A Day in the Life of a Data Scientist Conquer Machine Learning Lifecycle on Kubernetes Brian Redmond • Cloud Architect @ Microsoft (18 years) • Azure Global Black Belt Team • Live in Pittsburgh, PA Repeatable/consistent • CI/CD • This has worked well for App Dev. Now time for AI/ML • But, must ensure data scientist are not hindered by structure Why Containers, Kubernetes & Helm? • Container • Contains Scalable • Easy to explore hyper-parameters space • Easy to do distributed training But really, Data Scientists shouldn’t have to care about containers, kubernetes and all that stuff • Pachyderm can

Apache Beam TensorFlow Service Communication Management Container Orchestration Data Processing Pipelines Data Flow Graphs for Machine Intelligence Kubernetes Contributors opensource.google.com & On-Prem ● vSphere 6.5 for Alpha ● Simple CLI installation ● Online and Offline installation ● Private container registry support ● Latest 3 versions of k8s ● High-availability control plane ● Auto-repair Generate CA certificate and key Please enter your numeric choice [2]: 2 Enter the path to the SSH private key to use (leave empty to generate): /Users/karangoel/.ssh/vsphere Enter proxy without "http://"

Upgrades - Failures Kubernetes (As a Bosh Release) BOSH (Deploys/Manages VMs) CPI CNI Harbor Private Container Registry Worker Worker Worker Kubernetes Cluster Services API Worker Worker Developer Structured Data Metrics Alerts Events VMware vRealize Operations Capacity, Performance and Configuration Management Events Launch in Context Unstructured Data Logs Messages VMware Engine (GKE) 33 CONFIDENTIAL Google Container Engine (GKE) Pivotal Container Service (PKS) Private and Public Cloud Public Cloud (on GCP) Certified for every K8s release Container Service

研发中 - Amazon EKS on Fargate - Managed Nodes - Managed add-ons - DNS resolution of Amazon EKS private endpoints - New Amazon EKS Regions: Sao Paulo, Canada Central - Next-generation CNI plugin © 加固 监控 NETWORK 配置 VPC 网络策略 路由表 NACLs 数据 网络流量保护 客户端加密 服务端加密 EKS CONTROL PLANE CONTROL PLANE 配置 PRIVATE CONTROL RBAC 策略 © 2019, Amazon Web Services, Inc. or its Affiliates. All rights reserved. Amazon password) { • business core data • Personal Identifiable Information (PII) • gotchas: leaks, GDPR (in Europe) { host container dependencies code config user data © 2019, Amazon Web Services

com/developer/article/1627330 Step1: 安装 docker-ce # 安装需要的支撑软件 yum install -y yum-utils device-mapper-persistent-data lvm2 # 添加 yum 源 yum-config-manager --add-repo https://download.docker.com/linux/centos/docker-ce configmap vi 3-configmap.yaml kind: ConfigMap apiVersion: v1 metadata: name: traefik-config data: traefik.yaml: |- serversTransport: insecureSkipVerify: true api: 的登录密码(登录用户为 admin) kubectl -n rook-ceph get secret rook-ceph-dashboard-password -o jsonpath="{['data']['password']}" | base64 --decode && echo Step4: 配置 storageclass download https://github.c

Container Containerd 用户VPC SLB NAS ECS RDS OSS SLS NAT Gateway Container Registry Private Zone ENI Pod 使用场景 Use cases • Multimedia processing • IoT sensor messages processing • • HTTP REST APIs and web applicated • Mobile backends • Continuous Integration Pipeline • Data Analytics • … Asynchronous Event Driven Stateless High dynamic Short Duration 14 -

- 1 Utility/Public subnet # - 1 NGW for the private subnet to NAT to # - 1 Route Table Association to the Main Route Table # - 1 Private subnet (to hold the instances) # --------------- - 1 Utility/Public subnet # - 1 NGW for the private subnet to NAT to # - 1 Route Table Association to the Main Route Table # - 1 Private subnet (to hold the instances) # --------------- - 1 Utility/Public subnet # - 1 NGW for the private subnet to NAT to # - 1 Route Table Association to the Main Route Table # - 1 Private subnet (to hold the instances) # ---------------

����Volume����! • shared: ����! • slave�����! • private������K8S 1.6����! �����/ rbind / Shared /mnt/a! /mnt/a! Slave /mnt/a! /mnt/c! /mnt/a! Private /mnt/a! /mnt/c! /mnt/a! /mnt/b! • ���������/mnt/a rbind��� /mnt/c! 1. ��A� RO �����RBD volume: /var/lib/kubelet/plugins/xxxx! 2. node-exporter � make-private ����������! 3. ��A������RBD volume����rbd umap����node-exporter� � RO ��! 4. ��B� RW �������RBD ����! • node-exporter! • node-exporter�K8S daemonset������������������! • K8S���mount make-private��������volume��������! • ceph��! • ceph rbd client����������! • ��lsblk����������! • K8S! •

Instance Pod Viking agent Container Container Pod Get Pod status Service/Ingress DNS Entry SLB Private Zone ASK-Scheduler K8S API Server Metrics API CloudMonitor, Prometheus HPA Controller Cloud Server • 基于云产品控制器降低Kubernetes集群的复杂度 • 使用PrivateZone代替coredns服务发现 • 使用SLB layer-7（ALB）作为默认Ingress Private Zone ALB Serverless容器基础设施 - ECI • 更低的计算成本：弹性成本要低于ECS，long run应用成本要接近ECS包年包月 • 更高的弹性效率：ECI扩容速度要远高于ECS

your GOPATH • Go into your project path and init your project ‘your-domain’ would be like your private tenant name. • Then initialize your own resource group, version and kind. • Your API server could