architecture challenges in cloud computing
DESCRIPTION
Session Presented @IndicThreads Cloud Computing Conference, Pune, India ( http://u10.indicthreads.com ) ------------ The Java EE 6 platform is an extreme makeover from the previous versions. It breaks the “one size fits all” approach with Profiles and improves on the Java EE 5 developer productivity features. It enables extensibility by embracing open source libraries and frameworks such that they are treated as first class citizens of the platform. NetBeans, Eclipse, and IntelliJ provide extensive tooling for Java EE 6. But how can you leverage all of this on a cloud ? GlassFish v3, the Reference Implementation of Java EE 6, can easily run on multiple cloud infrastructures. This talk will provide a brief introduction to Java EE 6 and GlassFish v3. The attendees will learn how to create a simple Java EE 6 sample application and deploy them on GlassFish v3 running locally. Then it will deploy that sample using Amazon, RightScale, Joyent, and Elastra cloud infrastructures. It will also show how servers are dynamically provisioned in some environments to meet the demand. The talk will also explain the advantages of each approach enabling you to choose the optimal strategy for your environment. Takeaways from the session The attendees will be able to learn how to deploy a Java EE 6 application in different cloud environments. They’ll also learn about the pros/cons of these infrastructures.TRANSCRIPT
1
Architecture Challenges in Cloud Computing
Prabodh Navare SAS
2
Introduction -
Prabodh Navare
Solution Architect – Manufacturing
About SAS -
Leader in Business Analytics
$2.3 billion revenues
SAS R&D Pune
Magarpatta
3
Cloud apps vs. In-premise apps
• Exceptional cost saving.
• Exceptional fast deployment.
• High performance is an expectation.
4
#1 Design for
Auto-Scaling
5
Linear Scaling
• Elastic design• Parallelization of tasks• In-memory execution• Caching
6
In an average computer, it takes the CPU approximately 200ns (nanoseconds) to access RAM compared to 12,000,000ns to access the hard drive.
This is equivalent to what's normally a 3 1/2 minute task taking 4 1/2 months to complete!
7
#2Design for
High Performance
8
Lamp Stack
(Linux, Apache, MySQL and PHP)
9
Lamp Stack (Current)
(Linux, Apache, MySQL and PHP)
• Memcache
• Hadoop
10
#3Design for
Failover
11
Command Pattern for Failover
12
Next…
Is customer Lock-in good or bad?
13
#4Design for
Data Portability
14
Standards
• Dataportability.org
• ISO/TS 8000-110:2008
15
#5Design for
Pay-as-you-go
16
Costing of my Service ?
Amount of shared memory used, tech support levels, CPU cycles, hard disk space, bandwidth used, electricity, the ROI for the customer etc…
17
Pay-as-you-go
• Support a billing system with stats of user pattern.
• Design different flavors of service.
18
Architecture Challenges!!
#1 Design for Auto-scaling
#2 Design for High performance
#3 Coding for failure
#4 Design for Data portability
#5 Design for Pay-as-you-go
19
Thank you!