From Zero to Cloud in 30 minutes
Download as PPT, PDF8 likes930 views
Simone Brunozzi, an AWS technology evangelist, gave a presentation titled "From Zero to Cloud in 30 minutes" about architecting applications in the AWS cloud. Some of the key points discussed included: 1. Designing applications to be fault-tolerant and avoid single points of failure using AWS tools like availability zones and EBS. 2. Using loose coupling between application components to improve scalability and allow components to scale independently. 3. Designing applications to dynamically scale in response to changes in usage or failures. 4. Leveraging AWS security features like security groups and encryption to build security into applications from the start.
From Zero to Cloud in 30 minutes
- 1. From Zero to Cloud in 30 minutes Simone Brunozzi - AWS Technology Evangelist Europe
- 2. A presentation is not an obligation. Seth Godin It's a privilege.
- 3. Pleased to meet you Technical background Loves presenting Loves traveling Loves people Simone, Italian, AWS technology Evangelist Twitter: @simon (tag is #next09 ) Linkedin Facebook Dopplr Friendfeed Amazon Flickr SmugMug
- 4. Next09, May 6th, 2:00pm 01 - Audimax: "When Money Talks", Jeff Jarvis, Humair Haque 02 - K2: "Social Media", Gary Hoff 04 - K4: "Startup presentations", Victor Henning, Daniel Shaffeld What is happening right now? 03 - P1: "From zero to Cloud in 30 minutes", Simone Brunozzi
- 6. Share Economy Martin Weitzman (1986): the more we share, the better
- 7. Examples of "sharing"... The Merchant business, the "Risky Bet" Amazon Business: Retail Merchants Developers (AWS) Amazon Web Services: > 540,000 registered users > 52 Billion objects
- 8. Cloud Computing To help you share your ideas with the world
- 9. The one slide sales pitch And yes: I hate bullet points :) Amazon S3: durable storage on internet Amazon CloudFront: Content Delivery Service Amazon EC2: virtual servers on demand EBS for EC2: persistent storage Public Data Sets: Human Genome, Census, Science, etc Amazon SQS: messaging system Elastic MapReduce: instant Hadoop cluster http://aws.amazon.com
- 10. La Pedrera - Casa Milà, Barcelona - Antonio Gaudi Architecting in the Cloud
- 11. Architecting in the Cloud 1. Design for failure 2. Loose coupling sets you free 3. Design for dynamism 4. Security is everywhere 5. Don't fear constraints 6. Many storage options 7. AWS ecosystem and community I share lessons learned at Amazon.com
- 12. 1. Design for Failure "Everything fails, all the time" Werner Vogels , CTO Amazon.com and nothing will really fail Avoid single points of failure Assume everything fails, and design backwards
- 13. Design for Failure with AWS Tools to make your life easier Elastic IP Availability Zones (AZ) Elastic Block Store (EBS) Real time monitoring Heartbeat, Linux-HA, NFS, RAS: Reliability Availability Serviceability, Beowulf ZFS, BTRFS, cluster file system, AndrewFS, CODA, Cluster Resource Manager
- 14. 2. Loosely Coupled Systems "Low, Loose, Weak" beats "High, Tight, Strong" Loosely Coupled in time (message oriented middleware) Loosely Coupled in format (data transformation) In Web Services: implementation is hidden from the caller Everything is a Black Box De-coupling for Hybrid models Load-balancing clusters Better scaling SQS prevents failures
- 15. 3. Design for Dynamism No assumption on health, location Bootstrap, dynamic configuration Management components to scale Architect for change Panta Rei (everything flows, by Heraclitus Simplicius) Scaling-out and scaling-in will change servers you talk to
- 16. 4. Security is everywhere Physical is free Network is easy The rest can be added Use it Security groups (EC2 cluster) and IP ranges Group-based rules to control access between App layers Encrypt S3, data transfer, file systems (efs)
- 17. 5. Don't fear constraints More RAM? Shared distributed cache Architectural constraints? Don't embrace, break them Better DB performances? Multiple read-only / sharding / DB clustering Your server is better? EC2 on demand. Static IP? Boot script for software reconfiguration from SimpleDB
- 18. Example: memcached Distributed Memory Cache System ( mem-cache-dee ) No auth/security; DB speedups with read reduction f get_x (int id) { result = db_sel("SELECT * FROM users WHERE id = ?", id); return result; } function get_x (int id) { result = memcached_fetch("userrow:" + id); if (!result) { result = db_sel("SELECT * FROM users WHERE id = ?", id); memcached_add("userrow:" + id, result); } return result; }
- 19. 6. Many Storage options Amazon S3: large static objects Cloudfront: distribution SimpleDB: simple data indexing/quering Amazon EC2: local disc drive Amazon EBS: persistent storage Lessons from Amazon.com
- 20. 7. AWS community and Ecosystem AWS Ecosystem AWS Community Find help, guidance, assistance when you need it
- 21. AWS community and Ecosystem An example?
- 22. AWS community and Ecosystem An example?
- 23. AWS community and Ecosystem An example?
- 24. Warning! Ugly diagram ahead
- 26. Architecting in the Cloud Did you like it? 1. Design for failure 2. Loose coupling sets you free 3. Design for dynamism 4. Security is everywhere 5. Don't fear constraints 6. Many storage options 7. AWS ecosystem and community
- 27. Twitter: #next09 @simon Did you like it?
- 28. Simone Brunozzi - simoneb@amazon.lu - Twitter: @simon Thank You! http://bit.ly/aws06