What Google, RedHat, Oracle, and others announced at KubeCon + CloudNativeCon 2018

Earlier this month, 4000+ developers attended the Cloud Native Computing Foundation’s flagship event, KubeCon + CloudNativeCon 2018 conference, held at Copenhagen, Europe from May 2nd to 4th. This conference focussed on a series of announcements on microservices, containers, and other open source tools for building applications for the web.

Top vendors including Google, RedHat, Oracle, and many more announced a myriad of releases and improvements with respect to Kubernetes. Read our article on Big vendor announcements at KubeCon + CloudNativeCon Europe.

Let’s brush through the top 7 vendors and their release highlights in this conference.

Google released Stackdriver Kubernetes Monitoring and open sourced gVisor

Released in beta, the Stackdriver Kubernetes Monitoring enables both developers and operators to use Kubernetes in a comprehensive fashion and also simplifies operations for them.

Features of Stackdriver Kubernetes Monitoring include:

Scalable Comprehensive Observability: Stackdriver Kubernetes Monitoring sums up logs, events and metrics from the Kubernetes environment to understand the behaviour of one’s application. These are rich, unified set of signals which are used by developers to build higher quality applications faster. It also helps operators speed root cause analysis and reduce mean time to resolution (MTTR).

Seamless integration with Prometheus: The Stackdriver Kubernetes Monitoring integrates seamlessly with Prometheus--a leading Kubernetes open source monitoring approach--without any change.

Unified view: Stackdriver Kubernetes Monitoring provides a unified view into signals from infrastructure, applications and services across multiple Kubernetes clusters. With this, developers, operators and security analysts, can effectively manage Kubernetes workloads. This allows them to easily observe system information from various sources, in flexible ways. Some instances include, inspecting a single container, or scaling up to explore massive, multi-cluster deployments.

Get started on-cloud or on-premise easily: Stackdriver Kubernetes Monitoring is pre-integrated with Google Kubernetes Engine. Thus, one can immediately use it within their Kubernetes Engine workloads. It is easily integrated with Kubernetes deployments on other clouds or on-premise infrastructure. Hence, one can access a unified collection of logs, events, and metrics for their application, regardless of where the containers are deployed.

Also, Google has open-sourced gVisor, a sandboxed container runtime. gVisor, which is lighter than a Virtual machine, enables secure isolation for containers. It also integrates with Docker and Kubernetes and thus makes it simple to run sandboxed containers in production environments. gVisor is written in Go to avoid security pitfalls that can plague kernels.

RedHat shared an open source toolkit called Operator Framework

RedHat in collaboration with Kubernetes open source community has shared the Operator Framework to make it easy to build Kubernetes applications. The Operator Framework is an open source toolkit designed in order to manage Kubernetes native applications named as Operators in an effective, automated and scalable manner.

The Operator Framework comprises of an:

Operator SDK that helps developers in building Operators based on their expertise. This does not require any knowledge of the complexities of Kubernetes API.

Operator Lifecycle Manager which supervises the lifecycle of all the operators running across a kubernetes cluster. It also keep a check on the services associated with the operators.

Operator Metering, which is soon to be added, allows creating a usage report for Operators providing specialized services.

Oracle added new open serverless support and key Kubernetes features to Oracle Container Engine

According to a report, security, storage and networking are the major challenges that companies face while working with containers. In order to address these challenges, the Oracle Container Engine have proposed some solutions, which include getting new governance, compliance and auditing features such as Identity and Access Management, role-based access control, support for the Payment Card Industry Data Security Standard, and cluster management auditing capabilities.

Scalability features: Oracle is adding support for small and virtualized environments, predictable IOPS, and the ability to run Kubernetes on NVIDIA Tesla GPUs.

New networking features: These include load balancing and virtual cloud network.

Storage features: The company has added the OCI volume provisioner and flexvolume driver.

Additionally, Oracle Container Engine features support for Helm and Tiller, and the ability to run existing apps with Kubernetes.

Kublr announced that its version 1.9 provides easy configuration of Kubernetes clusters for enterprise users

Kublr unleashed an advanced configuration capability in its version 1.9. This feature is designed to provide customers with flexibility that enables Kubernetes clusters to meet specific use cases.

The use cases include:

GPU-enabled nodes for Data Science applications

Hybrid clusters spanning data centers and clouds,

Custom Kubernetes tuning parameters, and

Meeting other advanced requirements.

New features in the Kublr 1.9 include:

Kubernetes 1.9.6 and new Dashboard

Improved backups in AWS with full cluster restoration

An introduction to Centralized monitoring, IAM, Custom cluster specification

Read more about Kublr 1.9 on Kublr blog.

Kubernetes announced the availability of Kubeflow 0.1

Kubernetes brought forward a power-packed package for tooling, known as Kubeflow 0.1. Kubeflow 0.1 provides a basic set of packages for developing, training, and deploying machine learning models.

This package:

Supports Argo, for managing ML workflows

Offers Jupyter Hub to create interactive Jupyter notebooks for collaborative and interactive model training.

Provides a number of TensorFlow tools, which includes Training Controller for native distributed training. The Training Controller can be configured to CPUs or GPUs and can also be adjusted to fit the size of a cluster by a single click.

Additional features such as a simplified setup via bootstrap container, improved accelerator integration, and support for more ML frameworks like Spark ML, XKGBoost, and sklearn will be released soon in the 0.2 version of KubeFlow.

CNCF(Cloud Native Computing Foundation) announced a new Certified Kubernetes Application Developer program

The Cloud Native Computing Foundation has successfully launched the Certified Kubernetes Application Developer (CKAD) exam and corresponding Kubernetes for Developers course.

The CKAD exam certifies that users are fit to design, build, configure, and expose cloud native applications on top of Kubernetes. A Certified Kubernetes Application Developer can define application resources and use core primitives to build, monitor, and troubleshoot scalable applications and tools in Kubernetes.

Read more about this program on the Cloud Native Computing Foundation blog.

DigitalOcean launched managed Kubernetes service

DigitalOcean cloud computing platform launched DigitalOcean Kubernetes, which is a simple and cost-effective solution for deploying, orchestrating, and managing container workloads on cloud. With the DigitalOcean Kubernetes service, developers can save time and deploy their container workloads without the need to configure things from scratch. The organization has also provided an early access to this Kubernetes service.

Read more on the DigitalOcean blog.

Apart, from these 7 vendors, many others such as DataDog, Humio, Weaveworks and so on have also announced features, frameworks, and services based on Kubernetes, serverless, and cloud computing.

This is not the end to the announcements, read the KubeCon + CloudNativeCon 2018 website to know about other announcements rolled out in this event.