BACKGROUND IMAGE: iSTOCK/GETTY IMAGES
SEATTLE -- Serverless computing is usually thought of as a developer tool, but some IT pros see the Knative project and imagine an event-driven future for themselves.
Knative, launched by IBM and Google in July 2018, provides a set of middleware components that manage the container back end for serverless functions and link together workloads under an event-driven approach. The project, which forms the basis for the Pivotal Function Service and garnered developer preview support in Red Hat OpenShift, could also serve as a framework for DevOps teams to offer their own on-premises versions of AWS Lambda and support the portability of workloads between different cloud providers' function-as-a-service environments.
"We think this is the future of how we will build cloud-native services," said Knative early adopter Ram Gopinathan, principal technology architect at wireless service provider T-Mobile, which uses Knative to manage a store locator application for consumers.
T-Mobile will also investigate Knative to offer an event-driven application deployment pipeline to its software developers and to automate data uploads to its Google BigQuery analytics service and Elasticsearch indices. Knative could also automatically scale Kubernetes pods and replicas as T-Mobile's data stores grow and scale compute resources to process large files as needed, Gopinathan said in a presentation here at KubeCon this week.
Users see strong potential from an early project
Knative is too advanced for most mainstream enterprises, but IT ops pros here this week are eager to try its event-driven automation for common infrastructure management tasks.
"I can see using it for 'fire and forget' tasks, like expanding a storage volume," said Mark DeNeve, systems engineer for HR services company Paychex in Rochester, N.Y., which uses Red Hat OpenShift. "It could help our day-to-day operations stay ahead of our developers."
DeNeve said he is also interested in the long-term potential to move serverless functions between on-premises and public cloud infrastructures, though the company has not yet moved to a hybrid cloud environment.
Other enterprise DevOps pros expressed interest in the Knative project to help add multi-cloud and multi-region portability to serverless functions.
"We also see the possible use case for internal IT ops to automate workloads among different management teams," said Dale Ragan, principal software design engineer at SAP's Concur Technologies Inc., an expense management SaaS provider based in Bellevue, Wash.
A serverless primordial soup has yet to coalesce
Tom Petrocellianalyst, Amalgam Insights
While Knative holds promise, the market for open source serverless utilities is still highly fragmented and chaotic. And numerous open source projects, from Iron.io -- now Oracle Fn -- to, most recently, GitLab Serverless, have said they can make serverless functions accessible on premises and portable between clouds.
"Right now, the advantage for Knative is that it has the broadest [IT vendor] support," said Tom Petrocelli, analyst at Amalgam Insights in Arlington, Mass. "It also has an association with Kubernetes, which means Kubernetes users don't have to manage an entirely different platform."
Still, the major problem with serverless isn't the many players in the market, which is to be expected in the early stages of a technology's development, Petrocelli said. Rather, it's the confusion that lingers over the basic definition of serverless computing.
"I can ask 10 people what serverless is and get 11 different answers," Petrocelli said. "It's bad for the market when there's so much contradictory information out there."