SAN FRANCISCO, March 26, 2019 (GLOBE NEWSWIRE) -- Lightbend today announced that Kubeflow, the industry-standard framework for machine learning workflows in Kubernetes-based environments, is now a certified component of Lightbend Platform, the industry’s leading system for building streaming data pipelines and reactive microservices.
“A major challenge when leveraging machine learning and artificial intelligence is integrating your data science processes with production deployments. In the world of Kubernetes, including industry-leading OpenShift, Kubeflow is an open-source system for bridging these gaps,” said Dean Wampler, VP Fast Data Engineering at Lightbend. “By certifying Kubeflow with Lightbend Platform, organizations have a single platform for building, deploying, and running the spectrum of applications powering the real-time enterprise from reactive microservices, to streaming data pipelines, to machine learning and artificial intelligence workflows.”
Lightbend Platform on OpenShift provides the production tools needed to build streaming data pipelines and reactive microservices:
- Streaming data tools: Spark, Kafka, Akka Streams, and Kafka Streams;
- Reactive microservices: Akka, Play, Lagom;
- Production monitoring: Lightbend Console.
Kubeflow brings an integrated suite of tools for data science and production deployments of ML/AI applications:
- Research: JupyterHub;
- Model development and training: Katib, PyTorch, MXNet, and TensorFlow;
- Production model serving support: Argo, Minio, ModelDB, PyTorch Serving, and TensorFlow Serving.
With Kubeflow and Lightbend Platform, teams can build and deploy a range of applications, from stateful microservices to manage user interactions, to ETL jobs for moving data between datastores with required transformations, to ML applications like user recommendations in real time, all running on industry-leading OpenShift production platform.
Instructions for using Kubeflow are included with the Lightbend Platform documentation. Boris Lublinsky, Lightbend Fast Data Architect, wrote this 9-part blog on installing and using Kubeflow with OpenShift, lightbend.com/blog/how-to-deploy-kubeflow-on-lightbend-platform-openshift-introduction.