Inaugural address manjusha - Indicthreads cloud computing conference 2011
2 likes724 views
The document discusses the cloud ecosystem's evolution and its significance for businesses, emphasizing trends in cloud computing and its projected market growth. It highlights factors driving cloud adoption, such as cost, flexibility, and technology support, while also presenting a case study on developing a platform for managing competitions. The speaker, Manjusha Madabushi, underscores the importance of leveraging cloud capabilities to enhance competitiveness in the market.
Inaugural address manjusha - Indicthreads cloud computing conference 2011
- 1. The Cloud Ecosystem Its more than just services Manjusha Madabushi Talentica Software 1
- 2. About Me Co-Founder and CTO of Talentica Software At Talentica, we help companies transform their ideas to successful products. We specialize in building scalable products using cutting edge technologies Clouds -Amazon, Rackspace and Microsoft Azure Technology areas – NLP, Analytics , mobile Databases – Vertica, Mongo DB, Hbase, Cassandra on Amazon BI – Microsoft (SSAS), Mondrian, jaspersoft 2
- 3. Have you used a cloud before? For Developing Apps? IT infrastructure? Personal? 3
- 4. Ist Indicthreads conference 5-10 participants Around 3-5 participants < 5 participants 4
- 5. Trends Gartner puts cloud computing as the top of the list growth areas to watch for in 2011 5
- 6. Hosting.com Cloud Computing is exploding “...software vendors can no longer simply push customers to their products; rather, vendors’ products need to be where their customers want to be—in the cloud.” 6
- 7. How big is big? G lobal Market value - estimated to be $241 billion by 2020 as compared to 40.4 billion in 2010 –Forrester Research 7
- 9. Cloud Models 9
- 10. Main factors • Cost – still the main factor • Business continuity and Disaster Recovery ( new in 2011) Automatic Failover, multiple data centers • Easy migration – Quick Deploy Amazon Elastic Beanstalk • Flexibility • Elasticity • Technology / Tools support 10
- 11. Flexibility Flexibility is the key • Can use any language or programming model • Use any service individually or together Amazon email services Cloudera – Hadoop clusters • Pay as you use - Use as little or as much as you need and pay for that On- demand instances, reserved instances, spot instances NY times wanted to convert about 11 million articles from TIFF to PDF ( 4TB ) data 100 EC2 instances running for 24 hours All data on S3 – Hadoop, iText, JetS3t • Use existing management tools or infrastructure Private Clouds 11
- 12. Elasticity Is the fundamental property of the cloud • Turn off anytime – stop paying • Let apps sleep during off peak hours Shrinks itself when there are no users accessing it • Follow the sun – turn on servers at different locations based on the users • Automation Provision and attach a 1 TB storage in 2min from the back of an auto rickshaw 12
- 13. Technology and tools Your Application Libraries and SDK’s Web Interface Tools Command Line Tools to access Heroku, Cloud Foundry Search, Eclipse Services plugins Authentication and Authorization Monitoring Deployment and Automation Cross platform Amazon MFA RightScale features Cloud trust Bean Stalk Compute Comp Parallel Processing Payments Content Delivery Messaging Email High Level ute Hadoop - MapReduce Amazon FPS Cloud Front SQS Amazon building blocks Paypal SES Compute Storage Network Database Low Level Auto Scaling Elastic Storage VPC DB Clusters building blocks Mongo DB, Cassandra Global Physical Infrastructure 13
- 14. Case Study Platform for managing competitions 1. Creating, running and tracking competitions 2. User generated content – videos, audio, blogs, etc. 3. Complex competition mechanics 4. Winner selection based on reviews / ratings internal and external to the system 5. Award points to each participant by the engagement value, sentiment on the web and community influence 6. Analytics to measure the performance of each of the competitions 14
- 15. What can you achieve? Virality Web Portal Community UGC Influencer Ratings / Engagement Comments Competition Value Mechanics External Admin Portal Ratings / Sentiment Collaboration 15
- 16. Typical stack Build a custom platform using MVC Model Vertica MongoDB + HBase Separate at Design Time View Ruby on Rails Separate at Deploy Time HTML 5/ CSS + Mobile Baked together at Run Controller Competition mechanics Time BI – Vertica, Mondrian, Jasper ETL 16
- 17. Deployment Load balance Mongo DB Ruby on Ruby on Rails(FE) Ruby on Rails(Admin) Rails(reports) S3 Queue Vertica CM – java BI - Analytics Crawlers 17
- 18. MVC in the cloud Model View Controller Third parties Data Anywhere in any form operating on Many Delivery data Channels 18
- 19. Model = data Data could be internal or external Internal External •User uploaded content •Shared on Youtube, Facebook, twitter •Comments / Ratings by •Comments / Likes by users friends •Reviews by users •Blog posts and comments 19
- 20. View = Channels Content could be viewed • Web portal • Iphone Application • Android application • Kiosk 20
- 21. Controller = Business Logic The core business logic • Competition mechanics • Analytics Engagement, Sentiment, community influence • User Maintenance and Reporting 21
- 22. Internet as the Operating System of Services 22
- 23. Cloud as a platform Messaging S3 Queue Analytics Vertica, Cloud Metrics Conversations Data Management Cloudera Infrastructure Infrastructure & Management 23
- 24. What does it mean • Focus on the core Competition mechanics and Analytics measures • Leverage the Cloud Ecosystem • Companies are successful when they: Focus on the core competency Proactively manage the interfaces with other parts of the value chain 24
- 25. Today you can be wildly successful as a full stack (M+V+C) Tomorrow you will need to make yourself indispensable in the value chain – otherwise someone may eat your lunch. 25
- 26. Become a platform? Go the API way 26
- 27. Thank You 27