Kubernetes & Cloud-Native für IT Entscheider - Sebastian Scheele & Tobias Schneck, Loodse

Für die meisten Mitarbeiter außerhalb der IT-Abteilungen ist die richtige Aussprache von Kubernetes der erste Schritt von vielen, um zu verstehen, was die Vorteile von zu Cloud-Native Technologien und welche Auswirkungen es auf das Business haben kann. Ziel dieses Workshop ist es, Kubernetes und Cloud-Native für Entscheider in IT-Abteilung vorzustellen und ihnen zu erläutern, wie sie von der Unterstützung bei der Einführung von Cloud-Native Technologien profitieren können und wie sie auch nicht IT-Abteilungen begeistern können. Der Workshop beginnt mit einem Überblick über Cloud-Native, führt zu einer nicht-technischen Erklärung von Kubernetes unter Verwendung von Häusern als Metapher und endet mit entscheidenden Vorteilen für Nicht-IT-Abteilungen bei der Einführung von Kubernetes. Die Teilnehmer lernen die Vorteile von Kubernetes & Cloud-Native Technologien, wie man Kubernetes auf organisatorischer Ebene erklärt und sich für sie einsetzt.

Transcript

1. Kubernetes & Cloud-Native für IT-Entscheider
2. Agenda • Intro • Cloud Native for Business • The Journey of Cloud Native • Container Fundamentals • Kubernetes: The Cloud Native Operating System • From Legacy to Cloud Native and how to modernize you application • Open discussion
4. #1 IT challenge is organizational change
5. Dev Ops
6. IT Business
7. Kubernetes
8. Cloud Native
9. IT Business
10. Cloud Native
11. A lot of words
14. A Look At IT Adoption
15. IT Business
16. IT Business
17. apiVersion: kubermatic.k8s.io/v1 kind: Cluster metadata: name: g8gpmr9fw3 spec: cloud: dc: hetzner-nbg1 hetzner: token: SOME_SECRET_TOKEN clusterNetwork: dnsDomain: cluster.local pods: cidrBlocks: - 172.25.0.0/16 services: cidrBlocks: - 10.10.10.0/24 humanReadableName: demo version: 1.10.8 address: adminToken: 975w4b.7tcjkwl5zzhjrsg5 externalName: g8gpmr9fw3.europe-west3-c.dev.kubermatic.io ip: 35.198.93.90 url: https://g8gpmr9fw3.europe-west3-c.dev.kubermatic.io:32489 status: health: apiserver: true controller: true etcd: true machineController: true openvpn: true scheduler: true namespaceName: cluster-g8gpmr9fw3 Code
18. Business • Cloud spending will top $203.4 billion by 2020 - IDC • 67% of enterprise IT infrastructure and software spending will be for cloud by 2020 - Gartner
19. apiVersion: kubermatic.k8s.io/v1 kind: Cluster metadata: name: g8gpmr9fw3 spec: cloud: dc: hetzner-nbg1 hetzner: token: SOME_SECRET_TOKEN clusterNetwork: dnsDomain: cluster.local pods: cidrBlocks: - 172.25.0.0/16 services: cidrBlocks: - 10.10.10.0/24 humanReadableName: demo version: 1.10.8 address: adminToken: 975w4b.7tcjkwl5zzhjrsg5 externalName: g8gpmr9fw3.europe-west3-c.dev.kubermatic.io ip: 35.198.93.90 url: https://g8gpmr9fw3.europe-west3-c.dev.kubermatic.io:32489 status: health: apiserver: true controller: true etcd: true machineController: true openvpn: true scheduler: true namespaceName: cluster-g8gpmr9fw3 • Cloud spending will top $203.4 billion by 2020 - IDC • 67% of enterprise IT infrastructure and software spending will be for cloud by 2020 - Gartner Business Code ≈
20. Better Faster Cheaper S ~Business Code~
21. What is Cloud Native? Cloud native technologies empower organizations to build and run scalable applications in modern, dynamic environments such as public, private, and hybrid clouds. Containers, service meshes, microservices, immutable infrastructure, and declarative APIs exemplify this approach. These techniques enable loosely coupled systems that are resilient, manageable, and observable. Combined with robust automation, they allow engineers to make high-impact changes frequently and predictably with minimal toil.
22. What is Business Cloud Native? Cloud native technologies help businesses build modern software that delivers a good and/or service to their customers. These technologies create better systems that allow engineers to make changes faster at a lower cost. S
23. History Lesson: The Journey to Cloud Native
24. Software Application Servers 01000011011011110110010001100101
25. Person House
26. Hosting Continuum Data Center Cloud Virtual Machine Container Serverless Home Construction Renting Apartment Flat Share Hotel
27. Cloud Native Continuum These technologies create better systems that allow engineers to make changes faster at a lower cost. Simplicity Agility Density Isolation Customization Control
28. Build Your Own Data Center Home Construction
29. Non-Coder Translation Build Your Own Isolation Customization Control
30. Non-Coder Translation Build Your Own Problems Better Faster Cheaper Two Years CapEx
31. Rental Cloud Landlord
32. Non-Coder Translation Rental Better Faster Cheaper Two Weeks ↓ Two Days CapEx ↓ OpEx
33. Non-Coder Translation Rental Problems Better Faster Cheaper Two days Dedicated costs Manual set-up process
34. Split Space Virtual Machine Furnished Apartment Guest OS Server Binaries /Libraries App 1 Host OS Guest OS Binaries /Libraries App 1’ Guest OS Binaries /Libraries App 2 Hypervisor
35. Non-Coder Translation Split Space Better Faster Cheaper Two Days ↓ Ten Minutes Dedicated ↓ Per-second-billing Manual set-up process ↓ Consistent deployments
36. Non-Coder Translation Split Space Problems Better Faster Cheaper Ten Minutes Single use Clashing dependencies
37. Shared Space Container Flat Share Server Host OS Binaries /Libraries Binaries /Libraries App 1 App 1’ App 2'’ App 2’ App 2
38. Non-Coder Translation Smaller Sharing Better Faster Cheaper Ten Minutes ↓ 1/2 Second 15% CPU utilization ↓ 80% CPU Utilization Separation of dependencies
39. Non-Coder Translation Smaller Sharing Problems Better Faster Cheaper Set up time Always on Set up required
40. On-demand Serverless Hotel
41. Non-Coder Translation On-demand Better Faster Cheaper Always ready Pay per Execution No setup required
42. Hosting Continuum Data Center Cloud Virtual Machine Container Serverless Home Construction Renting Apartment Flat Share Hotel Simplicity Agility Density Isolation Customization Control
43. Cloud Native Continuum Simplicity Agility Density Focus on business value, not the infrastructure
44. What is Business Cloud Native? Cloud native technologies help businesses build modern software that delivers a good and/or service to their customers. These technologies create better systems that allow engineers to make changes faster at a lower cost. S
45. Building a Business Case
46. The Paradox 91% of companies use hybrid and multi-cloud 36% 10% use orchestration use automated orchestration 72% cite complexity as major challenge
47. The New World Is Here Old World New World Infrastructure: Vendor lock-in → Continuous freedom of choice Dev Experience: Waiting for Ops → Self-service platform Operations: Resource-intensive → Automated Resilience: Costly downtimes → Highly available, self-healing infrastructure Cloud costs: Cloud-provider dependant → Dynamically move workloads for best price
48. Use Case: Application on VM or on Containers? Infrastructure (100 nodes*$ 30) $ 30,000 p.a. Total $ 30,000 p.a. Scenario 1: “Application as VMs” Infrastructure (15 nodes*$ 60) $ 10,800 p.a. Total $ 10,800 p.a. Scenario 2: “Application as Containers in Kubernetes” Total Cost Saving $ 19,200 p.a. Enterprise X runs numerous services. They want to evaluate running the Applications as dedicated VM or as Containers in a Kubernetes cluster. Cuts Costs
49. Infrastructure (100 nodes*$ 30) $ 30,000 p.a. Application management (4 FTE) $ 800,000 p.a. Total $ 908,000 p.a. Scenario 1: “Application as VMs” Infrastructure (15 nodes*$ 60) $ 10,800 p.a. Application management (1 FTE) $ 200,000 p.a. Total $ 210,800 p.a. Scenario 2: “Application as Containers in Kubernetes” Total Cost Saving $ 619,200 p.a. Simplify Management Use Case: Application on VM or on Containers? Enterprise X runs numerous services. They want to evaluate running the Applications as dedicated VM or as Containers in a Kubernetes cluster.
50. Infrastructure (100 nodes*$ 30) $ 30,000 p.a. Application management (4 FTE) $ 800,000 p.a. Downtime (1 hour) $ 302,800 Total $ 1,210,800 p.a. Scenario 1: “Application as VMs” Infrastructure (15 nodes*$ 60) $ 10,800 p.a. Application management (1 FTE) $ 200,000 p.a. Total $ 210,800 p.a. Scenario 2: “Application as Containers in Kubernetes” Total Cost Saving $ 922,000 p.a. Fix Faster Use Case: Application on VM or on Containers? Enterprise X runs numerous services. They want to evaluate running the Applications as dedicated VM or as Containers in a Kubernetes cluster.
51. Container Fundamentals
52. Why the Container Hype? The software industry is going through a transition • From "Pets": • Monolithic applications • Long development cycles • Environment-dependency • Low scalability • To "Cattle": • Distributed applications • Fast, iterative improvements • Multiple, isolated environments • Faster scalability
53. The Deployment Challenge • Many different stacks: • Languages • Frameworks • Databases • Numerous targets: • Individual development environments • Pre-production, QA, staging... • Production: on premise, cloud, hybrid
54. Bananas and Coffee
55. Multiplicity of goods Interaction between goods (bananas next to spices Multiplicitiy of transportation and storage methods Quick and smooth transportation
56. The Rise of Shipping Containers Multiplicity of goods Interaction between goods (bananas next to spices Multiplicitiy of transportation and storage methods Quick and smooth transportation A standard container that can be loaded with any good and that is sealed until final delivery On it‘s way, it can be loaded, unloaded, stacked, moved over long distances....
57. USP: storage and transport tool
58. Handling and stackability
59. Standardization
60. Flexibility Container Types Standard Container High-Cube Container Hardtop Container Open Top Container Flat Platform (Plat) Ventilated Container Cooling Container Bulk Container Tank Container
61. Logistic chain
62. General cargo freighter vs. container ship Container Ship “Marit Maersk” Port of Hamburg Feb. 2009 Length: 366m width: 48,2m Depth: 15,5m Loading capacity: 11 400 TEU (Twenty Foot Equivalent Unit) Cargo Freighter “Santa Ines” 1951 Length: 148,04m Width: 18,60m Depth: 8,16 Loading capacity: 11 710 t
63. What working in the port used to look like...
64. ...and what it looks like now:
65. The Shipping Container Ecosystem
66. Same thing!
67. The Deployment Challenge Numerous stacks Appropriate interaction of services and apps Numerous hardware environments Smooth and quick migration Static website User DB Background workers Web frontend Queue API endpoint Analytics DB Deployment VM QA server Customer Data Center Public Cloud Disaster Recovery Production Server Production Cluster Contributor‘s laptop
68. A Nightmare Development VM QA server Single Prod Server Onsite Cluster Public Cloud Contributor‘s labtop Customer Servers Static website ? ? ? ? ? ? ? Web frontend ? ? ? ? ? ? ? Background workers ? ? ? ? ? ? ? User DB ? ? ? ? ? ? ? Analytics DB ? ? ? ? ? ? ? Queue ? ? ? ? ? ? ?
69. A Container System for Applications Numerous stacks Appropriate interaction of services and apps Numerous hardware environments Smooth and quick migration Static website User DB Web frontend Queue Analytics DB Deployment VM QA server Customer Data Center Public Cloud Production Cluster Contributor‘s laptop An isolated, lightweight, portable, self-sufficient container.... ...that can be manipulated using standard operations and run on virtually anx platform
70. Escape the Nightmare Development VM QA server Single Prod Server Onsite Cluster Public Cloud Contributor‘s labtop Customer Servers Static website Web frontend Background workers User DB Analytics DB Queue
71. Outcome • Standardization • Reduced dev-to-prod time • Reduced continuous integration job time • Declarative • leads to reproducibility • Abstraction and Isolation • contracts • Independence • Ease • Lots of containers!
72. A Timeline
73. What is a container? Really, in technical terms, what is it?
74. Containers vs VMs Containers are isolated but share OS kernel and, if appropriate, bins/libraries • Faster deployment • Less overhead • Easier migration • Faster restart
75. • Lightweight • Hermetically sealed • Isolated • Easily deployable • Introspectable • Runnable Linux processes • Improves overall developer experience • Fosters code and component reuse • Simplifies operations for cloud native applications A quick recap of containers
76. Demo https://www.katacoda.com/loodse/courses/ecns
77. Kubernetes: The Cloud Native Operating System
78. Containers are awesome! Let’s run lots of them!
79. Everything at Google runs in containers…. > 2 billion containers per week • Gmail, Web Search, Maps, ... • MapReduce, batch, ... • GFS, Colossus, … • Even Google’s Cloud Platform: VMs run in containers!
80. Software Applications Servers 01010010011010101001010001 10101010010101001010010101 01010001010101011010101010 10101010101010101111010101 01010100101000100101010101 01010100101010100101111010 Enterprise IT
81. People Housing City
82. What is Business Kubernetes? Kubernetes governs cloud native IT A City governs the housing market
83. What is Kubernetes? Kubernetes is a portable, extensible open-source platform for managing workloads and services, that facilitates both declarative configuration and automation.
84. What is Kubernetes? Greek - meaning helmsman or pilot Root of Governor cybernetics (study of communication and control)
85. Who “Manages” Kubernetes? Vendor neutral governance group of many “cloud native” projects
86. Working Definitions
87. Kubernetes Project Stats • Over 42,000 stars on Github • 1800+ Contributors to K8s Core • Most discussed Repository by a large margin • 50,000+ users in Slack Team
88. Kubernetes Project Stats
89. What is Kubernetes for the Business? Kubernetes is the governor of cloud native IT
90. Bare Metal Any Cloud Cloud Native Stack Kubernetes KubeVirt/Virtlet Software Applications Container Engine Containers Housing Stack Governor People Land Bare Metal Any Cloud Cloud Native Stack VMs Kubernetes KubeVirt/Virtlet Serverless Knative/OpenFaaS Software Applications Container Engine Containers Homes Hotels Apartments City Services
91. Bare Metal Any Cloud Cloud Native Stack Kubernetes KubeVirt/Virtlet Software Applications Container Engine Containers Housing Stack Governor Cloud Native Stack Kubernetes
92. What is Kubernetes? Kubernetes is a portable, extensible open-source platform for managing workloads and services, that facilitates both declarative configuration and automation.
93. What is the problem to solve? • We are going to talk about a solution… • Can we agree on a problem first? A problem can be defined as any situation in which a gap is perceived to exist between what is and what should be. - Arthur B. Van Gundy Jr.
94. What is the problem to solve? • We are going to talk about a solution… • Can we agree on a problem first? A problem can be defined as any situation in which a gap is perceived to exist between what is and what should be. - Arthur B. Van Gundy Jr. “Actual State” “Desired State” This! Exactly! Reconciliation
95. Problem Solved “Actual State” “Desired State” Reconciliation
96. Declarative Configuration Reconciliation Scheduler Controller
97. Kubernetes etcd API server Scheduler Controller You Reconciliation
98. City City Plan Mayor Builder Regulator You Reconciliation
99. Cloud Native City etcd API server Scheduler Controller City Plan Mayor Builder Regulator You
100. K8s Master API Server Dashboard Scheduler kubelet kubelet kubelet kubelet Container Registry etcd Controllers web browsers kubectl web browsers Config file Image CNI - Plugin Network (e.g. Fanal, Calico) Developer Developer
101. Bare Metal Any Cloud Cloud Native Stack Kubernetes KubeVirt/Virtlet Software Applications Container Engine Containers Housing Stack Governor Land Bare Metal Any Cloud Cloud Native Stack Kubernetes Building Blocks
102. What is Kubernetes? Kubernetes is a portable, extensible open-source platform for managing workloads and services, that facilitates both declarative configuration and automation.
103. Building Blocks Nodes Land
104. Bare Metal Any Cloud Cloud Native Stack Kubernetes KubeVirt/Virtlet Software Applications Container Engine Containers Housing Stack Governor Land Bare Metal Any Cloud Cloud Native Stack VMs Kubernetes KubeVirt/Virtlet Serverless Knative/OpenFaaS Container Engine Containers Homes Hotels Apartments City Services
105. What is Kubernetes? Kubernetes is a portable, extensible open-source platform for managing workloads and services, that facilitates both declarative configuration and automation.
106. Scaling Services Workloads and Services Houses Bare Metal Any Cloud Cloud Native Stack VMs Kubernetes KubeVirt/Virtlet Serverless Knative/OpenFaaS Container Engine Containers
107. Bare Metal Any Cloud Cloud Native Stack Kubernetes KubeVirt/Virtlet Software Applications Container Engine Containers Housing Stack Governor People Land Buildings Services Bare Metal Any Cloud Cloud Native Stack VMs Kubernetes KubeVirt/Virtlet Serverless Knative/OpenFaaS Software Applications Container Engine Containers Bare Metal Any Cloud Cloud Native Stack Kubernetes KubeVirt/Virtlet Software Applications Container Engine Containers Housing Stack Governor People Land Homes Hotels Apartments City Services
108. What is Kubernetes? Kubernetes is a portable, extensible open-source platform for managing workloads and services, that facilitates both declarative configuration and automation.
109. Day 2 IT City Shopping Mall Hospital Church University Clinic Museum Opera Docks CI/CD Observability Service proxy Service Mesh Streaming/messaging Networking Container registry Software distribution
110. Most Importantly... Use the SAME API across bare metal and EVERY cloud provider!!!
111. Demo https://www.katacoda.com/loodse/courses/ecns
112. FROM LEGACY TO CLOUD NATIVE 3 Patterns to Modernize your Application Portfolio
113. Monolithic Release Cycles Ops replace old with new Release Plan Intg, QA & UAT
114. Monolithic Release Cycles Ops replace old with new Release Plan Intg, QA & UAT Bugs Bugs
115. Microservice Release Cycles Feature to Prod Feature to Prod Feature to Prod Old feature still running Release Plan Release Plan Release Plan
116. Microservice Release Cycles Release Plan Release Plan Release Plan Feature to Prod Feature to Prod Feature to Prod Old feature still running Feature to Prod
117. Real Life Examples
118. The Result • Handled Christmas loads at scale without breaking a sweat • Quickly scale up and down. Burstable • Interchangeable front-end containers = automated sites with custom DNS and HTTPS • Developers have taken ownership of deployment, can update, roll-back, w/o sysadmins #1 Large Scale Web Applications Solution • Containerized tiered components • Small containers deployed at scale • CI/CD to manage and orchestrate containers • Development and prod have zero shift Problem • Content Management and E-Commerce needed at large scale • Servers did not scale under load • Multi-server states hard to synchronize • Development environment difficult to replicate
119. The Result • From <50 concurrent to <5,000 concurrent requests • Exposed data and functionality to Mobile, Salesforce and partner channels • No big bang rewrite upfront. Retiring legacy incrementally. • Enabled DevOps #2 Legacy Middleware with Huge Data Problem • Data and functionality were closely coupled • User session = DB connection • Database choked during the day • Hard to add new features Solution • New microservices layer as new entry point • Data caching to avoid expensive hits to DB • Containerized microservices talk to legacy • Messaging backbone
120. Modernizing Portfolio of Applications A Structured Approach
121. How we bring the App to production on Kubernetes? Continuous Integration (CI) Continuous Deployment (CD)
122. 3 Main Patterns in Modernizing Workloads Lift & Shift Augment with new layers Complete rewrite • Containerize existing components • Deploy them on a PaaS • Keep external integrations and data on legacy • Legacy applications have to be well written and suited • Legacy remains intact • New layer enables new capabilities • Deploy on PaaS • New integration points between legacy and new layers • Legacy is totally replaced • New interfaces and data • Use PaaS to run • Some data and features can be re-wrapped, but mostly are retired
123. Lift and Shift Modernization
124. Lift and Shift Modernization
125. Lift and Shift Step 1 • Lift existing runtimes into container • Keep data, messaging and integrations where they are • Enable development and ops to deliver value on a new platform
126. Lift and Shift Step 2 • Slice functionality by vertical business concerns • Shift feature after feature into vertically aligned microservices and APIs • Rinse and repeat. Retire legacy when feature parity is attained
127. Augment Based Modernization
128. Augment Based Modernization
129. Augment Based Modernization Step 1 • Create integrations points to expose legacy data and functionality. Legacy as backend • Create new capabilities through passthrough services • Deliver new business value
130. Augment Based Modernization Step 2 • Slice backend in vertical business aligned services • Refactor backend functionality into business aligned microservices • Rinse and repeat. Retire backend when feature parity is attained
131. Rewrite Based Modernization
132. How Do They Compare Complete ReWrite Augment with new layers Lift and Shift cost of migration Time Generally the most expensive and longest
133. Application Modernization Pathways Enabling modernization from legacy applications Starting Point Open Source & Enablement Cloud Native Enablement Desired State Lift and Shift Modernization Non Open Source Middleware Applications Migrate to Open Source Stack Enable on Kubermatic container platform Modernize applications to API and Microservices cloud architecture on containers Augment with new functionality Monolithic / Mainframe applications that are hard to change without breaking functionality Enable integrations to expose data and functionality using Open Source stack Enable on Kubermatic container platform Modernize applications to API and Microservices cloud architecture on containers Complete re-write Monolithics or applications on non Open Source middleware about to be retired Architecture and Design mapping old system capabilities to new architecture Set-up, train on and get hands-on experience with modern container platform from Loodse Create new set of applications on modern cloud-native applications
134. Methodology
135. Migration Methodology Iterative, managed service, factory scale up
136. Methodology and Process Deploy Phase: Knowledge Base Migration Teams (1….n) Challenge Backlog Center of Excellence C B A C B D E F