Pivotal launched a Kubernetes-based serverless functions platform that uses components from the Knative project to manage deployment and operation of serverless functions across private and public cloud providers. The move is the latest by a vendor in tapping the container orchestrator to break out of what has been traditionally a siloed approach to serverless deployments.
The Pivotal Function Service (PFS) supports the deployment and operations of database, batch jobs, web APIs, legacy applications, and event-driven functions across cloud platforms. PFS takes function source code and deploys it. It also supersedes any previous versions of that function but keeps that previous version in case of a rollback. Those deployed functions can scale down to zero when not in use to preserve resources and scale up automatically based on need.
PFS uses Cloud Native Buildpacks to consistently and securely package functions. Those buildpacks can detect dependencies and automatically build a function into a running application without any interaction from a developer. The buildpack also uses a layered approach to build those functions that allows for the patching of images without impacting the function itself.
Onsi Fakhouri, senior vice president for cloud research and development at Pivotal, explained that PFS is the first multi-cloud packaging of the Knative project. He noted that Knative was the foundation for the Pivotal-led Project riff open source project that extends Knative with developer and operator tooling.
“Riff simplifies the Knative installation experience and adds key user experience components,” Fakhouri noted in a blog post.
Pivotal was one of the founding members of the Knative project alongside Google, IBM, Red Hat, and SAP. The open source project is based on Kubernetes and provides a set of components that allows for the building and deployment of container-based serverless applications that can be transported between cloud providers. Basically Knative is using the market momentum behind Kubernetes to provide an established platform on which to support serverless deployments that can run across different public clouds.
Knative is still in its infancy, having only recently released the 0.2 version of the platform, but is seen as a platform that could pave the way as an alternative for building serverless on top of Kubernetes.
James Governor, analyst and co-founder at RedMonk, noted in a blog post that, “Knative will almost certainly become the standard plumbing for functions-as-a-service (FaaS) on Kubernetes.”
PFS operates through the Pivotal Cloud Foundry (PCF) platform alongside the Pivotal Application Service (PAS) and Pivotal Container Service (PKS). Pivotal is initially offering installation documentation for PKS, Google Kubernetes Engine (GKE), and some local environments. It will eventually add Azure Kubernetes Service (AKS) and VMware Cloud PKS support.
Despite the name, serverless functions do rely on servers but hides the running of those servers from the developer. A deployed function automatically configures the needed infrastructure resources and then shuts down those resources once the function is complete. This is designed to save money on infrastructure resources as the function uses only what is necessary and also helps speed the DevOps cycle as developers don’t need to deal with infrastructure provisioning.
Most cloud providers today offer a serverless option. However, they are typically tied for use strictly on that cloud provider’s infrastructure and can’t be ported to another cloud platform. Those include Amazon Web Services (AWS) Lambda, Microsoft Azure Functions, and Google Cloud Functions.
The Pivotal platform joins a handful of others launched recently that look to break out of that siloed approach. This is becoming more important as organizations look to port their serverless functions across cloud deployments.
For instance, TriggerMesh last month launched a software-as-a-service (SaaS) serverless management platform that runs on top Knative. It allows developers to manage their serverless and function-as-a-service (FaaS) deployments from the platform to work across different infrastructure environments. This includes the ability to define exactly what runtime is being used by the deployment and the ability to trigger events across different cloud platforms.
Others that have launched similar platforms include Stackery, Platform9, and Serverless Inc.