Amazon AWS Presentation Outline

• Cloud Services
  • Amazon: Amazon Web Services (Hadoop, Linux)
  • Google: AppEngine (Python + Java)
  • Microsoft: Azure (Visual Studio, .Net, VB, C#)
  • many more (http://groups.google.com/group/cloud-computing/web/list-of-cloud-platforms-providers-and-enablers?pli=1)

• Why Amazon AWS
  • easier for Universities
  • supports hadoop
  • offers Linux VM
  • storage + computing = utility: pay as you go

• Amazon's customers

(from http://techcrunch.com/2008/04/21/who-are-the-biggest-users-of-amazon-web-services-its-not-startups/ )

• • Some startups
  • Web services based on hosting large collections of photos
  • banks departments
  • large pharmaceutical companies
  • large companies

• Amazon's services
  • Elastic Compute Cloud (EC2)
  • Simple Scalable Storage (S3)
  • Elastic MapReduce (Hadoop)
  • all accessible from Web GUI and 3rd Party

• EC2
  • "Amazon EC2 reduces the time required to obtain and boot new server instances to minutes, allowing you to quickly scale capacity, both up and down, as your computing requirements change. "
  • Start up the server you need
  • Pick the # of cores, the RAM, and the disk storage
  • Pick the OS
  • Elastic IP addresses
  • several architectures: The unit used is 1 ECU = 1.0-1.2 GHz 2007 Opteron or 2007 Xeon processor.
    • Small Instance: 1 32-bit ECU, 1.7 GB Ram, 160GB disk (32-bit instance)
    • Large Instance: 4 64-bit ECU, 7.5 GB Ram, 850 GB disk
    • Extra-Large Instance: 8 64-bit ECU, 15 GB Ram, 1.7 TB disk
    • High-Memory Extra-Large Instance, 6.5 ECU, 17.1 GB Ram, 420 GB disk
    • High-Memory Double Extra-Large Instance, 13 ECU, 34.2 GB Ram, 850 GB disk
    • High-Memory Quadruple Extra-Large Instance, 26 ECU, 68.4 GB Ram, 1.6 TB disk
    • High-CPU medium Instance, 5 ECU, 1.7 GB Ram, 350 GB disk (32-bit instance)
    • High-CPU Extra Large Instance, 20 ECU, 7 GB Ram, 1.6 TB Disk.

• S3
  • Write, read, and delete objects containing from 1 byte to 5 gigabytes of data each. The number of objects you can store is unlimited.
  • Objects are redundantly stored on multiple devices across multiple facilities in an Amazon S3 Region.
  • Simplified pricing: $0.15 per GB/month, stored or downloaded.

• Elastic MapReduce
  • Hadoop on top of EC2 and S3
  • Store data in S3
  • Start Job Flow of some number of instances in Elastic Map Reduce
  • Logs available
  • EC2 instances shut down automatically once Job Flow is over.
  • Supports Karmasphere Studio for Hadoop IDE