, Identifying and Protecting Assets Against Ransomware and Other Destructive Events, addresses how integrity controls function as a defense layer, enabling organizations to detect and recover from integrity attacks including ransomware and malicious data modification.

Digital Preservation

Long-term data integrity introduces the additional challenge of preserving records across changing storage media, software formats, and organizational contexts. Digital preservation extends integrity assurance over timescales of decades, which exceeds the operational lifetime of any individual storage medium or application. Standard preservation practices include format migration (converting records from obsolete formats to current ones while documenting the transformation), fixity checking (periodically recomputing checksums and comparing against stored reference values to detect silent data corruption), and geographic replication across independent storage facilities. Provenance metadata, recording who created a record, when, and under what conditions, is an essential complement to fixity controls because it allows custodians to distinguish intentional authorized revisions from unauthorized modifications. The intersection with data governance is direct: a governance program's data stewardship roles and audit trails provide the organizational infrastructure that makes preservation technically meaningful.

Quality Assurance and Continuous Monitoring

Data quality assurance (QA) applies systematic testing and monitoring to verify that data meets defined standards for accuracy, completeness, consistency, and timeliness throughout its lifecycle. Statistical process control techniques adapted from manufacturing quality control are used to set control limits on key data quality metrics and trigger alerts when readings fall outside expected ranges. Automated data profiling tools scan production datasets to surface anomalies such as unexpected null rates, value distributions outside historical norms, or referential integrity violations. According to the IEEE Computer Society guidance on data lifecycle management best practices, continuous monitoring integrated into data pipelines is more effective than periodic audits because it catches integrity failures close to their source, reducing the cost of remediation.

Applications

Data integrity has applications in a wide range of fields, including:

• Financial systems, where transaction records must be tamper-evident for auditing and regulatory compliance
• Healthcare, where patient record accuracy affects clinical decisions and legal liability
• Scientific research, where reproducibility depends on provenance and fixity of experimental datasets
• Legal and regulatory archives, where records must remain unaltered to serve as evidence
• Supply chain systems, where inventory and logistics data consistency affects operational decisions across trading partners