Cloud Native + Open Source Virtual Summit China 2020

Android have been spread to some use cases such as cloud and edge computing. One example is Android Cloud gaming, which runs many Android instance in Cloud and streaming encoded video to remote devices. This use case requires a solution to consolidate many Android runtime into single platform with little overhead, strong isolation and lessen management overhead.
This presentation proposes an innovate idea to run Android in Container to fit for the uses case of cloud computing and edge computing. Trough enhancement docker and Kubernetes, millions of Android instances are deployed and managed on Cloud easily. Prometheus is used to monitor the status of Android instance and managed instance simultaneously. Performance data show its impressive performance and good scalability.

Kubevirt, as a plugin to K8S, uses CRD extensions to provide K8S with the ability to manage virtual machines like containers.

However, VM cluster in Baidu not only manages the VM, but also manages VM related resources, such as network, custom images and image repositories, and user configuration. And thus, Kubevirt's support for those resources is not sufficient.

This talk will introduce how to use macvtap to implement a dual network adapters, connect to a centralized image service to manage VM images, and inject user configuration and password during the initialization process in a Kubevirt-based lightweight VM cluster. This talk will also cover how to deploy K8S and kubevirt in the edge environment to realize the virtual machine cluster management capabilities at the edge in Baidu.

SIG-Multicluster is focused on solving common challenges related to the management of multiple Kubernetes clusters, and applications that exist therein. We are responsible for designing, discussing, implementing and maintaining API’s, tools and documentation related to multi-cluster administration and application management. This includes not only active automated approaches such as Cluster Federation, but also those that employ batch workflow-style continuous deployment systems like Spinnaker and others. Standalone building blocks for these and other similar systems (for example a cluster registry), and proposed changes to kubernetes core where appropriate are in scope. In this intro we will give you an overview of the projects we're currently actively working on, how best to get involved, and what our future plans look like.

As Kubernetes is one of the biggest open-source projects worldwide, the creation of a new release can be an overwhelming topic. The major burden is not only to release a new version every three months, but also to provide stability and security fixes over the whole support period of a release.

In this talk, we will cover the details behind the lifecycle of a Kubernetes release. The presentation will contain everything from assembling the Release Team before the actual cycle begins up to maintaining the release branches, which provide continuous patch releases to the community. The talk will outline the technical process of branch management by covering all of the necessary stages during the release cycle. We will also demonstrate our Release Engineering tooling and dive into the design decisions we’ve made as a group that shape the overall process.

Federated Machine Learning (FML) is a one of the most promising machine learning technologies to solve data silos and strengthening of data privacy and security, which allows collaborating organizations to create models without leaking data privacy. FATE is an open source project providing a secure MPC framework to support FML architecture. The challenge of FML is how to coordinate different parties usually reside in different geographical location.

This talk will introduce how we do FML in Kubernetes by Kubeflow. The flexibility and scalability of Kubernetes helps a lot on customizing our platform requirements to fit FML. A FATE-Operator is designed to run FML jobs on Kubernetes, which fit FML into existed Kubeflow ecosystem.

In this session, audience will learning:
1. What is FML?
2. How to design a Kubernetes Operator;
3. The experience of running FML in Kubernetes.

Cloud computing is fast moving. For Java app, Spring Cloud simplifies distributed, microservice-style architecture by implementing proven patterns (e.g. service discovery, circuit breaker) to help you easily build a cloud-native app. However, you still need to face several critical cloud challenges to run and manage your cloud-native apps.
How to orchestrate resources to meet dynamically growing business needs? How to have more flexible process to accelerate the app deployment and management? How to achieve a seamless pipeline with advanced cloud monitoring and tracing solutions? Come to hear best practices to host Spring Cloud production on top of Kubernetes. The speaker will demonstrate how to easily deploy cloud-native apps to Microsoft Azure and accelerate your digital transformation with the power of Spring Cloud, Kubernetes, Cloud-Native Buildpacks and the cloud platform.

This is a presentation by the recently formed Kubernetes VMware User Group. This group addresses running all forms of Kubernetes on VMware infrastructure. Kubernetes is in the process of deprecating the cloud provider and storage plugins that have historically been built into the source tree and directly packaged with many distributions. The new out-of-tree cloud provider and CSI storage plugin are now recommended for production use. The legacy plugins are no longer getting feature upgrades. However, when running on older hypervisor or Kubernetes releases, choices may be constrained. This session will cover the in vs out-of-tree feature and compatibility matrix, offer guidance on migration to out-of-tree - and cover documentation and support resources.

The Provider IBM Cloud is a subproject of the Cloud Provider SIG. In this session, the Provider IBM Cloud leads will provide an introduction of the project, and brief about the activities and GitHub repo on Kubernetes Cluster API Provider IBM Cloud. This will be followed by an overview of IBM Provider used by IBM Cloud Kubernetes Service (IKS) and Red Hat OpenShift on IBM Cloud, and how the provider leverages the Kubernetes Cloud Controller Manager. The session should be informative to anyone with interest in the Kubernetes Provider IBM Cloud subproject.

Serverless compute engines，like Fargate on AWS，ACI on Azure，CloudRun on Google and CCI on HuaweiCloud， emerges on public cloud platforms since 2018，with these services， users can easily launch their computing task on clouds， without provisioning and managing their servers.Although the billing can be computed by the second, which can reduce many costs for users, there is still no end for cloud providers when refering to reducing costs while perserving a high performance.

In this session, we will share how we reduce our management costs through offloading, management slimming, and schedule optimizing. we provides container service with extremly low overhead. Also, we will give a breaf introduction on how we improve the performance of tasks from different industris such as gene computing, big data processing, and deep learning.

Volcano is a batch system built on Kubernetes. It provides a suite of mechanisms that are commonly required by many classes of batch & elastic workload including: machine learning/deep learning, bioinformatics/genomics and other "big data" applications. These types of applications typically run on generalized domain frameworks like TensorFlow, Spark, PyTorch, MPI, etc, which Volcano integrates with. This deep dive session will will give an introduction on project Volcano, including the features that required by many batch/elastic workload, e.g. big data, AI, and HPC. And demonstrate the benefit of using Volcano in those scenarios. At the end of the session, the roadmap and future plan will be introduced and the target scenarios.