Documentation

  • N/A
  • 2.6
  • 09/10/2021
  • Public Content

How it Works

Edge Insights for Industrial
is a set of pre-integrated ingredients designed to accelerate the development and deployment of solutions for the industrial sector. The package is meant to be deployed on the device closest to the data generation, such as the tool or machine assembling a product. This enables the ingestion of video and time-series data, storage of data, performing analytics, closing the loop by transmitting a control message, and publishing the results.
Edge Insights for Industrial
is based on Open Edge Insights.
The Edge Insights for Industrial User Guide (EII User Guide) details the capabilities of EII and serves as a reference for developers. It is accessible at: https://open-edge-insights.github.io/.
This section provides an overview of the modules and services featured with
Edge Insights for Industrial
.

Modules and Services

For an overview of the EII modules, see below.

Video Ingestion and Analytics

Video Analytics
Classifier algorithms are executed in the Video Analytics container. The container uses the
Edge Insights for Industrial
Data Bus to subscribe to the input stream published by the Video Ingestion container. The algorithms included in the container are optimized for Intel hardware using the
Intel® Distribution of OpenVINO™ toolkit
(
OpenVINO™
and the
Intel® Math Kernel Library
. You can create multiple instances of the Video Analytics container to classify multiple video input sources.
  • The is a software development kit (SDK) for deploying deep learning computer vision inference applications. It optimizes inferencing on your edge IoT device for Intel® architecture.
  • The is a library of highly-optimized, threaded, and vectorized math routines designed to maximize performance on Intel® processors.
Watch a video on Video Analytics on the Edge Insights for Industrial Architecture & Technical Training page.
Video Ingestion
Edge Insights for Industrial
supports ingestion from:
  • Video files.
  • GigE cameras.
  • USB
    cameras.
  • RTSP
    cameras.
Edge Insights for Industrial
uses and OpenCV to provide a preconfigured ingestion pipeline that you can modify for different camera types and preprocessing algorithms. The Video Ingestion container publishes video data, consisting of metadata and frames, to the
Edge Insights for Industrial
Data Bus. You can create multiple instances of the Video Ingestion container to ingest video from multiple input sources. In addition, the Video Ingestion container itself can run UDFs performing analytics. This is recommended for faster processing of frames to avoid transmission delays.
Watch a video on Video Ingestion on the Edge Insights for Industrial Architecture & Technical Training page.
Training and Learning Suite (TLS)
The Training and Learning Suite has a web-based user interface for training deep learning models. The key TLS features are:
  • Remote deployment of a TLS-trained model into an
    Edge Insights for Industrial
    system.
  • Visualization of Edge Insights for Industrial video data classification results.
For details, see the Training and Learning Suite (TLS) section.
Image Store
The image store provides storage and retrieval of images as binary blobs. You can store the images persistently by using: In persistent storage with , an open source, distributed, object storage server.
Watch a video on Image Store on the Edge Insights for Industrial Architecture & Technical Training page.

Time Series Analytics

Edge Insights for Industrial
leverages
Telegraf*
for time-series data ingestion into
InfluxDB*
.
Kapacitor*
reads data from
InfluxDB*
, generates alerts, and takes the appropriate action.
  • is an agent for collecting, processing, aggregating and writing performance metrics.
  • is a distributed datastore for metrics, events, and real-time analytics.
  • is a data processing engine that handles both stream and batch data from InfluxDB, acting on this data in real time via its programming language, TICKscript.
Watch a video on Time Series Data on the Edge Insights for Industrial Architecture & Technical Training page.

Security and Configuration

The Data Security Framework enables security through a two-stage process involving provisioning and execution on Edge Compute Node. Its primary objective is to prevent unauthorized access of generated data from within the system or via external network interfaces.
Edge Insights for Industrial
uses for configuration management.
etcd*
is a consistent, distributed key-value store that provides a reliable way to store data that needs to be accessed by a distributed system or cluster of machines.

Manageability

Edge Insights for Industrial
uses , Azure* and ThingsBoard* to support over-the-air updates and to enable cloud-to-edge manageability of application services.
For details, see the Manageability: Updating an Application-Over-The-Air section.

Docker

is a container framework widely used in enterprise environments. It allows applications and their dependencies to be packaged together and run as a self-contained unit.

Edge Insights for Industrial
Data Bus

Edge Insights for Industrial
Data Bus is an abstraction over
ZeroMQ*
, which is used for all inter-container communication.
is a brokerless message bus that transfers data from the source directly to the destination.
ZeroMQ*
is used in TCP and IPC mode with pub-sub and request-response patterns.

Product and Performance Information

1

Performance varies by use, configuration and other factors. Learn more at www.Intel.com/PerformanceIndex.