Windows Azure For Architects
1 like424 views
Windows Azure is a cloud computing platform that provides Infrastructure as a Service (IaaS), Platform as a Service (PaaS), and Software as a Service (SaaS). It allows developers to build, deploy, and manage applications through Microsoft-managed data centers. The key services include compute, storage, networking, and development tools. Architectural principles for building reliable cloud applications on Windows Azure emphasize statelessness, redundancy, loose coupling between components, and scaling automatically based on demand.
Windows Azure For Architects
- 1. Windows Azure for Architects Anko Duizer February 6nd 2010
- 2. Introducing… • Anko Duizer – Chief Software architect – Blog: http://www.ankoduizer.nl – Twitter: ankoduizer • TellUs – Lead generation – > 5000 sites – World wide – Microsoft platform
- 3. Introduction
- 4. Cloud computing is Internet ("cloud") based development and use of computer technology ("computing”)
- 5. Cloud Attached platforms IaaS services SaaS (PaaS) Extra Application Cloud ? Application functions Platform On- Browser/ ? Application client Application premises
- 6. Storage Integration Development environment Identity Management Application Infrastructure services Local support Operating system Applications Foundation Application services
- 7. Cloud From on-premises to cloud platforms Cloud infrastructure application services Cloud foundation Cloud application services On-premises On-premises infrastructure application services OS On-premises application services
- 8. Part 1: Windows Azure Platform
- 15. Windows Azure platform AppFabric Applications SQL Azure Windows Azure Live Services Applications Windows Windows 7/ Windows Others Server Vista/XP Mobile
- 16. Application Compute Storage Config Fabric
- 17. HTTP Blobs Tables Queues
- 18. Access Control (ACS) ? Service Bus
- 19. 2) Discover endpoints Service bus Registry 1) Register 3) Access Endpoints endpoints application Application Application Organization X Organization Y
- 20. SQL Data Services Others (future)
- 21. Windows Azure service lifecycle Coding and Provisioning Deployment Maintain modeling goal state - Mapping and deploying to - Monitor New services Desired actual - React to configuration hardware events and updates - Network configuration Developer Developer/ Automated Automated Deployer
- 23. Simplicity
- 25. Hardware fails
- 28. A few design choices
- 31. Lots and lots of commodity machines
- 32. A few challenges
- 33. What do you do about that thing called state?
- 35. Go horizontal Go stateless
- 36. Store state in Windows Azure storage
- 37. How do you deal with unreliable components?
- 39. Default.aspx Worker.cs LB (web role) (worker role) Queue
- 40. Public Internet Web role LB Storage services
- 41. Public Internet Worker role Storage services
- 43. Use Windows Azure queues for separation of work
- 44. How do you deal with varying load?
- 46. How do you deal with failures?
- 47. Use Windows Azure storage for valuable data
- 48. ... Entity Entity Entity ... Property Property Property Name Type Value
- 49. What about the SQL Azure?
- 50. Application Application Browser ODBC, OLEDB, application REST Client SQL Client* ADO.Net PHP, Ruby, … Cloud HTTP+REST HTTP TDS Windows Azure REST (Astoria) Web App Data center ADO.Net + EF SQL Client* TDS + TSQL Model SQL Azure
- 51. Cloud Data in the application cloud On-premises application
- 52. Retry on transient failures But…
- 53. Be idempotent
- 55. Part 4: Managing
- 58. Production Stage Deployment Deployment
- 59. +
- 61. How do I debug the cloud?
- 62. Logging
- 63. How do I get notified when something bad happens?
- 65. Part 4: Useful?
- 66. “On and Off “ “Growing Fast“ Compute Compute Inactivity Period Average Usage Average Usage Time Time On & off workloads (e.g. batch job) Successful services needs to grow/scale Over provisioned capacity is wasted Keeping up w/ growth is big IT challenge Time to market can be cumbersome Complex lead time for deployment “Unpredictable Bursting“ “Predictable Bursting“ Compute Compute Average Usage Average Usage Time Time Unexpected/unplanned peak in demand Services with micro seasonality trends Sudden spike impacts performance Peaks due to periodic increased demand Can’t over provision for extreme cases IT complexity and wasted capacity
- 67. Is it cheaper?
- 68. Compute : $0.12 / hour Storage : $0.15 GB / month Bandwidth: $0.10 GB in, $0.15 GB out
- 69. Retry on failures Recap Loose coupling