Introduction to Streams development for edge analytics

Edit me

IBM® Streams can be used to develop and build applications for edge analytics using IBM Edge Analytics Beta on Cloud Pak for Data.

For more information about Edge Analytics, see the IBM Edge Analytics Beta on Cloud Pak for Data documentation.

Getting Started

To get started creating a Streams application for edge analytics, view one or more Streams edge samples:

  • Basic Edge Samples - simple, hello-world like application that you can deploy to the edge
    • Edge Rolling Average sample - A Python notebook that computes the rolling average of a set of sensor values on an edge system.
    • Edge Application Control sample - A Streams Processing Language (SPL) application that performs some basic analytics on the edge. This sample also illustrates how to pass parameters and set runtime options for edge applications.
  • Multi-tier Samples - an application that runs on edge systems, and another application that runs on the Cloud Pak for Data control plane (CP4D Hub).
    • Edge Digit-Recognition notebook sample - An edge application written in a Python notebook uses a machine-learning model to identify hand-drawn integers. Results from the model scoring at the edge are sent to another notebook running in the CP4D Hub.
    • Edge Digit-Recognition Streams Flows sample - Similar to the Edge Digit Recognition sample except that the edge application is constructed using the Streams Flows visual builder.
    • Edge Network Analysis sample - The Netflow sample collects and aggregates network traffic data. It is written in Streams Processing Language (SPL) and can be deployed in many different configurations (on-premise, cloud, edge, etc.) Instructions have been added explaining how to build the sample using VS Code and deploy the collection and aggregation portion of the sample to edge systems.

For more information about develop and building edge applications using IBM Streams, see Moving analytics to the edge with Edge Analytics.

Transferring data from the edge

One way to communicate data from edge applications to a central cloud platform, service, or application is through a data feed or messaging system like Apache Kafka. See Apache Kafka options for edge applications for a list of a few possible Kafka installation options and how to interface with those options in your edge applications using Streams toolkits.