Earthscale Architecture Overview

Earthscale provides software to efficiently serve and visualize geospatial datasets from cloud storage.

Earthscale’s architecture consists of the following services:

• Earthscale Explorer: a web application to visualize geospatial datasets registered with Earthscale.

• Metadata Service: allows users to register datasets with Earthscale, search for datasets, and query dataset metadata.

• Data Query Service: efficiently reads and serves data from registered geospatial datasets.

• Cache Service: manages a smart dataset cache to optimize serving latency. Writes to the Earthscale Cache object store.

The Earthscale Explorer and Metadata Service form the Earthscale control plane. The Data Query Service, Cache Service, and the Cache form the data plane.

First, users submit requests to the Metadata Service to register a new dataset given its location in cloud storage. The service reads basic metadata like spatial extent.

To serve data in applications, users make API calls to the Data Query service.

There are two common methods for deploying Earthscale within an organization. In both cases, any communication between the customer’s cloud and Earthscale’s cloud is governed by secure short-lived OIDC federated credentials.

Standard Deployment

Here, the control plane and data plane both run in Earthscale’s cloud account. Earthscale’s services read data from the customer’s cloud, and optionally cache this data in the Cache object store to reduce serving latency. Earthscale deploys the control and data planes in the same cloud region as the customer to minimize egress across cloud regions.

Bring-Your-Own-Cloud Deployment

In a “bring-your-own-cloud” deployment, the Earthscale data plane runs in the customer’s cloud account. Only metadata leaves the customer's network to register datasets with the Earthscale control plane.