2026.4 version release (July 2026)

Failover support for easier testing, disaster recovery

DCT 2026.4.0 introduces disaster recovery for the DCT appliance's Continuous Data and Administrative capabilities. You can pair two appliances so that a secondary continuously replicates the primary's application database. If the primary is lost, you promote the secondary to take over with minimal data loss. Disaster recovery is configured and monitored from a new Disaster Recovery tile on the appliance Setup Dashboard.

Key capabilities

• Continuous streaming replication from a primary appliance to one or more secondaries.

• One-click promotion of a secondary to primary during a disaster, with a split-brain safeguard.

• Automatic recovery from common replication interruptions, plus manual resynchronization.

• Optional full TLS certificate validation of the primary, managed through the existing Network Security tile.

• Health monitoring through the Disaster Recovery tile, including replication state and lag.

For additional information, see the Automated disaster recovery documentation.

GCP and KVM appliance support

The DCT appliance can now be deployed on Google Cloud Platform (GCP) as a Compute Engine instance. Delphix publishes a public GCE Custom Image for each DCT release, which can be referenced directly when creating an instance — no download or manual import is required. For air-gapped or restricted environments, a manual import path using a tar.gz file and the gcloud CLI is also available.

Installation of the DCT appliance is now also supported on KVM, including KVM-based hypervisors such as Nutanix AHV and Proxmox. As with Continuous Data, deploy the qcow2 image that Delphix publishes and import it into your KVM environment (e.g., virt-manager or virsh).

To get started, see the GCP appliance and KVM appliance documentation.

Column filter default matching updated to "Contains" with case insensitivity

The default matching behavior for column filters across all DCT data tables has been updated from "equals" to "contains” with letter casing ignored. Column filters now return partial text matches by default, providing a more flexible and intuitive filtering experience when working with tabular data.

Offline (air-gapped) Telemetry simplification

The Delphix Telemetry's offline registration and upload flow is now simpler. The Admin > Telemetry registration wizard pre-populates the registration command with your instance detail; you paste in your Delphix Account API Token and copy the completed command, then run it in your terminal. Recurring exports now upload directly through the upload.delphix.com UI instead of manual API calls.

DCT Toolkit documentation improvements

DCT Toolkit has been overhauled to make it easier to learn and use. Within the CLI itself, command help is reorganized: commands now carry summaries and descriptions, groups and arguments are ordered logically, and required versus optional fields are clearly distinguished.

Alongside this, the published toolkit documentation has been rebuilt with a new getting-started guide, a broader set of worked examples, and a dedicated reference mapping DX toolkit commands to their DCT toolkit equivalents. See more in the DCT Toolkit topics.

Timeline bookmark import

New API endpoint, POST /bookmarks/transfer-engine-timeflow-bookmarks, that migrates timeflow bookmarks created directly on a Delphix Continuous Data Engine into DCT-managed bookmarks, so they become visible in DCT and usable for Provision, Refresh, and other operations. These were API-only bookmarks that were frequently used for custom self-service implementations. The endpoint supports a dry-run preview before you commit and can be scoped to a dataset group, a single VDB, or a single dSource.

SSH key authentication for SFTP file connectors

SFTP file connectors now support public key authentication using SSH private keys as an alternative to username and password. Register an SSH private key in DCT using the API, then select it when creating an SFTP connector. No password is required when using this method. For more information, see Register an SSH key and Create a file connector.

Password vault support for FTP, FTPS, and SFTP file connectors

FTP, FTPS, and SFTP file connectors can now retrieve their login credentials from an external secret manager at run time instead of storing a username and password directly in DCT. HashiCorp Vault, CyberArk, GCP Secret Manager, and Azure Key Vault are supported. Configure a password vault and credential path under Compliance > Delphix Infrastructure > Password Vaults, then reference the credential path from the file connector. For more information, see Password vaults and Configure a file connector to use a password vault.

File masking on cloud storage

DCT Compliance now supports sensitive data discovery and masking directly on cloud object stores without requiring local file systems or mounts. Connectors can be created for Amazon S3 (with AWS_SECRET key-based or AWS_ROLE instance profile authentication), S3-compatible services such as MinIO and Ceph, and Azure Blob Storage (with AZURE_SECRET access key or AZURE_MANAGED_IDENTITY authentication). Credentials can be stored in DCT or retrieved from cloud provider identity services at runtime. For more information, see Cloud file connectors.

Mount support for file connectors

Compliance engines can now mount remote file systems to access files for masking and discovery operations. VM-based engines support NFS and CIFS mounts, which are created through the DCT API and managed by DCT throughout the job lifecycle. Containerized engines support external Kubernetes-managed mounts that are discovered and tracked automatically. Mounts are created once and reused across multiple jobs. For more information, see Mounts.

Support for multiple file formats

File masking now supports five file formats: delimited, fixed-width, JSON, XML, and Parquet. Delimited files support configurable delimiters (including control characters), text enclosure, and escape strategies. Fixed-width files support precise column position definitions. JSON, XML, and Parquet files are parsed and exposed as structured fields for targeted masking. File format is immutable after connector creation and is defined in the Resource Files section. For more information, see Resource Files.

Discovery and masking jobs for files

Create discovery jobs to analyze file data and automatically assign masking algorithms based on content patterns and discovery policies. Create masking jobs to apply the algorithms and generate masked output files. Both job types work across all supported file formats (delimited, fixed-width, JSON, XML, Parquet) and connector types (FTP/SFTP, mounts, and cloud storage). Compliance nodes receive ephemeral copies of connectors only during job execution, minimizing credential exposure. For more information, see Create a discovery job for files and Create a masking job for files.

Database Document Store support

Data Control Tower now supports masking structured data stored inside a single database column by treating JSON and XML values as document stores. Instead of masking the entire column as one opaque value, you can decompose eligible columns into named inner fields and assign a masking algorithm to each field independently. For more information, see Edit masking rules in a rule set and Refresh a rule set.

Extended database support

Continuous Compliance in DCT now supports all standard file types and custom JDBC drivers for compliance workloads. Custom driver support removes the previous limitation to the built-in driver set, so new applications can connect non-standard databases. For new applications onboarded onto DCT, all standard databases and file types are now supported, except Mainframe files.

Additional authentication modes for DCT Compliance

Continuous Compliance in DCT now supports Kerberos and Password Vault-based authentication for compliance workloads. Password vaults are registered centrally with DCT, which then ensures the right credentials reach the correct orchestrated workload.

Hyperscale Compliance on DCT — Oracle (Early Access)

DCT now includes an integrated user interface for managing Hyperscale Compliance workflows for Oracle databases. Users can configure and execute Hyperscale masking jobs, monitor orchestrator flows, and track job execution and status, all from within DCT, without needing to interact with Hyperscale directly.

This release brings together orchestrator flow management, job creation and monitoring, and real-time execution tracking into a unified DCT experience, with algorithm synchronization across Hyperscale nodes handled automatically in the background.

To participate in the Early Access program, please contact delphix-early-adopters@perforce.com.

Added support for canceling, restarting, and cleaning up Hyperscale executions

Starting with 2026.4, you can:

• Cancel a Hyperscale job and use Execution Status to monitor the cancellation process.

• Restart a job if a load or post-load step fails.

• Restart a job if an execution is canceled.

• Clean up data generated during a job run.

To learn more, see Clean up data after a Hyperscale job execution.

Synthetic Data (Early Access)

Synthetic Data is now available as an Early Access feature in Data Control Tower. For more information, see Synthetic Data in DCT.

Synthetic Data enables teams to generate realistic and representative datasets for non-production environments without using production data. Using schema metadata and generation rules, you can create synthetic datasets for development, testing, and AI initiatives while supporting privacy and compliance requirements.

Key capabilities

• Application-based synthetic data generation workflows.

• AI-assisted discovery to identify relevant tables based on testing scenarios.

• Configurable generators for producing realistic synthetic values.

• Referential integrity preservation across generated datasets.

• Job-based execution and delivery of synthetic data to target environments.

• Bring your own LLM for AI features.

• Conversational AI (Delphix AI Assistant) — converse with the assistant for the entire workflow.

In addition, for all customers entitled to DCT Enterprise, we have included the ability to configure 2 reference connectors. This allows you to discover two source databases and generate any number of synthetic copies. We encourage you to try it out today!

To provide feedback or learn more about the EA program, please contact delphix-early-adopters@perforce.com.

BYO AI provider

Synthetic Data now supports a Bring Your Own Provider (BYOP) model for LLM-powered features. You can configure external LLM providers and assign them to features such as AI Assistant and LLM-based discovery workflows. For more information, see Configure LLM providers.

AI Assistant for Synthetic Data

The AI Assistant is now available in Synthetic Data to help create and configure resources through natural language interactions. The assistant supports guided create and update operations and uses confirmation workflows for mutating actions. For more information, see AI Assistant.

Discovery enhancements with AI agents

Synthetic Data discovery now includes enhanced AI-driven workflows that combine AI-assisted table identification with LLM-based generator assignment for unclassified fields. These enhancements improve dataset onboarding by reducing manual table and generator selection effort. For more information, see Create a dataset and Synthetic Data discovery methods.

Synthetic Data licensing controls by tier

Synthetic Data licensing now supports tier-aware access controls and enablement behavior across Core, Self-Service, and Enterprise licenses. Enterprise supports controlled access limits when Synthetic Data is disabled, while enabling Synthetic Data unlocks full functionality. For more information, see Synthetic Data licensing.

Generator enhancements

This release introduces a set of Synthetic Data generator enhancements delivered. Seed List, Regex, Multi Data, and Fixed Value generators now automatically trim generated values that exceed the target field's maximum length, precision, or scale. Functional Generator now supports single-value and multi-value generation through Multi-Value Output, supports Seed List inputs in single-value mode, and supports JSON/XML string output. Functional input terminology is updated to use Field, Structure, and Connector. Multi Data methods that are locale-sensitive now support an optional per-field Locale override for generated output.

Delphix MCP Server — Complete CRUD operations (Preview)

The Delphix MCP Server has been extended with full CRUD operations for the entire Delphix platform, including Continuous Data, Continuous Data for PaaS, Continuous Compliance, Synthetic Data, and Administrative actions. For additional information, see the MCP Server GitHub repository.

Delphix Terraform Provider — GCP support

The delphix_engine_configuration Terraform resource now supports Google Cloud Platform (GCP), enabling infrastructure-as-code deployment and configuration of Delphix Engines on GCP. For more information, see the Terraform Provider release notes on Ecosystem Hub.