Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.

MBrace: The Past, The Present and The Future

426 views

Published on

Talk delivered as part of the F# Exchange 2017, https://skillsmatter.com/conferences/8053-f-sharp-exchange-2017#program

You live in a time when it's never been easier to write distributed systems capable of running across hundreds or potentially thousands of machines in the cloud. For many people, however, the goto solutions all lie on the JVM but the CLR and F# has a secret weapon in the form of MBrace. In this talk, you will discover how MBrace fits into the distributed computing space and what features it offers above and beyond some of it's closest competitors. You'll also see how MBrace was used on a project, breaking down some aspects which went well and some aspects which could have gone better. Finally you'll see how MBrace is set to evolve through 2017 and some of the new features which are being brought to the platform.

Published in: Software
  • ⇒ www.WritePaper.info ⇐ is a good website if you’re looking to get your essay written for you. You can also request things like research papers or dissertations. It’s really convenient and helpful.
       Reply 
    Are you sure you want to  Yes  No
    Your message goes here
  • Be the first to like this

MBrace: The Past, The Present and The Future

  1. 1. MBRACE THE PAST, THE PRESENT AND THE FUTURE ANTHONY BROWN COMPOSITIONAL IT @BRUINBROWN93
  2. 2. INTRODUCTION ABOUT ME ▸ Consultant at Compositional IT specialising in F#, IoT and distributed systems ▸ Writing F# for ~ 4 years ▸ Author of Reactive Applications with akka.Net
  3. 3. COMPOSITIONAL IT FUNCTIONAL FIRST, CLOUD READY @COMPOSITIONALIT
  4. 4. THE PAST MBRACE: THE PAST, PRESENT AND FUTURE
  5. 5. WHAT IS MBRACE?
  6. 6. MBRACE IS: 1. A RUNTIME FOR DEPLOYING DISTRIBUTED APPLICATIONS
  7. 7. MBRACE IS: 2. A LIBRARY PROVIDING ABSTRACTIONS FOR MULTIPLE CLOUD PROVIDERS
  8. 8. MBRACE IS: 3. ABSTRACTIONS FOR PERFORMING DISTRIBUTED COMPUTATION
  9. 9. THE PRESENT MBRACE: THE PAST, PRESENT AND FUTURE
  10. 10. HOW ARE WE USING MBRACE?
  11. 11. THE PROBLEM
  12. 12. CUSTOM VALIDATION ENGINE FOR A SPECIFIC DOMAIN
  13. 13. THE SPECTRUM OF SOLUTIONS
  14. 14. AN EXISTING C# CODEBASE 1.
  15. 15. DISTRIBUTE WITH MBRACE 2.
  16. 16. MIGRATE TOWARDS CLOUDFLOW 3.
  17. 17. WHAT WENT WELL?
  18. 18. EASE OF DEVELOPMENT
  19. 19. EASE OF DISTRIBUTION
  20. 20. INTEGRATION WITH THE REST OF THE F# ECOSYSTEM
  21. 21. WHAT WENT LESS WELL?
  22. 22. SERIALIZATION OF C# CLASSES
  23. 23. EXECUTION IS COUPLED TO THE DATA
  24. 24. THE FUTURE MBRACE: THE PAST, PRESENT AND FUTURE
  25. 25. WHAT’S NEXT FOR MBRACE?
  26. 26. 2 AVENUES FOR EXPANSION
  27. 27. AWS MBRACE FABRIC - CLOUD WORKFLOWS CLOUDFLOW SQL GCP ?AZURE STREAMING GRAPHS ML
  28. 28. 1. BUILD DOWNWARDS
  29. 29. NEW CLOUD AND HOSTING PLATFORMS
  30. 30. AZURE CLOUD SERVICES
  31. 31. AZURE APP SERVICE
  32. 32. AZURE RESOURCE MANAGER DEMOS
  33. 33. AZURE VIRTUAL MACHINES
  34. 34. AZURE SERVICE FABRIC
  35. 35. OTHER PROVIDERS: GOOGLE CLOUD PLATFORM?
  36. 36. OTHER PROVIDERS: APACHE MESOS?
  37. 37. 2. BUILD UPWARDS
  38. 38. ABSTRACTIONS FOR COMMON BIG DATA WORKLOADS
  39. 39. THE FUTURE SQL ▸ Ease of exploration of data stored in MBrace storage ▸ Uses full SQL language features ▸ Opens MBrace to a broader range of stakeholders interested in the business data
  40. 40. MBRACE.SQL DEMOS
  41. 41. WEDNESDAY MAY 24TH A DEEP DIVE INTO MBRACE.SQL
  42. 42. THE FUTURE GRAPHS ▸ A lot of big data problems involve graphs ▸ Problems involving graphs are hard to distribute ▸ MBrace.Graph aims to solve the problem through a Pregel like graph API ▸ Still in early incubation stage with an evolving API ▸ Get involved at https://github.com/mbraceproject/ MBrace.Core/pull/172
  43. 43. THE FUTURE MACHINE LEARNING ▸ Machine learning is becoming more and more important ▸ Machine learning in the small is easy ▸ Distributing a machine learning algorithm is harder ▸ Potential for a SparkML style API on top of MBrace CloudFlows ▸ Early analysis stage of development
  44. 44. THE FUTURE MBRACE.STREAMING ▸ Provide a high throughput IoT streaming platform ▸ Abstraction similar to CloudFlow tailored towards temporal streaming data ▸ Abstraction over IoT solutions for AWS and Azure ▸ Early analysis stage of development
  45. 45. MBRACE IN 2017
  46. 46. A PLATFORM FOR DISTRIBUTED DATA SCIENCE IN .NET
  47. 47. A PLATFORM FOR CROSS- CLOUD DEVELOPMENT
  48. 48. WHAT’S NEXT?
  49. 49. OPEN ROADMAP DISCUSSION ON GITHUB
  50. 50. COMMUNITY CONTRIBUTES TO DISCUSSION
  51. 51. CONTRIBUTE WHERE YOU CAN -CODE -DOCUMENTATION -ADVICE
  52. 52. HTTPS://GITHUB.COM/ MBRACEPROJECT
  53. 53. COMPOSITIONAL IT ANTHONY@COMPOSITIONAL-IT.COM FUNCTIONAL FIRST. CLOUD READY.
  54. 54. Q&A.

×