File-Based Delivery

This page describes how transactional data can be delivered to Fyre using file-based methods.

File-based delivery is typically used when partners do not integrate using the standardized API and already generate transactional exports in an existing format. In these cases, Fyre adapts to the partner’s current export structure and delivery process.

While file-based delivery is fully supported, API-based delivery using standardized formats remains the recommended long-term approach whenever possible.

When file-based delivery is used

File-based delivery is most commonly used when transactional data is already exported on a schedule, existing exports are produced in non-standard formats, or introducing a new API-based flow is not feasible in the short term. It is also well suited for transferring large historical datasets efficiently.

In these scenarios, the goal is to reuse what already exists and minimize disruption to partner systems.

How file-based delivery works

Regardless of how files are delivered, the underlying model is the same.

Partners generate transactional data exports and make them available to Fyre through an agreed mechanism. Files are stored as received and processed asynchronously at a later stage as part of the ingestion pipeline.

Validation, enrichment, and aggregation do not occur at the time of delivery. This approach allows Fyre to handle large volumes of data efficiently and apply consistent processing rules across all partners.

Delivering files to Fyre

There are multiple ways to make files available to Fyre. The exact setup is agreed during onboarding, based on partner infrastructure and operational preferences.

Uploading files to Fyre-managed storage

Partners may upload export files directly into a dedicated storage bucket managed by Fyre.

Each partner is assigned a dedicated bucket or bucket path that is used exclusively for their integration. Fyre provides a service account with permissions scoped specifically to that location. These permissions are limited to managing files within the assigned bucket and do not grant access outside of it.

The structure of the exported files is flexible and can follow the partner’s existing export logic. Files may be organized by location, by date, grouped into time windows, or structured in any other consistent way that already exists in the partner’s systems. As long as the required data fields are present and identifiable, Fyre adapts to the provided structure during onboarding.

Partners can upload files using standard tools and SDKs provided by Google Cloud. Refer to the Google Cloud Storage upload documentation for implementation details and examples.

Retrieving files from partner-managed storage

In some cases, transactional exports already exist in the partner’s own cloud storage environment. Rather than re-uploading data, Fyre can retrieve these files directly from the source storage.

This is implemented using the Storage Transfer Service, which enables secure and scheduled transfers from partner-managed storage into Fyre-managed storage. The service is designed for large-scale data movement and works well when exports are already stable, read-only access is preferred, or large historical datasets need to be transferred.

The Storage Transfer Service supports multiple cloud providers, including:

• Amazon Web Services

• Microsoft Azure

• Google Cloud

This approach is secure, auditable, and does not require partners to change their existing export pipelines.

Implementation details and supported configurations can be found in the Storage Transfer Service documentation.

Supported file formats

Fyre supports common structured file formats, including but not limited to:

• CSV

• JSON

• Parquet

• ndJSON

Existing export formats can usually be reused. File-based delivery does not require a strict predefined schema, as long as required information can be clearly identified and mapped during onboarding.

Security considerations

All file-based integrations follow strict access-control principles.

Access is limited to dedicated buckets or paths, permissions follow the principle of least privilege, and credentials are scoped to the minimum access required for the agreed delivery method. Data is transferred and stored using secure, encrypted channels.

All access details are reviewed and agreed during onboarding.

Choosing file-based delivery

File-based delivery is a practical and reliable option when standardized API-based integration is not feasible.

In most cases, file-based delivery relies on existing export formats and structures that are specific to the partner’s systems. Fyre adapts to these formats during onboarding so that integration can proceed without requiring changes on the partner side.

File-based delivery and API-based delivery should be considered distinct integration models. The selected model is agreed during onboarding and reflects the partner’s technical capabilities and operational constraints.

Long-term considerations

File-based delivery is often used as a stable, long-term solution.

While API-based delivery using standardized formats is the recommended approach where possible, file-based integrations are fully supported and can operate reliably over time when properly aligned.

Any changes to the delivery model are discussed and agreed explicitly as part of the ongoing partnership.