Downloaded 64 times
Uploaded byVMware Tanzu
Communication Amongst Microservices: Kubernetes, Istio, and Spring Cloud - Angela Chin
This document provides an overview of communication amongst microservices using Kubernetes, Istio, and Spring Cloud. It discusses how Kubernetes is a container orchestrator that allows developers to run applications across infrastructure, and how Pivotal Container Service (PKS) provides managed Kubernetes. Istio is introduced as a platform that connects, secures, and observes microservices, utilizing sidecar proxies. Spring Cloud services are also discussed as providing abstractions for common patterns in distributed systems. The presentation explores how Istio and Kubernetes can work together to provide capabilities like retries, load balancing and mutual TLS for microservices, and compares this to features provided by Spring Cloud.
More Related Content
Similar to Communication Amongst Microservices: Kubernetes, Istio, and Spring Cloud - Angela Chin
Communication Amongst Microservices: Kubernetes, Istio, and Spring Cloud - Angela Chin
- 1. © Copyright 2017Pivotal Software, Inc. All rights Reserved. Version 1.0 Communication Amongst Microservices: Kubernetes, Istio, and Spring Cloud Angela Chin, Senior Software Engineer, Pivotal Twitter: @AngelaSChin
- 7. I’m Angela andI love ● (Software) Networking ● Hot yoga ● Hiking ● Desserts ● Food
- 8. Thanks for joiningme. Let’s get started!
- 9.
- 10.
- 11.
- 12.
- 13.
- 14. Confidential Kubernetes ● Container orchestratorbuilt around Docker
- 15. Confidential Kubernetes ● Container orchestratorbuilt around Docker ● “Have a datacenter on their laptop”
- 16. Confidential Kubernetes ● Container orchestratorbuilt around Docker ● “Have a datacenter on their laptop” ● Dev env == production env
- 17. Confidential Kubernetes ● Container orchestratorbuilt around Docker ● “Have a datacenter on their laptop” ● Dev env == production env ● Multi-cloud
- 18. Confidential Kubernetes ● Container orchestratorbuilt around Docker ● “Have a datacenter on their laptop” ● Dev env == production env ● Multi-cloud ● Community convergence & 3rd party integrations
- 19. Confidential Kubernetes ● Container orchestratorbuilt around Docker ● “Have a datacenter on their laptop” ● Dev env == production env ● Multi-cloud ● Community convergence & 3rd party integrations ● Flexibility
- 20. Confidential Kubernetes ● Runs applications ●Reduces infrastructure code ● Reduces toil
- 21.
- 22.
- 23. Confidential PKS ● Managed Kubernetesfor multiple clusters
- 24. Confidential PKS ● Managed Kubernetesfor multiple clusters
- 25. Confidential PKS ● Managed Kubernetesfor multiple clusters
- 26. Confidential PKS ● Managed Kubernetesfor multiple clusters ● Focused on Day 2 Operations
- 27. Confidential PKS ● Managed Kubernetesfor multiple clusters ● Focused on Day 2 Operations ● Uses BOSH
- 28. Confidential BOSH ● Day 2Operations ● Large-scale cloud software ● Security fixes within 48 hours ● Security aware ● Monitors and resurrects VMs ● Reproducible deployments
- 29. Embedded OS (Windows &Linux) NSX-T CPI (15 methods) v1 v2 v3 ... CVEs Product Updates Java | .NET | NodeJS Pivotal Application Service (PAS) Application Code & Frameworks Buildpacks | Spring Boot | Spring Cloud | Steeltoe Elastic | Packaged Software | Spark Pivotal Container Service (PKS) >cf push >kubectl run YOU build the containerWE build the container vSphere Azure & Azure StackGoogle CloudAWSOpenstack Pivotal Network “3Rs” Github Concourse Concourse Pivotal Services Marketplace Pivotal and Partner Products Continuous delivery Public Cloud Services Customer Managed Services Repair — CVEs Repave Rotate — Credhub Pivotal Cloud Foundry
- 30. Confidential I haz aKubernetes. What does that mean?
- 31.
- 32.
- 33.
- 34.
- 35.
- 36.
- 37. The Kubernetes APIis the single source of truth.
- 38.
- 39.
- 40.
- 41.
- 42. Confidential Master Master Master Master Master Worker kube-apiserver etcd kubectl apply -fmyApp.yml scheduler kube-controller- manager
- 43. Confidential Master Master Master Master Master Worker kube-apiserver etcd kubectl apply -fmyApp.yml scheduler kube-controller- manager kubelet
- 44. Confidential Master Master Master Master Master Worker kube-apiserver etcd kubectl apply -fmyApp.yml scheduler kube-controller- manager kubelet
- 45. Confidential Master Master Master Master Master Worker kube-apiserver etcd kubectl apply -fmyApp.yml scheduler kube-controller- manager kubelet myApp
- 46.
- 47.
- 48. Confidential But why doI care again?
- 49. Confidential What am Ideploying??
- 50.
- 51.
- 52. So you’re buildingsome microservices...
- 53.
- 54.
- 56. Microservices ● Smaller, simplerand easier to test ● Develop, deploy and run independently ● Organization can scale to many teams
- 57. Tradeoffs with Microservices ProsCons Strong Modular Boundaries Distributed Systems Independent Deployment Eventual Consistency Diversity of Technology Operational Complexity
- 58. Fallacies of Distributedsystems ● The network is reliable. ● Latency is zero ● Bandwidth is infinite ● The network is secure. ● Topology doesn't change. ● There is one administrator. ● Transport cost is zero. ● The network is homogeneous.
- 60.
- 61. Things your microservicesought to do... • Client-side retries • Load balancing • Mutual TLS • Configurable timeouts • Collecting metrics
- 62.
- 63. Confidential Kubernetes ● Container orchestratorbuilt around Docker ● “Have a datacenter on their laptop” ● Dev env == production env ● Multi-cloud ● Community convergence & 3rd party integrations ● Flexibility
- 64. Confidential Kubernetes ● Container orchestratorbuilt around Docker ● “Have a datacenter on their laptop” ● Dev env == production env ● Multi-cloud ● Community convergence & 3rd party integrations ● Flexibility
- 65.
- 66. Project to connect,secure, manage and observe microservices. ● Open source ● Platform agnostic ● Polyglot services What is Istio?
- 67. Project to connect,secure, manage and observe microservices. ● Utilizes sidecar proxies ● Service mesh What is Istio?
- 69. Things your microservicesought to do... • Client-side retries • Load balancing • Mutual TLS • Configurable timeouts • Collecting metrics
- 70.
- 71. Retries Proxy Frontend Proxy Backend 3 Proxy Backend 2 Proxy Backend1 Istio subscribe!
- 72. Retries Proxy Frontend Proxy Backend 3 Proxy Backend 2 Proxy Backend1 Istio backends!
- 73.
- 74.
- 75.
- 76.
- 77. Istio Terminology • Service:unit of (destination) application. • Service Versions: variants of (destination) application binary. • Service Registry: keeps track of pods/VMs of a service • Virtual Service: rule defining what service to send a request to • Destination Rule: rule defining what to do after destination service is identified
- 78. apiVersion: config.istio.io/v1alpha3 kind: VirtualService metadata: name:bananagram spec: hosts: - banana http: - route: - destination: host: banana Example Virtual Service Config
- 79. apiVersion: config.istio.io/v1alpha3 kind: DestinationRule metadata: name:bananagram spec: hosts: - banana trafficPolicy: connectionPool: tcp: maxConnections: 1 Destination Rule
- 80. Confidential Istio & Kubernetes:Better Together
- 82. $ helm templateinstall/kubernetes/helm/istio --name istio -- namespace istio-system > $HOME/istio.yaml $ kubectl create namespace istio-system $ kubectl apply -f $HOME/istio.yaml
- 83.
- 85.
- 86.
- 87. What about SpringCloud Services?
- 88. Spring Cloud Mission toeliminate boilerplate associated with distributed systems problems for Spring Boot applications. Started by using established libraries and servers (nothing was baked into platforms) ● Netflix OSS (Eureka, Ribbon, Hystrix) ● Hashicorp Consul ● Apache Zookeeper
- 89. Level of abstraction Loremipsum dolor sit amet, consectetuer adipiscing elit. Aenean commodo ligula eget dolor. Aenean massa. Cum sociis natoque penatibus et magnis dis parturient montes, nascetur ridiculus mus. Donec quam felis, ultricies nec, pellentesque ● Spring and Istio address the problems of distributed systems at different layers of abstraction ● “Istio helps decouple operations of a cluster from the application developer” - Eric Brewer, Google (VP, Infrastructure) Application Platform Or Container Orchestrator IaaS
- 90. Istio or SpringCloud (or both?) ● If the capability is provided by the platform (Istio), use it. Except for … ○ Fallbacks ○ Tracing Propagation ○ Security ● Istio or a service mesh architecture works better for polyglot environments ● But it also depends on your risk propensity ○ Istio GA’d in July 2018 ○ Spring Cloud has been around for 4 years ● Performance is also a factor For more info, watch: A Tale of Two Frameworks
- 91.
- 92. We Covered ● Kubernetes ●Pivotal Container Service (PKS) ● Microservices ● Istio ● Spring Cloud Services
- 93. Its been ajourney gif
- 98.
- 102.
- 103. Retries: Istio apiVersion: networking.istio.io/v1alpha3 kind:VirtualService metadata: name: foo spec: hosts: - foo http: - route: - destination: host: foo subset: v1 retries: attempts: 3
- 104. Retries: Spring @Service class Service{ @Retryable(RemoteAccessException.class) void service() { // ... do something } @Recover void recover(RemoteAccessException e) { // ... panic } } ● Annotation or Template ● Stateless or Stateful ● Flexible Retry Policies ● Backoff Policies (e.g. exponential)
- 105. Confidential In process vsout of process architecture
- 106. Frontend Backend A Library! In-processarchitecture
- 110. Multiple languages, Multiplelibraries Each with… • Different features • Different configuration • Different quirks Polyglot shouldn't be painful!
- 111. Outline • What yourmicroservices need • How sidecars help • Envoy & Istio • Kubernetes Integration • Cloud Foundry Integration
- 112. Out-of-process architecture Your AppSome service Separate process!
- 113. Out-of-process architecture Your App -retries - load balancing - mutual TLS - timeouts - metric collection - etc. Your business logic here
- 114. Your App (acting as aservice) Client app Out-of-process architecture
- 115.