> ## Documentation Index
> Fetch the complete documentation index at: https://docs.databunker.org/llms.txt
> Use this file to discover all available pages before exploring further.

# Frequently asked questions

### How does Databunker Pro manage sensitive data tokenization?

* Tokenizes entire user records (e.g., PII, PHI, KYC, PCI data) using UUID tokens or format-preserving tokens for specific data records (e.g., credit cards).
* Encrypts data with AES-256 and stores it in a secure vault.
* Provides access via RESTful APIs with role-based access controls (RBAC).
* Supports multi-tenancy for secure management of multiple clients.

### How is tokenization handled for multi-cloud or hybrid environments?

* Supports multi-cloud and hybrid deployments via Docker Compose, Helm charts, or a cloud-hosted version.
* Uses a stateless architecture for consistency across environments with a centralized secure vault for token mapping.
* Enables multi-tenancy for secure data separation in shared cloud or hybrid setups.
* Offers APIs for seamless integration with cloud-native or on-premises systems.

### Does Databunker Pro implement tokenization for PII, card data, or account information? What approach is used?

* Supports PII, card data (PCI), and account information (KYC).
* Tokenizes entire records with UUIDs or uses format-preserving tokenization for specific records (e.g., credit card numbers).
* Maintains data usability while ensuring security.

### Is there a secure vault for storing token-to-data mappings, or is a deterministic, format-preserving method used without a vault?

* Uses a secure vault to store AES-256-encrypted token-to-data mappings.
* Offers format-preserving tokenization for specific records (for example for credit cards).
* Ensures secure storage and retrieval with isolated vaults for different clients via multi-tenancy.

### Are encryption, secure APIs, and access controls in place during tokenization workflows?

* Encrypts data with AES-256 encryption.
* Uses secure RESTful APIs with RBAC to handle tokenization.
* Supports mutual TLS and certificate pinning for legitimate API access.
* Requires time-based tokens for special handling of bulk requests.
* Defines retrievable fields with a masking policy, masking others.

### What are the authentication, encryption, and access control mechanisms for the token vault?

* **Authentication**: Uses temporary UUID-based access tokens (more secure than JWT, as user identity like email/ID isn’t encoded). Supports passwordless options (e.g., one-time codes via email/SMS) for the optional user portal.
* **Encryption**: Employs AES-256 for data and vault storage with secure indexing for searches.
* **Access Controls**: Restricts vault access with RBAC and a masking policy defining retrievable fields (others masked). Tracks operations with audit trails. Ensures secure data isolation for different clients via multi-tenancy.

### Does the tokenization process align with regulations like RBI, PCI DSS, GDPR, etc.?

* Engineered to operate inside compliance programmes including **RBI, DPDPA, PCI DSS, GDPR, HIPAA, FERPA, ISO 27001, and SOC 2**. Databunker Pro provides the technical controls (AES-256 encryption, hashed indexing, CRBAC, audit trail, tenant isolation, key management) these frameworks require.
* Databunker Pro has historically been self-hosted, so customers typically inherit cloud-provider certifications (AWS, Azure, GCP) and add their own organisational SOC 2 / ISO 27001 on top.
* Product-level **SOC 2** attestation is in progress for the managed cloud offering (Databunker Portal) — expected within \~2–3 months. **ISO 27001** product-level certification is roadmapped alongside SOC 2.
* Supports data minimization, user consent management, audit trails, and a User Privacy Portal for data subject rights.
* Meets RBI's data localization and GDPR's privacy requirements via self-hosted in-region deployment.
* Provides multi-tenancy for compliance in multi-client environments.

### What about IRAP for Australian deployments?

* Databunker Pro is not currently IRAP-assessed at the product level.
* Because Databunker Pro is self-hosted inside your own cloud tenant, it inherits the IRAP boundary of the underlying platform — **AWS Sydney (`ap-southeast-2`) and Azure Australia East are IRAP-assessed at PROTECTED**, and Databunker Pro runs cleanly inside that boundary.
* A product-level IRAP assessment can be scoped under professional services if your contract requires it. Australian public-sector and university customers typically combine the cloud platform's IRAP assessment with their own organisational IRAP scope and treat Databunker Pro as a controlled component inside it.

### What is the de-tokenization policy? Is it role-based, audited, and strictly controlled?

* Role-based de-tokenization requiring RBAC permissions.
* Defines retrievable fields with a masking policy, masking others.
* Audits all operations with strict controls for compliance.
* Ensures tenant-specific de-tokenization via multi-tenancy.

### What audit and monitoring capabilities are provided for tokenization activities?

* Provides comprehensive audit trails, logging all tokenization and de-tokenization activities (user, timestamp, data accessed, data before and after change).
* Currently offers no special monitoring capabilities beyond audit logs.
* Segregates audit logs per client via multi-tenancy.

### How is high availability and disaster recovery ensured for the tokenization engine?

* Ensures high availability as a stateless service through containerized deployments (Docker, Helm) with load balancing.
* Supports disaster recovery via database backups (PostgreSQL/MySQL/Oracle) and replication.

### Is the tokenization system scalable to support large transaction volumes (e.g., millions of transactions per day)? What are the performance benchmarks?

Yes. Databunker Pro is built in Go on a stateless, horizontally-scalable architecture. A recent internal benchmark — a single Databunker Pro instance on AWS EC2 `m6i.2xlarge` backed by a **dedicated AWS RDS PostgreSQL** database (10,000,000 records, 120-field user profiles) — sustained **\~5,800 records/sec** on bulk writes with linear scaling, ingesting 10 M records in \~28 minutes. Detokenisation (`UserGet`) ran at \~14 ms p50. Throughput **scales by adding stateless Databunker Pro instances** against the same database. At small table sizes this is near-linear (a short 6-instance test reached \~34,400 records/sec); at large row counts the **database CPU** becomes the limit (a sustained 50 M, 4-instance run drove RDS to \~80% CPU at \~11,800 records/sec), so size up the database as you grow. See the [performance page](/pro/get-started/performance) for the full sizing guide.

For the full benchmark methodology, latency / throughput / storage numbers, and capacity-planning guidance, see [Performance & benchmarks](/pro/get-started/performance).

### How is token uniqueness ensured? Do you use randomization, hashing with salt, or cryptographic mapping?

* Ensures token uniqueness by checking for duplicate records in the database and regenerating UUID tokens if duplicates are found.
* Uses cryptographic mapping for format-preserving tokens with hash-based indexing and salts for deduplication.
* Maintains unique tokens per tenant via multi-tenancy.

### What cryptographic algorithms are used in token generation? Are they NIST-compliant?

* Uses AES-256 for encryption, SHA-256 for secure indexing, and cryptographic UUIDs or format-preserving methods for token generation.
* Aligns with NIST standards for encryption and key management.
* Supports secure multi-tenant environments.

### How are keys managed in the tokenization process? Are they stored in an HSM? What is the key rotation and lifecycle management policy?

* Manages keys securely with a **master key** (never exposed) encrypting sensitive data in the vault using AES-256.
* Protects the master key with a **wrapping key**, storable as a Kubernetes secret or retrievable from AWS Key Vault, HashiCorp Vault, or HSMs (requires custom development).
* Supports **Shamir’s Secret Sharing** for generating wrapping keys, requiring 3 out of 5 key shares to reconstruct.
* Configures key rotation following best practices for lifecycle management.

### What mechanisms prevent token mapping leakage or reverse engineering of the token? Is there protection against brute-force or pattern analysis?

* Prevents leakage and reverse engineering by using tokens (UUID-based or format-preserving) as pointers to AES-256-encrypted data in a secure vault.
* Ensures tokens contain no inherent data, making reverse engineering infeasible without vault access.
* Protects against brute-force or pattern analysis with RBAC, audit logs, and optional mutual TLS.
* Prevents cross-tenant leakage via multi-tenancy.

### How does Databunker Pro handle tokenization for structured and unstructured data? Is it applicable to database fields, documents, or images (e.g., OCR data)?

* Tokenizes structured data (e.g., database fields like PII, or credit cards in PCI).
* For unstructured data (e.g., documents, OCR-extracted data), recommends generating a random password, saving it in the user profile, and using it to encrypt the original file.

### What happens if underlying data is updated after a token is generated and sent to the cloud? Is a new token generated, or is the old token updated?

* Maintains the existing token’s validity, mapping to the updated data in the vault.
* Generates no new token unless a new record is created (checked via deduplication).
* Audits and encrypts updates, ensuring tenant-specific updates via multi-tenancy.
* Record versioning is supported as an optional mode (see [Record Versioning](/pro/concepts/record-versioning)). When enabled, every create and update retains an immutable version of the record with integrity checks, supporting forensic review, compliance audits, and rollback.