pushing enterprise software to the next level
play

Pushing Enterprise Software to the Next Level Self-contained Web - PowerPoint PPT Presentation

Pushing Enterprise Software to the Next Level Self-contained Web Applications on In-Memory Platforms Micha Nosek Starcounter AB Who am I? Micha Nosek Software Engineer, Technical Sales Engineer Starcounter


  1. Pushing Enterprise Software to the Next Level 
 Self-contained Web Applications on In-Memory Platforms Micha ł Nosek 
 Starcounter AB

  2. 
 Who am I? ▪ Micha ł Nosek 
 Software Engineer, Technical Sales Engineer – Starcounter 
 http://starcounter.com 
 ▪ Github: mmnosek 
 LinkedIn: https://www.linkedin.com/in/mmnosek 
 E-mail: michal@starcounter.com 
 Twitter: @mmnosek

  3. On Today’s Agenda 01 Setting the Stage RAM Memory Modern WEB SCS Architecture 02 In-Memory Application Platform Architecture Single App Integration Demo Future

  4. Enterprise Software of Today Micro-Services Monolith ▪ Orchestration ▪ Bad maintainability ▪ Eventual consistency ▪ Long builds ▪ Communication problems ▪ Technology lock-in ▪ Complexity ▪ Long TTM ▪ Poor scalability

  5. Wirth’s law “What Intel giveth, Microsoft taketh away.” “What Andy giveth, Bill taketh away”

  6. On Today’s Agenda 01 Setting the Stage RAM Memory Modern WEB SCS Architecture 02 In-Memory Application Platform Architecture Single App Integration Demo Future

  7. RAM Prices Price of 1MB in USD over time 1000,00000 100,00000 10,00000 1,00000 USD 0,10000 0,01000 0,00100 0,00010 0,00001 1980 1990 2000 2010 2020 Year https://jcmit.net/memoryprice.htm

  8. Conventional In-Memory

  9. Conventional In-Memory

  10. Pros and Cons Cons Pros ▪ Communication isn’t faster ▪ Getting faster ▪ It’s not durable ▪ Better utilised by modern CPUs ▪ Not getting cheaper anymore?

  11. On Today’s Agenda 01 Setting the Stage RAM Memory Modern WEB SCS Architecture 02 In-Memory Application Platform Architecture Single App Integration Demo Future

  12. Pros and Cons Cons Pros ▪ Still not implemented everywhere ▪ Ubiquitous (no native, separate process) ▪ Global scope (one app can break ▪ Semantics (content) vs Presentation something in another) ▪ Modularity as priority (reusability) ▪ Online requirement

  13. On Today’s Agenda 01 Setting the Stage RAM Memory Modern WEB source: 
 SCS Architecture scs-architecture.org 02 In-Memory Application Platform Architecture Single App Integration Demo Future

  14. SCS Architecture

  15. SCS Architecture

  16. SCS Architecture

  17. SCS Architecture System 1 System 2

  18. Pros and Cons Cons Pros ▪ Integration ▪ Modularisation ▪ Common look and feel ▪ Maintainability ▪ Inconsistency ▪ Loose coupling

  19. On Today’s Agenda 01 Setting the Stage RAM Memory Modern WEB SCS Architecture 02 In-Memory Application Platform Architecture Single App Integration Demo Future

  20. 
 In-Memory Application Platform 
 For Building Self-Contained Systems

  21. General Platform Architecture Front-end Framework 
 React, Polymer Communication 
 Palindrom - REST, Web Sockets Application 
 View Models, Entities, App Logic In Memory App Platform Mapping, Persistence, Queries Starcounter

  22. Traditional Stack vs Starcounter Stack

  23. Data Storage • In-Memory database • ACID compliant • Snapshot isolation • Flexible

  24. VMDBMS U.S. Patent No. 8,266,125

  25. Business Logic • Polyglot • Simplified • Platform-agnostic • Real-time

  26. User Interface • Web native • Web socket communication • Design agnostic • Thin

  27. Demo: 
 Simple SCS app

  28. Integration: Data Level

  29. UI A UI B UI C App A App B App C Model A Model B Model C Mapper Starcounter

  30. Integration: UI Level

  31. UI B UI A UI C UI A UI B UI C App A App B App C Model A Model B Model C Starcounter

  32. Outcomes Cons Pros ▪ Integration ▪ Modularisation ▪ Different look and feel ▪ Maintainability ▪ Inconsistency ▪ Loose coupling ▪ Platform lock-in? ▪ Full and easy integration ▪ Common look and feel ▪ Consistency

  33. Storage Engine Benchmark Full-Stack Benchmark ▪ 1.5 mln. accounts, 500 K remote clients transfer. ▪ YCSB load 5% writes, 95% reads. ▪ Money between accounts (5%) and read totals ▪ 1 x E5-2680v2, 1 machine (10/20 cores/ (95%). threads). ▪ Transfer and read operations are mixed randomly. ▪ 8 threads: 3.5 mln. Ops/sec. ▪ Starcounter on .NET (1 x EC2 c3.8xlarge): 1 M ▪ 16 threads: 5.4 mln. Ops/sec. OPS. ▪ MariaDB Galera Cluster 5 nodes with Node.js app ▪ c3.8xlarge – 60 GiB RAM, 32 vCPUs server (5 x EC2 c3.2xlarge, EBS root volume and ▪ c3.2xlarge – 15 GiB RAM, 8 vCPUs high network throughput, stored procedures): 55 K OPS. ▪ https://www.ec2instances.info/ ▪ Ratio suffers for MariaDB doing more writes.

  34. On Today’s Agenda 01 Setting the Stage RAM Memory Modern WEB SCS Architecture 02 In-Memory Application Platform Architecture Single App Integration Demo Future

  35. Currently vs Future Registers Registers Cache Cache Main Memory Main Memory Solid State Disk Non Volatile Memory Solid State Disk Magnetic Disk Magnetic Disk Current Future

  36. Starcounter in the Future

  37. Enterprise ▪ Simplified Software of ▪ Near real-time Tomorrow ▪ Easy to maintain ▪ Reusable/modularised ▪ Fully web-based ▪ Fast data ▪ HTAP or HOAP

  38. THANK YOU! Questions?

Recommend


More recommend