GenRocket Inc.

Built around synthetic data rather than copies of live data, GenRocket sells a test data platform to software quality teams, which use it to generate the data their tests need. The platform is designed to replace static, copied production data with data that adapts to the application under test.

Test data designed from rules rather than copied

Instead of copying records from a live system, the platform designs data from the rules, logic and scenarios a tester defines. The company calls this design-driven synthetic data: the structure and the scenarios are specified first, and the data is generated to fit them. The firm argues that copied production data only shows what has already happened, so it cannot cover every workflow, edge case and exception a release has to handle.

Data for complex workflows and transaction formats

The platform builds data for complex workflows and for transaction formats such as X12 EDI, the standard used to exchange business documents between trading partners. Volumes run from small test sets to very large ones, and the company says it can generate millions to billions of rows in minutes.

On-demand provisioning for testers

Self-service provisioning delivers fresh, controlled data in real time, in whatever volume, variety or format a team asks for. A self-service portal, which the company calls G-Portal, lets teams request the data they need themselves instead of waiting on a central data group.

Connecting to automation tools and release pipelines

The data is designed to work with test automation tools, frameworks and virtual test environments, and it can feed continuous integration and delivery (CI/CD) pipelines. The company says the platform can be rolled out across an enterprise with centralized data modeling and distributed self-service, so that several teams can share one set of data models while each draws the data it needs.

Data masking, subsetting and profiling

Alongside synthetic generation the platform offers data masking, data subsetting and data profiling. For subsetting, the company says sensitive personal data is never looked at or stored, because it is replaced synthetically. Subsets and synthetic sets can be run in parallel across one test environment or many, as needed. The platform includes sixteen methods for automating data modeling from metadata and integrates with third-party profiling tools that scan entire data sets.

Database support and delivery rates

The standard functionality covers six SQL databases: Oracle, DB2, MS SQL, MySQL, PostgreSQL and Sybase. The company reports delivery of 2.5 million rows of data a minute, and 100 million rows in 30 minutes.

Patent and reported coverage gains

The company holds US patent number 9,552,266 B2, which covers its method of generating test data. It reports coverage increases in the examples on its site, such as system integration and performance testing rising from 30 percent to 80 percent, and regression and API performance coverage rising from zero to 70 percent.

Enterprise adoption across large companies

More than fifty Forbes Global 2000 companies, by the company's own count, use the platform, and global systems integrators recommend it. Those figures describe the size of the organizations that have adopted the approach, and they are the company's own reported numbers rather than independent audits.