Compute and storage capacity

ArcGIS Analytics for IoT allows you to create, share, and manage real-time and big data ingestion and analysis tasks, which utilize compute capacity in the Analytics for IoT system. Analytic items can also write output to feature layers which are stored in the Analytics for IoT spatiotemporal big data store.

Compute limits

An Analytics for IoT subscription is a real-time and big data processing system with a given amount of compute. This paradigm enables predictable subscription costs for users.

Each Analytics for IoT item consumes compute resources while running, and this consumption varies by data characteristics, velocity, and volume. For example:

  • Feed types capable of ingesting tens of thousands of observations per second will consume more compute capacity than feed types that support lower velocities.
  • Real-time analytics support auto-scaling, which means that if their initial compute profile does not accommodate the incoming velocity or volume of data, the item will be incrementally scaled until the workflow can be sustained. Real-time analytics processing higher velocity data will therefore consume more compute capacity than analytics processing lower velocity data streams.
  • Big data analytics can be configured with different run settings based on the volume of data, which will run the analytic with different levels of compute.

The amount of compute capacity available in the Analytics for IoT subscription can be augmented with additional IoT Compute Units. An IoT Compute Unit provides additional compute capacity to accommodate more use cases, where each use case can consist of multiple feeds, real-time analytics, and big data analytics. When compute reaches 75% capacity, an alert will be sent to the organization administrator so they can determine if add-on IoT Compute Units need to be purchased.

Administrators can view the percentage of compute utilized on the Home page in the Analytics for IoT application. Additionally, the Compute Utilization page reports the percentage of compute utilized by feeds, real-time analytics, and big data analytics, as well as the percent compute utilized by each user item. For details, see Manage and monitor compute and storage.

Data storage limits

An Analytics for IoT subscription includes a given amount of feature storage for IoT data and analytic results. This feature storage can support hundreds of millions of records depending on the schema of the data (number of attributes) and the complexity of the geometry.

The amount of feature storage can be augmented with additional IoT Storage Units. When feature storage is at 50% capacity, an alert will be sent to the organization administrator, so they can determine if add-on IoT Storage Units need to be purchased.

Administrators can view the percentage of storage being utilized on the Home page in the Analytics for IoT application. Additionally, the Storage Utilization page reports the percentage of storage utilized by different system components, as well as the percent storage utilized by each feature layer. For details, see Manage and monitor compute and storage.