Skip to main content

The Decentriq platform

The enterprise SaaS platform for secure data collaboration

Decentriq is a cloud-based enterprise SaaS platform offering data clean rooms that enable its users to work with and collaborate on data assets with minimal risk. As the first data collaboration platform where users do not have to trust each other, the platform operator, or the cloud provider, Decentriq mitigates the risks of collaborating on sensitive datasets and helps organizations unlock the full potential of their data.

Data can be used for collaboration without being shared, data can only be used for a specific limited purpose, and compliance can be remotely verified.

Data clean room collaboration

Decentriq enables users to collaborate on data with the guarantee that their contributed data sets can only be used for the explicitly specified intended purpose. The platform uses the metaphor of a data clean room, a highly controlled environment free from outside influence, to describe a specific act of collaboration.

Basic workflow

Each data collaboration is configured in a data clean room. By definition, a data collaboration includes at least two participants, each of which can take on one or multiple roles as DCR Creator, Analyst or Data Owner. The standard workflow follows these steps:

  1. The DCR Creator creates a new data clean room and defines the terms of the collaboration:
    • Table schemas and file definitions - what data will be included and how it will be structured.
    • Computations - what operations will be performed on the data (e.g. SQL queries, Python code, etc).
      Learn about each of these in the Computations section.
    • Participants and permissions - who is responsible for contributing data (Data Owner) and who can see the results (Data Analyst).
  2. The DCR Creator publishes the data clean room, which means:
    • These parameters are tied together into a verifiable format (DCR configuration) which allows each participant to review and understand the scope of collaboration.
    • The Decentriq platform uses cryptographic protocols to indelibly bind these parameters to the code that executes them.
    • This is the "contract" for the collaboration – readable by humans, but enforceable by the platform.
    • If the interactivity feature enabled, participants can request changes based on a request approval flow.
  3. The participants collaborate in the data clean room:
    • The Data Owners can provision / delete their datasets.
    • The Analysts can run computations and retrieve results.
    • All participants can view the data status and the audit log.

Additionally, every data clean room configuration contains the root certificate authority used for user authentication as well as attestation specifications - the technical details required to ensure that the configuration is unambiguous and complete, including the verification policy for the hardware.

For more details about how Decentriq achieves its security guarantees, please refer to our technical whitepaper.

The Data computation graph

The Decentriq platform supports multi-stage data pipelines. While defining computations in a data clean room, users can specify the relationship between computations and which other participants can read the output.

Compute graph

Computations can be written in SQL, Python, and R programming languages, and the most popular machine learning, statistics, and visualization libraries and packages are available. The platform also supports arbitrarily structured data such as JSON, images, and trained ML models in Python and R using the standard techniques for those languages.

Learn about each supported compute node in the Computations section.

How to get started

  1. To get started, you need a Decentriq account. If you don't have one yet, please contact - also for any other question, we're always happy to help.

  2. To use the browser-based UI, log in to the Decentriq Platform and follow the Decentriq UI tutorial.

  3. If you want to use Decentriq's Python SDK instead, create an API token in the Decentriq UI and follow the Python SDK tutorial.