Pivotal hopes to burnish its image as a cloud-agnostic PaaS provider with the introduction of Pivotal Function...
Service, a serverless computing runtime that it said works with any combination of public and private clouds.
The serverless computing model supports event-driven application architectures, which are more common thanks to IoT and app notification streams. AWS set the pace for serverless computing in 2014 with AWS Lambda, its function-as-a-service offering, and other cloud providers have followed suit with similar services. Now, Pivotal Function Service targets companies with multi-cloud strategies that want a unified way to conduct serverless operations.
The company bases Pivotal Function Service on Knative, the open source middleware project for Kubernetes container-based applications that emerged earlier this year and quickly developed an ecosystem around Pivotal, Red Hat, SAP, IBM and others. The service is currently in an early stage trial.
On top of Knative, Pivotal built its open source Project riff to add developer and operational tools and improve installation packages. Pivotal also includes buildpacks, which collect related functions into executable artifacts, so developers need only worry about their business logic, according to Onsi Fakhouri, Pivotal's SVP of cloud R&D, in a blog post.
Pivotal Function Service will feel "truly serverless" to users, Fakhouri wrote. For example, it not only deploys function source code, but also tracks whether a previous version of the function exists and preserves it in case there's a need to roll things back. It invokes functions in response to an event, whether from an outside request or in response to the completion of another function. Fakhouri gave the example of a function that corrects a customer's postal code and then triggers another function that stores the mailing address.
Serverless is of keen interest -- and likely disruptive -- to enterprise IT shops, though Pivotal's move into serverless primarily helps those organizations already committed to Pivotal's ecosystem.
Ryan Marshserverless and DevOps consultant
"This is where the rubber will meet the road for enterprise IT and public cloud," said Ryan Marsh, a serverless and DevOps trainer and consultant based in Houston, who also works as chief evangelist for Xolv.io, a software quality and testing vendor. "It's one thing to deploy containers where you configure, support and control the full development stack. If I'm making a Java Spring app on Postgres, deploying that to a container is easy and with Pivotal Cloud Foundry and Pivotal Container Service, I could be forgiven for assuming what I have is equivalent to the public cloud, just in my control."
While Pivotal Function Service aims at multi-cloud installations, including on-premises, private cloud environments, serverless may be best suited for the public cloud.
"On-prem cloud was cute when what the cloud meant to enterprise IT was faster provisioning," Marsh said. "Now, cloud and, specifically, serverless -- which is much more than functions -- means a wealth of incredible products and services, run by the best ops teams in the world, available on a per-invocation billing. Your enterprise IT department cannot keep up with that. It's foolish to even think about it."