Intrinsic data framework

Intrinsic data framework enables third-party users to access historical data for the purpose of debugging, analysis and much more. Intrinsic has different pipelines for handling different types which each have unique characteristics. The concepts below lay out key concepts that help build a mental model about the vision around the data framework.

Retention model and access points

Generally speaking, all data pipelines are designed to work across on-premise (IPCs) and the cloud. Below are the pros and cons of both approaches.

On premise access

Pros

• Lower latency access without needing the internet
• Ease of integration with factory floor systems

Cons

• Low retention as we have constraints on available storage
• Remote access is harder

Cloud access

Pros

• Higher retention periods
• Ease of remote access
• Leveraging cloud compute for creating web based analysis experiences
• Access Control List based sharing

Cons

• Limited on internet availability for uploading from workcells
• More involved steps for Air-gapped operations

Data pipelines

Data pipeline	Description	Cloud enabled	Producers	Data Access Points
Structured logging	Protobuf based timeseries data. It is a collection multiple data streams, each of a different description and type. These data streams are called event sources	Yes. Data gets selectively uploaded to cloud services	Any compute component running on a workcell	Cloud and Onprem APIs to read data per event source.
Text logging	printf style strings emitted by containers running on workcells	Yes. Data gets selectively uploaded to cloud services	All processes running on a workcell	Onprem APIs to read text logs by skill, resource, hardware module. Cloud APIs are not yet available.
Pubsub	Protobuf based messaging data publishing and subscription framework	Not yet. Currently sandboxed to a workcell. Cloud based cross workcell communication not yet available	All processes running on a workcell	Per topic Publish and Subscribe APIs are available
Key Value store	Protobuf based storage service available to persist data on a workcell	Not yet	All processes running on a workcell	Set, Get, List APIs are available per key on workcells