2. Agenda
For “Cloud Computing” FDP Workshop….
Session 1 - Context
All the sessions are in discussion format.
Foundations – Hardware, Software Please feel free to stop the speaker and
1 and Economics initiate a relavant discussion any time during
the session .
Session 2 - Perspectives
Service provider and
2 Service consumer
Perspectives
General Disclaimer:
The views expressed in here are mine and
Session 3 - Challenges they do not represent the views of the
company or the customers I work for…
Challenges in Research
3 Open Discussion
Cloud Computing Prasad Chitta
3. Computing
Before “Cloud” ….. The traditional way
Hardware Software Economics
Servers come in multiple tProprietary software for the tHardware procurement
sizes hardware platform
Software licensing
Typical procurement times Procedure oriented
in weeks to months Support models
Custom built for customer
needs
1 Big-Iron – Mainframe - Server based computing accessed by dumb terminals
2 Client / Server computing – Thick client computing
3 Browser Based, thin client and n-tier computing - Internet Computing
Cloud Computing Prasad Chitta
4. Foundations - Hardware
Grid Computing and Virtualization
• Grid Computing
Parallel Processing Paradigm
Pooling multiple small computing resources look
like a big single computing resource
• Virtualization
Abstraction of underlying detail
Making a big computing resource appear as multiple
smaller resources (Multi-Tenancy)
• Bandwidth availability
Availability of cheap internet bandwidth
Parallel processing using multiple threads right from processor
with multiple cores all the way to servers and then abstraction
of underlying hardware to different sizes using virtualization
gives the first foundation to the ”Cloud Computing”
5. Foundations - Software
Service Orientated Architecture
• Open Standards
Open Standards in software and
Open source software
• Service Oriented Architecture
Computing as a mesh of loosely-coupled ”services”
N- tier architectures
Providing ”something” as a ”service” is the second foundation
to the ”Cloud Computing”
6. Foundations - Economics
CAPEX to OPEX shift
• CAPEX to OPEX shift
No or minimum capical expenditure
• Explosion of DATA
Big Data – Facebook has an average of 240 photos
per user!
• Time to Market
I would have got my idea implemented yesterday!
Businss economic need is the third foundation to the ”Cloud
Computing”
7. Cloud Computing
Finally Defined!
Standardise
Virtualize
Cloud Computing: Provide a self-
Automate service, pay-per use computing facility
that is scalable elastically with reliable
quality of service for the consumers.
Cloud Computing Prasad Chitta
8. Cloud Enablement
Not Cloud Washing… .
Hardware – Blades and grids
Standardise Platform - Open standards
Service Standardization – API, SOAP, REST
Hypervisors
Virtualize Abstraction
Multi-Tenancy
Self Service Provisioning
Automate Metered Usage
Management
Quality of Service
Cloud Computing Prasad Chitta
9. Perspectives
That Matter…
Cloud Service Provider Cloud Service Consumer
Large Enterprise tLarge Enterprise
1 2
Medium or small company Medium or small Company
Individual Contributor Individual consumer
I IaaS - Infrastructure as a service
P PaaS – Platform as a service
S SaaS – Software as a service
Cloud Computing Prasad Chitta
10. Sample Portfolio of a BFS Enterprise
http://www.tcs.com/resources/white_papers/Pages/Cloud-Computing-Strategic-Considerations-for-Banking-and-Financial-Institutions.aspx
14. A “list” of concerns
for
Consumer Provider
• Pricing
• Positioning vis-à-vis existing
offerings
15. Challenges
For research....
1. General Cloud Development
Related
Security & a) Provisioning, Load
Privacy Management
2. Business Models and Interaction
a) Performance, Service Levels
and Quality of Service
Provisioning, Multi-
Metering & Tenancy & 3. Core research
Quality of a) Security and privacy
Management Service concerns
b) Data Handling
Cloud Computing Prasad Chitta
17. Brewer’s CAP Theorem
And its implication on grids…
• Consistency (Atomicity of ACID)
• Availability
• Partition Tolerance
Only any two can be achieved in a given
grid so get on to “BASE”
18. Brewer’s CAP Theorem
And its implications on grids
Cloud Computing Prasad Chita
From: http://www.julianbrowne.com/article/viewer/brewers-cap-theorem
19. BigQuery powered by Dremel
From Google, taking map-reduce to next stage
http://research.google.com/pubs/papers.html
Your Data BigQuery
1. Upload 2. Process 3. Act
Cloud Computing Picture from: http://cloud.berkeley.edu/data/dremel.pptx Prasad Chita
20. THANK YOU!
Reach me on:
LinkedIn: http://in.linkedin.com/in/prasadchitta
Blog: http://technofunctionalconsulting.blogspot.com