IBM Poised To Compete on ‘Number of Levels’ in Serverless Space; Adds Updates to Bluemix OpenWhisk

The serverless space is enjoying growing attention and investments from mainline enterprise companies. IBM is the latest to add new features and extensions to its entry, dubbed IBM Bluemix OpenWhisk. IDN speaks with IBM Distinguished Engineer, Michael Behrendt.

Tags: analytics, Apache, APIs, Bluemix, cloud, IBM, Lambda, microservices, OpenWhisk, serverless. Watson,

Michael Behrendt, IBM
Michael Behrendt
IBM distinguished engineer

"The latest version of OpenWhisk helps developers to more easily embed cognitive intelligence, cloud data services, and IoT sensor data within apps."

Application Architecture Summit
Modern Application Development for Digital Business Success
Online Conference

The serverless space is enjoying growing attention and investments from mainline enterprise companies.


IBM is the latest to add new attributes and extensions to its entry, dubbed IBM Bluemix OpenWhisk. Big Blue’s serverless space boasts openness, portability, extensibility (via API and cloud) and multi-language support.  


It’s a package that IBM hopes will attract developers and IT for the nascent serverless sector, drawing attention away in particular from Amazon Lambda and open source projects.


“We’ve built OpenWhisk to be able to compete in the growing serverless market on a number of levels,” IBM Distinguished Engineer, Michael Behrendt, told IDN.


IBM’s OpenWhisk serverless architecture aims to accelerate and simplify app development by abstracting away infrastructure. Its serverless architecture promotes using quick, adaptable and scalable action sequences to meet the fluid demands of app performance, functionality and user experience, according to Behrendt.


IBM’s latest OpenWhisk extensions leverage Big Blue’s platform approach to provide developers easy ways to tie their serverless projects in with IBM’s portfolio extensibility technologies.  Notably, these include API Gateway, cognitive intelligence (i.e. Watson services) and more, Behrendt said. All these connections can be made automatically and on-demand -- without the need to manage and configure infrastructure, he added.


“Traditionally, developers that use serverless must configure and secure external endpoints manually, which takes a lot of time. The latest evolution of IBM’s serverless offering does this for developers automatically, helping them to more easily embed cognitive intelligence, cloud data services, and IoT sensor data within apps,” he added. The latest updates point to IBM’s intent to provide developers “choice and flexibility. . .  around how they pull in the outside tools and data they need to create apps,” he added.


Under the covers, OpenWhisk uses business rules to bind events, triggers, and small, distinct, and independent actions to each other. These OpenWhisk actions are built to run automatically only when needed. This approach lets developers work on different pieces of code simultaneously, leading to more rapid design-build-test-launch app lifecycles.


A brief review of IBM Bluemix OpenWhisk’s other capabilities reveals:

  • A broad range of connections to services to let developers differentiate their work product, including IBM’s API Gateway, IBM’s cognitive cloud services (Watson APIs and the Watson IoT Platform, for example).
  • An expanding ecosystem of partner tools, such as Kong’s open API connector and PubNub’s data stream network for real-time applications, he added.
  • Intention to support multiple languages and runtimes, including Node.js, Swift, Python and even arbitrary binary programs encapsulated in Docker containers.

IBM’s Bluemix OpenWhisk Focuses on Simplifying APIs, Microservices

Open Whisk

“We have now made OpenWhisk easier to use securely and in concert with outside code, data and systems, readying it for the enterprise,” said Jason McGee, VP and CTO for IBM Cloud Platform in a statement. Behrendt walked IDN through some use cases that demonstrate the value of this.


“An important use case for serverless is the ability to build microservices or APIs,” Behrendt said. OpenWhisk now makes it easier to build serverless APIs.

In specific, “this expansion of OpenWhisk allows developers to expose serverless actions as secure and controlled APIs on the cloud, making it easier and quicker to build apps such as IoT and cognitive solution,” he explained to IDN.


OpenWhisk’s new integration with API Gateway is key to this capability – as it enables developers to securely create and bridge external endpoints from multiple systems into serverless functions. The API Gateway on IBM Bluemix OpenWhisk acts as “gatekeeper” between an external source and a corresponding OpenWhisk action, Behrendt explained.  


“We’ve added an API Gateway so that developers can map to OpenWhisk actions and then have a running serverless API on the backend. It helps with getting started, as they can create an API in just a few minutes, and makes it easy for developers to experiment with the system with full control over HTTP traffic in and out,” he said.


As an example, he said think of the external source as being an uploaded image and a corresponding OpenWhisk action occurs when the Watson Visual Recognition service (connected via API) can identify and tag that image.


OpenWhisk also provides usage analytics for developers to help them better understand when and why different services are being used and how users are engaging with their apps. It’s all designed to provide deeper insight into how to improve the overall user experience, Behrendt said.

Beyond all these discrete features, IBM OpenWhisk supports the full application lifecycle -- design, test and run which lets developers contribute on their own, and tailor OpenWhisk to their individual needs,” he added.


To set the stage for further flexibility and extensibility, OpenWhisk is designed as an open platform to give developers greater flexibility and avoid vendor lock-in. “For example, if a developer chooses to run OpenWhisk somewhere else on a different cloud, or their own infrastructure, they can. It’s the same code as if they’re running in Bluemix, but there’s freedom of choice,” he said.


For IBM, OpenWhisk’s openness offers another key benefit.


“To us, ‘open’ also means that the code is not controlled by a single vendor. By having our code in Apache, it resides in a vendor-neutral foundation. Our competitors may have their code in the open, but not owned by a foundation,” he said. He shared a very real-world example open source developers are quite familiar with. “If a customer finds a bug in code that wasn’t working as expected, they can contribute to the open source code to fix it. With that, everyone can benefit from the change they made to our technology.”


IBM Bluemix OpenWhisk is based on the open source Apache OpenWhisk project. As its name implies, it runs on the IBM Bluemix cloud platform.