Kb12012011 amitava cloud_computing

Agenda

 Cloud computing – a brief history
 Solving scalability problem
 Features of a real cloud computing service
 Cloud service models
 Success stories
 Case study
 Cloud computing myths
 How to get started

Gartner's definition

 Cloud Computing:
A style of computing in which elastic IT-
enabled capabilities are delivered as a
service to external customers using
Internet technologies.

 Gartner, 2008

Looking back

1961 – John McCarthy’s vision

1999 – Salesforce

2002 – Amazon Web Services

2006 – Amazon Elastic Compute Cloud

2009 – Browser based cloud enterprise
applications (Google Apps, Windows Azure)

Let me start with a tough question

Where did it all begin?
 Scalability problem

Scalability problem

 Characteristics:
 Definitely data-intensive
 May also be processing intensive
 Examples:
 Crawling, indexing, searching, mining the Web
 “Post-genomics” life sciences research
 Other scientific data (physics, astronomers, etc.)
 Sensor networks
 Web 2.0 applications
 …

How much data?
 Wayback Machine has 2 PB + 20 TB/month
(2006)
 Google processes 20 PB a day (2008)
 “all words ever spoken by human beings” ~ 5 EB
 NOAA has ~1 PB climate data (2007)
 CERN’s LHC will generate 15 PB a year (2008)

640K ought to be
enough for anybody.

Some classic ignorance
 August 2008, USA - Hurricane Gustav : The surge of Obama
dollars crashed the Red Cross website in less than 15 minutes.
 Feb. 21, 2011, India - ICC Cricket World Cup : Furious cricket fans
slammed organizers of the World Cup as the official ticketing
website crashed amid a scramble for 1,000 tickets available for
the final
 June 2009, China - Government website for reporting corrupt
officials : crashed shortly after launching
 May 2009, India - World's largest Democracy’s election : The
election commission unveiled a brand new website for providing
real time results of the mega-poll which was well prepared to
handle 80.64 billion hits in 8 hours (2,800 hits/second). Guess
what happened? On the election result day media reported:
300,000 hits/second made Election Commission website crash.
 October 2009, France - A website launched by French first lady
Carla Bruni has crashed on its first day - overwhelmed by the
number of users trying to access it at once.

Agenda

 Cloud computing – a brief history
 Solving scalability problem
 Features of a real cloud computing
service
 Cloud service models
 Success stories
 Case study
 Cloud computing myths
 How to get started

Predicting Infrastructure Needs

Actual Usage

Customer
Dissatisfaction
Compute Power

Predicted
Usage

Waste

Time

3) No initial investment (No CapEx)

4) Pay as you go, pay for what you use

5) No Infrastructure maintenance

The 70/30 switch

30% 70%

On-Premise Your Managing All of the
Infrastructure Business “Undifferentiated Heavy Lifting”

Configuring
More Time to Focus on
Cloud-Based Your Cloud
Your Business
Infrastructure Assets

70% 30%

Cloud Service models(*aaS)

 IaaS (Infrastructure as a Service)
 Utility computing
 Why buy machines when you can rent cycles?
 Examples: Amazon’s EC2, GoGrid, AppNexus
 PaaS (Platform as a Service)
 Give me nice API and take care of the implementation
 Example: Google App Engine
 SaaS (Software as a Service)
 Just run it for me!
 Example: Gmail

 Source – NIST (National institute of Standards and Technology, U.S.
Department of Commerce) in its Special Publication 800-146

SaaS (Software as a Service)
 Consumer is free of any worries and hassles related to the service.
 Service Provider has very high administrative control on the
application and is responsible for update, deployment, maintenance
and security.
 Example, Gmail is a SaaS where Google is the provider and we are
consumers.
 Best use of SaaS is in productivity and collaboration apps in the cloud
likeGoogle Apps, Online Project Management like DeskAway
 CRM apps – Impel CRM, Salesforce.com, Microsoft Dynamics
 Storage and Sharing services like Dropbox, Skydrive(windows
live), Amazon S3, Google Docs

PaaS (Platform as a Service)
 Platform where software can be developed, tested and
deployed, meaning the entire life cycle of a software can be
operated on a PaaS.
 Provides everything you need to develop a cloud SaaS application
from the development environment, programming
languages, compilers, testing tools to deployment mechanism.
 Example:
 Microsoft Windows Azure targeting .NET developers
 Beanstalk from Amazon and GAE for Java, Python developers
 Heroku for Ruby on Rails (preferred for Facebook apps development)
 Dotcloud for multilingual application development

IaaS (Infrastructure as a
Service)
 Do you require virtual computers, cloud storage, network
infrastructure components such as firewalls and configuration
services?
 Usage fees are calculated per CPU hour, data GB stored per hour,
network bandwidth consumed, network infrastructure used per
hour, value added services used, e.g., monitoring, auto-scaling etc.
 Very useful for startup companies who don’t know how successful
their newly launched application/website will be.
 Example:
Amazon,Rackspace, GoGrid,Joyent, Rightscale and Terremark

Guess the company

 An online company
 It uses more than 12,000 servers
 They have more than 230 million
users

 What's the name of the company?
 What do they do?

Kb12012011 amitava cloud_computing

Zynga.com

 Creators of the most popular
Facebook games: Farmville, Mafia
Wars, Treasure Isle, etc.

 More than 12,000 servers on AWS

 More than 230 million monthly users

 100% on Amazon IaaS

Netflix

 More than 9 Billion USD market cap
 2 years back had their own data center
but now due to increase in demand they
are Migrating 100% on Amazon Web
Services
 So far migrated: movie lists, website
search, transcoding, recommendations, et
c.
 10 M subscribers, 100k DVD titles

Case 1
 Company : A Risk Management company
 Problem:
 They need to analyze data as soon as the stock exchange closes, to find
mistakes and correct them in time.
 However, this is usually very expensive.

 Traditional solution
 Buy a huge IT infrastructure
 Hire people to manage it
 Negotiate contracts, etc.
 Manage the IT infrastructure
 Every 2-3 years, major upgrade

 Benefits from Cloud Computing:
• No need for investments
• Massive cost savings
• IT staff can focus on product/service
• Higher availability
• Faster time to market

Case 2
 Company : A web startup company
 Problem:
 They want to launch a website/app, but they don't know how successful it will be.
 Need to serve customers worldwide.
 Usually very expensive.

 Traditional solution
 Buy servers
 Buy storage
 Sign a CDN contract (Content Delivery Network)
 Launch website/application
 Manage scaling and provisioning

 Benefits from Cloud Computing:
• No need to buy IT Infrastructure
• Deploy worldwide
• Scale up/down when needed
• Save time
• Focus on the business

Step 1: simple website

 Visitors per month: 10,000

 Monthly costs: Total: $18
 Server: $15
 Bandwidth: $3

Step 2: Adding Cloudfront

 Visitors per month: 100,000

 Server: $15
 Bandwidth: $5
 CDN: $30 (Cloudfront)

Step 3: Register in Web 2.0

 Visitors per month: 1,000,000

 Server: $120
 Bandwidth: $55
 CDN: $200 (Cloudfront)
 Storage: $10 (viz. Amazon S3)
 Elastic Load Balancer: $20

Step 4: Adding a scaling
database like Amazon RDS
 Visitors per month: 10,000,000

 Monthly costs: Total: $1,820
 Server: $240
 Bandwidth: $125
 CDN: $1,200 (Cloudfront)
 Storage: $50
 Elastic Load Balancer: $45
 Database: $160 (RDS)

Myth #1:
"the cloud isn't reliable"
 Answer:
 Multiple Regions (US, EU, AsiaPacific)
 Many companies achieve better
performance in the cloud
 Thousands of metrics are met with the
99.9% outlier

Myth #2:
"I'll lose control of my data"

 Answer:
 You own and control your own data
 You can control where the data is stored
 Cloud Service provider doesn't move your
data
 You can always get your data back

Myth #3:
"Cost is the only Cloud advantage"

 Answer:
 Resource Management is more efficient
 Faster time to market
 Being able to focus on your business
 Agility

How to get started

 Create an Amazon Web Service (AWS)
Account (you can use your Amazon
account)
 Learn the basics (EC2, S3)
 Start a small project
 Review results
 Expand your AWS usage

Rubber meets Road
 Large enterprises are building their own private clouds to resolve
the security problems
 Coding & development skills will take a back seat to project
management, quality assurance testing, business analysis, and
other high-level abstraction thinking
 IT departments will shrink as users go directly to the cloud for IT
resources
 Cloud computing providers will support information security as if
banks secure money
 Small-sized to medium-sized businesses as well as large
enterprises will be run on the cloud
 Large enterprises will become part-time cloud-computing
vendors
 The browser will be all the desktop software you need
 Game services will be one of the most interesting applications of
Cloud Computing

Kb12012011 amitava cloud_computing

Recommended

More Related Content

What's hot (14)

Viewers also liked (18)

Similar to Kb12012011 amitava cloud_computing (20)

Recently uploaded (20)

Kb12012011 amitava cloud_computing

Editor's Notes