I’m excited¬†to announce our new Edge Technical Preview.

Why Edge?

  • Data volume: the volume of data being produced by sensors and IoT devices is growing at a huge rate – transmitting all that data to the cloud for analysis is not practical
  • Rapid reaction times: for many scenarios especially those that are¬†safety critical, the round trip latency between the sensor and cloud takes too long. ¬†The speed of light becomes a limiting factor – so an immediate decision and action needs to be taken at the Edge.
  • Data Privacy and Security: ¬†Every time you move data, you increase the attack surface, giving more opportunity for data to be exposed. ¬†There may also be regulatory restrictions on where data is allowed to reside. ¬†For these cases processing the data at the Edge is a must. ¬†Data gravity is driving the computation closer to the data.

Edge devices are becoming more powerful

  • Even a few years ago edge nodes and gateways would only have a limited processing capability, with single low power cores and a small amount of memory. ¬†Now Edge nodes are capable of significant processing.

Offline operation

There are scenarios which are intermittently connected, or even long term disconnected – with limited or expensive connectivity over satellite links.

  • Actions must continue to take place when offline
  • Updates and synchronisation should occur when connectivity becomes available.

Watson IoT Platform Edge

The Watson IoT Platform Edge Technical Preview offers some key capabilities for the Edge:

  • Automated deployment of workloads and services to the Edge
    • After loading a small bootstrap agent to the Edge node, all other services and workloads are deployed automatically
    • Makes it easy to manage large infrastructures of Edge nodes
    • Makes it easy to maintain and update software on Edge nodes
    • Deploy custom, IBM and third party services to the Edge
  • IoT Core Edge service
    • Local MQTT broker with routing options
      • Communicate between edge components
      • Choose how messages are routed between Edge and Cloud
      • Store and Forward
    • Filtering at the edge

Workloads and services are defined and distributed using docker containers.

Edge nodes can vary in size enormously Рfrom a server farm consuming large amounts of data from an oil rig or wind farm, to a single core Edge node monitoring temperature.  At the small end the solution supports a single core ARM or Intel architecture system running Ubuntu Linux with a minimum of 256MB memory.

Want to get started?

Take a look at the documentation and sample here: https://console.bluemix.net/docs/services/IoT/edge/WIoTP_edge.html#watson-iot-platform-edge-components

This is a technical preview and we’re interested in your feedback. ¬†Please let us know what you think.

Join The Discussion

Your email address will not be published. Required fields are marked *