Converged Infrastructures on Kubernetes with Kubevirt
Download as PPTX, PDF0 likes16 views
AWS Ambassador Summit
Converged Infrastructures on Kubernetes with Kubevirt
- 2. Legacy Converged Infrastructures Transformation to ECS/EKS can be containerized cannot be containerized Database Freedom Commercial Database Application Modernization Framework Windows to Linux Windows Refactor the Monolith Event Storming Event Storming NoSQL CQRS Event Sourcing Microservices Replatform Refactor
- 4. 4
- 5. 5
- 6. 6
- 7. 7
- 8. 8
- 9. 9
- 10. Cloud-Native? ● Applications packaged in containers (e.g. Docker) ● Managed by an orchestrator like Kubernetes ● Run on public, private or hybrid clouds
- 11. How? ● AWS A2C (App2Containers) ● IDE functions (e.g. Visual Studio, Intellij …) ● DIY
- 12. Q&A?
- 14. Legacy Converged Infrastructures Transformation to ECS/EKS can be containerized cannot be containerized Application Modernization Framework
- 16. Wait?! You can manage VMs with cloud-native practices?
- 18. Legacy software
- 19. Legacy software
- 22. Managing VMs with cloud-native practices Server
- 23. Managing VMs with cloud-native practices Server Server Server Server Server Server Server Server Server Server Server Server
- 24. Managing VMs with cloud-native practices Year 2000 Server
- 25. Managing VMs with cloud-native practices Year 2000 Server VM
- 26. Managing VMs with cloud-native practices Year 2000 Server VM VM VM
- 27. Managing VMs with cloud-native practices Year 2000 Server VM VM VM Security Backup- restore Maintenance Load balancing Scalability Monitoring Personnel Skillset Know-how Licenses
- 28. Managing VMs with cloud-native practices Year 2000 Server VM VM VM Infra investment Security Backup-restore Maintenance Load balancing Scalability Monitoring Human investment Personnel Skillset Know-how Licenses
- 29. Managing VMs with cloud-native practices Infra investment Security Backup-restore Maintenance Load balancing Scalability Monitoring Human investment Personnel Skillset Know-how Licenses Year 2000 Server VM VM VM Year 2015 Server Infra investment Security Backup-restore Maintenance Load balancing Scalability Monitoring Human investment Personnel Skillset Know-how Licenses Containers
- 30. Managing VMs with cloud-native practices Year 2015 Server VM VM VM Server Containers
- 31. Managing VMs with cloud-native practices Year 2025 Server Containers
- 32. Managing VMs with cloud-native practices Year 2025 Server Containers
- 33. Managing VMs with cloud-native practices Year 2025 Server Containers
- 34. Managing VMs with cloud-native practices Server Containers
- 35. Managing VMs with cloud-native practices Server Container s
- 36. How?
- 37. Managing VMs with cloud-native practices Converged infrastructure Run legacy VMs in containers for significant reuse, reap benefits at enterprise scale • Modernization: Alternative “VMware Exit” strategy • Efficiency: 30-70% reduction in number of servers • Lean stack: 20-60% licensing cost optimization • Unification: Improved team performance
- 38. Q&A?
- 40. deWindowsification 40 This process involves migrating applications from Windows-dependent environments to more flexible, cost-effective platforms, such as Linux, to reduce licensing costs and improve scalability.
- 41. Why deWindows? ● Licensing and Infrastructure(OS/Container footprint) Costs ● Performance ● Security
- 42. Modernize what? ● .NET .NET8+ → ● SQL Server SQL Linux, MySQL/PostgreSQL → ● IIS nginx →
- 43. 43
- 44. How Amazon Q can contribute?
- 45. 1
- 46. 1
- 47. 1
- 48. 1
- 49. 1
- 50. 1
- 51. Q&A?
- 54. Why modernize? ● Licensing Costs
- 55. 55
- 56. 56
- 57. Don't Force Everything into RDBMS 57 What is RDBMS? Cost Implications: ● RDBMSs often scale vertically (more powerful single node = more $$$) ● High write/read throughput = increased IOPS = higher cloud bill ● Complex joins and ACID operations are resource-intensive When NOT to use RDBMS: ● Non-relational data? → Use DocumentDB ● Time-series data? Use time-series DBs like Amazon Timestream. → ● Graph relationships? Use graph DBs like Amazon Neptune. → ● Key-Value lookups? Use DBs like Redis. →
- 58. Managing VMs with cloud-native practices How : Into the cloud Swap pieces Optimize gains Don’t leave DB behind Modernize DB in Lockstep Database Freedom 1. SQL Server to Aurora PostgreSQL 2. 1500+ Stored Procedures 3. NoSQL decompositions Gearing up for further serverless and microservices 15x increase in release frequency
- 59. Q&A?
- 61. Converged Infrastructures Database Freedom Application Modernization Framework Windows to Linux Refactor the Monolith Event Storming Event Storming
- 62. Micro refactoring? ● No architectural changes
- 64. Stateless vs Stateful 64 ● Stateful applications (e.g., .NET keeping state, Kubernetes StatefulSet) rely on PVs and PVCs that need to remain attached. ● Stateless applications can be restarted on any instance, making them ideal for auto-scaling and spot cost savings. ● Stateless = Resilient + Scalable + Cost- efficient on spot instances.
- 65. Optimize Cache Layers 65 AWS CloudFront AWS API Gateway Caching HTTP Static Content Caching DAL Caching /get Redis DB Caching
- 66. Micro Refactoring? ● 12factor ● Stateful Stateless → ● Optimize/Introduce cache layers
- 68. Splitting the Monolith does not have to be Microservices!
- 70. %99+ of the global applications are synchronous!
- 71. ACID vs BASE ● ACID: More DB transactions Higher IOPS More → → compute Higher AWS bill → ● BASE: Async processing + Queues + Caching Reduced → load Lower bill → ○ https://github.com/Splitet/SplitetFramework
- 73. Don't Force Everything into RDBMS 73 What is RDBMS? Cost Implications: ● RDBMSs often scale vertically (more powerful single node = more $$$). ● High write/read throughput = increased IOPS = higher AWS bill. ● Complex joins and ACID operations are resource-intensive. When NOT to use RDBMS: ● Time-series data? Use time-series DBs like Amazon → Timestream. ● Graph relationships? Use graph DBs like Amazon → Neptune. ● Key-Value lookups? Use DynamoDB, Redis. →
- 74. non-ACID(Eventual Consistency) examples? Twitter likes/timelines Facebook news Amazon Cart/Payment
- 75. Consistency Stock = 1 Stock = 1 Replication Checkout Product Page
- 76. Inconsistency Stock = 0 Stock = 1 Replication Checkout Product Page
- 77. Inconsistency Stock = 0 Stock = 1 Broken link! Checkout Product Page
- 78. Consistency Availability Partition Tolerance N/A CA CP AP RDBMS CP AP CAP Theorem
- 79. “Availability over consistency is not a technical choice, it is a business choice.”
- 80. Q&A?
- 82. Application Modernization Framework Refactor the Monolith Microservices
- 84. Microservices is not Docker/Kubernetes
- 85. Microservices is not just smaller services
- 87. Microservices is a new data model
- 88. What if you have a Monolith?
- 89. Splitting the Monolith with Strangler-Fig Pattern
- 92. Q&A?
- 94. Application Modernization Framework Refactor the Monolith CQRS Event Sourcing
- 95. CQRS Command–query separation Read DB Write DB UI / API
- 96. CQRS Command–Query Responsibility Segregation Ref: https://altkomsoftware.pl/en/blog/cqrs-event-sourcing/
- 100. Entities & Events Account 1, +50 Account 1, +150 Account 2, +300 Account 1, -100 Account Number Balance Description 1 100 Bank A 2 500 Bank B …. Table: Account Events: Account Account 2, +200
- 101. CQRS with Event Sourcing Ref: https://altkomsoftware.pl/en/blog/cqrs-event-sourcing/
- 103. CQRS opens the path to Event- Sourcing
- 104. Event-Sourcing Recap: ● Every state change recorded as an event ● The event store becomes the principal source of truth ● State = rebuild using events ● Advantages: ○ No additional Audit Logs, Event are audit logs! ○ Highly-scalable!
- 106. Challenges of Event Sourcing ● Operational Consistency ● Configurable Snapshot Policies & Mechanisms ● Standardized Rollback Mechanisms ● Concurrency Resolving ● Topology Design ● Data Migration
- 107. Q&A?
- 108. YES! You can manage VMs with cloud-native practices! “With Amazon Q Developer transformation, we saved about 40% of our costs. With AWS Graviton-based processors, we can save an additional 10%.” Marc Pearce Head of Cloud Operations, Intelliflo
Editor's Notes
- #19: VM dependencysi olan da aynı
- #86: Servislerin DB uzerinden haberlestigi yapilara Monolith diyoruz
- #90: Incrementally migrate a legacy system by gradually replacing specific pieces of functionality with new applications and services. As features from the legacy system are replaced, the new system eventually replaces all of the old system's features, strangling the old system and allowing you to decommission it.