IBM OpenWhisk: Function-as-a-Service

OpenWhisk is an open-source serverless compute platform. It executes your code in response to events from triggers. You only pay for the time your code actually runs (rounded up to the nearest 100ms), which means no excess capacity or idle time. OpenWhisk scales to fit your exact demand, from once a day to thousands of parallel requests per second. Action instances scale to meet demand exactly, then disappear. It’s perfect to handle volatile workloads that can be split into smaller pieces and executed in parallel. Because OpenWhisk is open-source and is built around an open ecosystem of event providers, there is never a need to worry about vendor lock-in. The IBM Openwhisk block allows OpenWhisk to work with today’s stream-based applications; stream data to and from OpenWhisk and leverage PubNub BLOCKS for lightweight compute in the network.

Overview

The OpenWhisk block can be used for a number of serverless to serverless use-cases. Here are a couple examples:

Vehicle Dispatch (PubNub Blocks → OpenWhisk)

PubNub Blocks filters geo data and routes to OpenWhisk for computation. OpenWhisk runs business logic and machine learning, and broadcasts back to PubNub listeners on end-devices.

Disaster Alerts (OpenWhisk → PubNub)

Openwhisk detects weather anomalies and broadcasts to PubNub listeners on end-devices for real-time disaster alerts.

Social Chat App (PubNub → Watson + OpenWhisk)

PubNub filters chat messages through a PubNub Watson Translator block, and offloads audience analysis to OpenWhisk for async post-processing.

Input