Spaces

Spaces are secure, on-demand environments that bring together people, tools, data, and models to support safe collaboration and the development of data products and services. They are designed for teams working within and across organizations, enabling users to explore, analyze, and work with data in a governed, cloud-based workspace.

Each participant in a Space has access to their own isolated compute environment, while still being able to share data securely with others. Spaces can also execute background tasks, supporting automation and continuous data processing - ideal for operational analytics, machine learning workflows, or the development of reusable data assets.

What Are Spaces?

A Space is a configurable environment where users can work collaboratively with data using a set of predefined tools and compute resources. Within a Space, users can connect to data products and assets they have access to, write code, create visualizations, and produce new outputs — all within a governed, auditable framework.

Spaces are purpose-built for collaboration. They allow teams to safely explore and transform data, test hypotheses, and share results without compromising data security or governance.

Key Capabilities

• Provision on-demand compute resources for individual users.

• Enable secure collaboration and controlled data sharing.

• Support multiple tools for querying, transforming, and visualizing data.

• Allow background task execution for automated or scheduled workflows.

Space Types

The experience of using a Space depends on the permissions held by its users. Two main configurations are available:

VDI-less

The VDI-less, or Workbench experience is the default for users who hold full access and export permissions to all products and assets within a Space. It provides a fast, flexible way to interact with data directly through the browser, without needing a secure desktop.

• Tools can be launched directly from the Space page.

• Users can download data and connect to the internet.

• Best suited for advanced users performing deep analysis or product development.

• Typical activation time is between 6–20 minutes on AWS Data plane and 5-10 minutes on a Databricks workspace, depending on resource availability. This may also fluctuate depending on your environment’s configuration

VDI

The VDI, or Sandbox experience is used when a user has limited or view-only permissions for one or more data products or assets in a Space. It offers the highest level of protection and is designed for working with sensitive data.

• Users access the Space through a secure Virtual Desktop environment such as AWS Workspaces or Itopia.

• Required when the user holds Spaces-only permissions for any asset or product.

• Setup typically takes around 20 minutes before access is available. This time span may cary depending on your platform.

• Prevents data export and enforces strict data governance policies.

While the VDI experience takes longer to activate, it ensures full compliance and protection for controlled datasets, such as during trials, pre-sales, or regulated industry use cases.

Serverless in Spaces

Serverless compute means the underlying infrastructure that powers Spaces is managed and scaled automatically behind the scenes, rather than having to spin up and maintain a dedicated cluster for it. Compute resources are provisioned the moment they are needed and released the moment they are not.

Spaces can run on serverless compute instead of a dedicated cluster. This means your Space starts in a couple of minutes rather than five to ten, with no change to how you work once it's running.

• Faster start-up. Because there's no dedicated cluster to provision, activating a Space is much faster.

• No cost for idle time. With serverless compute, you're only billed for the time your queries and notebooks actually run - not for the cluster sitting idle while you're not using it.

• No change to your workflow. Notebooks, SQL Lab and Visualisations all work exactly as they do today. You won't notice any difference in how you use the Space day to day - only that it's ready sooner.

• Improved scalability. Serverless compute draws on a much larger shared pool of resources, so it's less prone to the delays that can happen when many Spaces need compute at once.

Tools

Each Space provides access to a suite of tools that support data access, analysis, transformation, and visualization. The exact tools available depend on the configuration defined by the Platform Operator.

These tools allow users to perform a wide range of tasks - from exploratory data analysis to creating dashboards and exporting derived outputs.

Creating and Activating Spaces

You can create a new Space directly from the Spaces app. Spaces can include any number of assets, products, and collaborators, and you can reorder or edit these configurations at any time. For more information on the process of creating Spaces, see the dedicated guidance page.

To create a Space:

1. Click the Spaces app on the navigation bar.

2. Click Create Space.

3. Define the Space specification by selecting:

   • The Products and Assets to include.

   • The Tools and machine configuration (e.g., node type and number of workers).

   • The Collaborators who should have access. (Only users visible to your organization can be added.)

4. Click Save to finalize the specification.

Each Space specification defines the compute configuration and data access setup. A user can create up to 15 Spaces, and you can modify an existing specification at any time. This number can be adapted based on your environment through liaison with Harbr Support

To activate a Space:

1. In the Spaces app, select your Space specification.

2. Click the Activate button on the right-hand side.

3. The platform provisions the compute environment and loads the specified tools and data.

4. Once active, tools change from grey to color and their status updates from Waiting to Launch.

5. Click a tool to open it. You may be prompted to enter login credentials before use.

Each activation spins up a new compute cluster, but database tables and results persist between sessions.

Data Access and Databases

Data access in a Space is governed by user permissions and the Space configuration. Users can connect to data through standard database interfaces, direct file access, or predefined platform connectors. Each Space includes several database environments:

This setup ensures that data movement is controlled, traceable, and aligned with platform-level governance and lineage rules.

Collaborators and Permissions

Spaces are collaborative by design, but access is carefully controlled.

• Only collaborators visible to your organization can be invited.

• Collaborators must have an active Spaces subscription to every asset and product included in the Space.

• If a collaborator is missing a required subscription, the Space will appear locked for them until they subscribe via the Exchange.

• Subscription templates and lineage rules determine whether derived assets require subscriptions to their source products.

This approach balances self-service collaboration with strong data governance and traceability.

Subscription Level Enforcement

Access to data within a Space is governed by the same subscription rules that apply across the platform, ensuring consistent enforcement of permissions and restrictions.

When an asset is accessed via a subscription with Row and Column, restrictions are enforced automatically and the asset cannot be separated from its active product context. 

A user can add a standalone asset to a Space only if one of the following conditions is met:

1. The user has direct Share permission for the asset

2. The user has an active subscription to a product that includes the asset without any row-level or column-level restrictions

3. Assets that are outside of the product cannot be added to the space.

Spaces do not introduce new access rules. Instead, they inherit permissions from subscriptions and direct asset shares, applying restrictions consistently to prevent the overexposure of data.

Derived Assets and Tasks

Spaces can produce outputs that become new data assets or products. These can take two main forms:

• Static artifacts, such as datasets, reports, or model definitions.

• Automated processes, created by converting code written in a Space into a Code Asset, which can then be used in a Task.

When lineage is enforced, users must have active subscriptions to all contributing products to access these derived outputs. If lineage is not enforced, assets created from other assets (but not products) can be accessed without additional subscriptions.

This capability allows teams to turn their analyses into governed, reusable outputs while maintaining clear traceability to source data.

Best Practices

Spaces make it possible to bring together people, data, and tools in one governed environment. They enable teams to:

• Safely explore, analyze, and visualize data.

• Build and share reproducible outputs.

• Automate processes and data transformations.

• Maintain full control over data lineage and permissions.

• Transform insights into governed, reusable data products.

To ensure efficient and governed use of Spaces, we recommend:

• Suspend Spaces when not in use to reduce resource costs.

• Select the correct specification for your workload and objectives.

• Stage intermediate results in the Collaborative Database to optimize performance.

• Monitor lineage to maintain transparency across derived assets.

• Check permissions regularly to ensure correct access control.

• Use private repositories - mirrors of PyPI and CRAN are preconfigured for security.

• Create purpose-built Spaces for specific projects to keep lineage clean.

• Persist valuable outputs as governed assets when work is complete.

• Delete old or inactive Spaces to avoid redundancy.

• Reuse artifacts such as dashboards or notebooks across Spaces or products where possible.