1. CLOUD COMPUTING

2. AGENDA
 Overview
 Popular Definitions
 Deployment Models
 Delivery Models – IaaS, PaaS, SaaS
 NoSQL
 Vendor Lock-in
 Security Concerns

3. WHAT IS CLOUD COMPUTING?

4. CLOUD COMPUTING OVERVIEW
 Sharing Resources
 Abstraction of Software Development Details
 Scalability
 High Processing Power
 Reliability / Availability
 Pay – as – you – use
 Peak load performance
 Simplified maintenance:

5. CLOUD COMPUTING DEFINITIONS
 OSSM
 5-3-4

6. DEFINITION : OSSM
 On-demand: the server is already setup and ready
to be deployed
 Self-service: customer chooses what they want,
when they want it
 Scalable: customer can choose how much they
want and ramp up if necessary
 Measureable: there's metering/reporting so you
know you are getting what you pay for

7. DEFINITION : 5 – 3 – 4
 5 Basic Characteristics
 On-Demand
 Ubiquity – (irrespective of location, app should be accessible)
 Location Independent Resource Pooling
 Elasticity
 Pay per Use
 3 Delivery Models
 IaaS
 PaaS
 SaaS
 4 Deployment Models
 Public Cloud
 Private Cloud
 Hybrid Cloud
 Community Cloud

8. DEPLOYMENT MODELS

9. PUBLIC CLOUD
 Most standard Cloud Computing Model
 Same infrastructure/resources are to be used by
other tenants/businesses
 Vendor makes Hardware/Software available as
services over internet
 Pay per use pricing model
 No CAPEX, Only OPEX
 No wasted resources, pay only for what you use
 Examples : Amazon EC2, IBM Blue Cloud, GAE,
Azure

10. PRIVATE CLOUD
 Not truly a cloud
 High CAPEX
 Rather, cloud like on-premise infrastructure with
horizontal scalability,availability,etc
 Or off-shore infrastructure behind a corporate
firewall
 Maintenance can still be outsourced
 Concerns of Security are minimized
 More control over data
 Examples : Amazon EC2 &S3(Simple Storage
Service)

11. HYBRID CLOUD
 Composition of atleast one Private Cloud and
atleast one Public Cloud
 All the involved clouds maybe provided by the
same or different vendors
 Allows scalability of Public Cloud, without exposing
sensitive data
 Use Case :
 Archived Data on Public Cloud, Operational Data on Private
Cloud

12. DELIVERY MODELS
IaaS PaaS SaaS

13. IAAS (INFRASTRUCTURE AS A SERVICE)

14. IAAS (INFRASTRUCTURE AS A SERVICE)
 On Demand Storage & Processing - Computing as
a Service
 Hardware, its
Software, Electricity, Cooling, Connectivity all
managed by Vendor
 Pay-as-you-go
 Difference between Shared Hosting v/s Cloud
Infrastructure
 Flexible

15. IAAS… (CONTD…)
 Provides infrastructure management tools.
 Vendors : Amazon EC2, Rackspace, etc
 Tools are available to monitor performance, peak
load conditions, etc
 Scaling, Monitoring , etc are still a responsibility of
the Service Buyer

16. HOW TO CHOOSE AN IAAS VENDOR?
 Support – Email, Phone, 24 x 7?
 Hardware, its configuration and Performance
 Partnership with Middleware Vendors
 Eg: Amazon EC2 has partnered with
IBM,Microsoft,RedHat, Oracle,etc.
 Licensing
 Windows Licences cannot be migrated from local data
center to Cloud
 IBM Websphere not available with EC2

17. PAAS (PLATFORM AS A SERVICE)

18. PAAS (PLATFORM AS A SERVICE)
 A platform to develop applications that can be
deployed on the cloud
 Framework / Platform is hosted on the Cloud
 Development language and platform is vendor
decided
 Infrastructure management by Vendor
 Scalability and management is provided by the
Vendor
 No Licensing hassles
 Vendors: Force.com, Google App Engine, Microsoft
Azure, Heroku
 Drawback : Vendor Lock in

19. FORCE.COM :
 Very easy to use Web Interface to create a web
application
 Code generation, compilation, deployment, etc happens
on Force.com servers
 Also has Eclipse Plugin based development
 Asks for Force.com credentials
 Code compilation and deployment happens on
Force.com servers, seamlessly
 Development in Apex(Dev) and VisualForce (UI)
 Supports Agile

20. HEROKU
 Development, deployment, scaling
 Heroku’s Ruby Platform lies on AWS
 Used by Facebook for scaling and application
development
 Dev Languages : Ruby on Rails, Java
 Now acquired by Force.com

21. GOOGLE APP ENGINE (GAE)
 SDK for Java and Python
 Provides Eclipse Plugin
 Big Table DataStore
 Simple App Configuration
 Automatic Scaling, No performance Tuning is
required
 Quotas - request count, bandwidth usage, CPU
usage, datastore call count, disk space used, emails
sent, even errors!

22. WINDOWS AZURE
 Windows Azure Tools for Visual Studio
 Technology Stack and Tools :
 Azure tools for Visual Studio
 Azure SDK
 Visual Web Developer (replacement for Visual Studio)
 ASP.NET MVC3 (framework)
 IIS

23. WINDOWS AZURE STACK
 Windows Azure – Cloud OS as a Service
 5 Services – Live Services, SQL
Azure, AppFabric, Sharepoint, Dynamic CRM
 SQL Azure – Cloud based SQL Server
 Azure AppFabric – Collection of Services
(Caching,Service Bus,Integration)
 Azure Platform Alliance (non-MS Datacenters)

24. AZURE DEVELOPMENT PLATFORM
 ASP.NET Framework
 PHP
 SDKs for Java and Ruby help integrate with
AppFabric Services

25. SAAS (SOFTWARE AS A SERVICE)

26. SAAS (SOFTWARE AS A SERVICE)
 Enterprise Application without installation overhead
 Applications that are available to be used over the
internet
 Pay per user Account
 Scalable to multiple users and tenants
 Applications like SalesForce.com, Gmail,Google
Apps,etc

27. DATABASE FOR THE CLOUD

28. PROBLEMS WITH TRADITIONAL DATABASES
FOR CLOUD
 Big Data - Big data are datasets that grow so large
that they become awkward to work with using on-
hand database management tools
 Difficulties include capture, storage, search,
sharing, analytics, and visualizing of vast data

29. NOSQL
 Can service heavy read/write workloads
 Usually Avoid Join Operations
 Scale Horizontally
 Works well even with cheap commodity servers
 Flexible – Schema changes are easy to make
 Examples : Big Table, Mongo
DB, Hadoop, Cassandra, Amazon SimpleDB

30. NOSQL - CHALLENGES
 Quota limits
 Max 1000 records per query
 Query times out in 5 seconds
 Maturity – Nascent compared to RDBMS
 Support – Mostly Open Source. Support driven by
startups with no true global reach
 Administration – Installation & Maintenance skill is
not easily available
 Expertise – Not easily unavailable, which is
unacceptable to businesses

31. VENDOR LOCK-IN
 Definition – Stuck with 1 vendor because of
complexity to move to another vendor
 Scenarios to move out of a Cloud Vendor:
 Cost
 Shut Down of Services
 New choice of Services offered by another Vendor
 How to avoid/minimize vendor lock-in chances?
 While architecting your app, ensure vendor-specific services
are judiciously used
 Evaluate Options
 Check ROI
 Read about upcoming features/services

32. VENDOR LOCK-IN
 IaaS
 Less lock-in
 Data Migration is of moderate difficulty
 PaaS
 Tightly coupled to vendor
 Migration to another vendor may require re-engineering
the application

33. SECURITY IN CLOUD COMPUTING
 Issues:
 Data Integrity
 Recovery
 Network security
 Access and Authentication procedures
 Data encryption techniques
 Tenant isolation
 Most issues are now controlled or resolved
 Still considered unsafe for Financial data

34. USE CASES

35. UNCERTAINTY OF LOAD
 Zynga has both Private and Public Cloud services
at disposal
 New Game launches on Public Cloud
 When usage is stabilized, then migrated to Private
Cloud

36. SHORT-TERM USAGE
 Seasonal Apps
 Event Websites
 OCR to Doc Conversion

37. NETFLIX
 Completely out of Cloud
 Highly fluctuating usage of large multimedia data
 Across different geographical locations
 Speedy Access

38. LIMITATIONS

39. LIMITATIONS OF CLOUD COMPUTING
 Connectivity is mandatory
 Security
 Skills – Development and Administration
 Network Bandwidth
 Not suitable for all businesses