Axa Assurance Maroc - Insurer Innovation Award 2024
Cloud Computing for Enterprise Architects
1. Cloud Computing
16 Dec 2011
jean-francois.caenen@capgemini.com
Chief Technology Officer
Capgemini France
2. Cloud Computing Controversy
Simply the most
The latest
awesome
fashion
paradigm shift
in a fashion-
in the last 50
driven industry
years of IT
Same old stuff
New disruptive
from last
Internet Technologies
generation Enterprise
Cloud and Business Cloud
repackaged for
Models
this generation
2
3. CFOs
love
Cloud Computing !
No investment before use.
Pay for what you need, when you need.
Instead of investing in rigidity, you buy flexibility.
3
4. Business
managers
know & love
Cloud Computing!
Implement
Business Applications we need
much quicker
Source:
HfS Research and London School
of Economics, Oct 2010
846 Entreprises
4
5. A Major Shift in Solution Sourcing
From This To This
Experimenting
with Cloud
Cloud-based
Some Open Source Services
(in commoditized or wherever
embedded stacks) viable
Mostly Open Source
Commercial Continual Software
Software reappraisal (on IaaS or PaaS)
on-premise
Commercial SW due to lack of
on-premise viable alternatives,
only where needed or for risk
mitigation
5
6. 100 000 users
June 2011
Salesforce.com expects $3 Billion Run Rate next Year
in 5 years 1Billion $
Revenue per year
Start
2006 2011
6
9. By 2015, 50 % or more of NEW enterprise IT spend
will be Cloud-based or Hybrid.
By 2015, 65 % or more of NEW enterprise IT workloads
will be Cloud-based or Hybrid.
By 2015, 25 % or more of TOTAL enterprise IT workloads
will be Cloud-based or Hybrid.
SaaS is driving adoption.
PaaS enables custom & composite solutions.
9
10. Solution Life Cycle
Evolution & Innovation
Software as Bits
Requests for
Enhancements
Develop Test Release Install Qualify Operate
Software
Debug Version
Users
Service
Usages & Feedbacks
Develop Test Operate Internet
Service
Debug
Users
10
12. The Amazon AWS pace of Innovation
Simple Storage Service (S3) 13 Mars 2006
Simple Queuing Service (SQS) 11 Juillet 2006
Elastic Cloud Computing (EC2) 23 Aout 2006
Flexible Payment Services (FPS) 2 Aout 2007
Simple DB
DevPay
3 major 13 Décembre 2007
16 Décembre 2007
Cloudfront
Innovations
Elastic Block Storage (EBS) 20 Aout 2008
18 Novembre 2008
Elastic Map Reduce per year
Virtual Private Cloud
2 Avril 2009
26 Aout 2009
Relational Database Services (RDS) 27 Octobre 2009
Versionning for S3 8 Février 2010
Simple Notification Services (SNS) 14 Avril 2010
12
13. Evolution of Computing Models
SLA 99.999 99.9 Always On
Scaling Vertical Horizontal
Hardware Custom Enterprise Commodity
HA Model Hardware Software
Software Centralized Decentralized Distributed
Centralized Shared
Consumption Service Service
Self-service
Mainframe Enterprise Cloud
Source: Randy DIAS – www.cloudscaling.com
13
15. AWS uses non-Enterprise IT patterns
Enterprise
SLA Target 99.999%(pretend) 99.95%(realistic)
(Uptime)
SLA Actual 99.9%(if lucky) 99.9%(or more)
Scaling Vertical Horizontal Scale Engineering
Hardware Vanity Commodity
Hardware Hardware Commodity Operational
HA Model Use Hardware & pray Use software HW Excellence
& plan for failure
Software Arch Centralized & brittle Decentralized &
isolated fault domains
Consumption
Model Call the help desk Self-service & APIs
Automation 1 admin per x00 srv 1 admin per x0 000 srv
15
16. Essential Understanding
Scale Engineering
1990’s Big App 2010’s Big App
in the Enterprise on the Internet
100K users? 800M+ users
‘‘ The cloud lets its users focus
on delivering differentiating
business value. ’’
Werner VOGELS
CTO, Amazon
16
17. A Continuum of Services
Cloud Computing
SaaS Platform Infrastructure
Rented Application CRM,ERP… Software Cpu, storage ‘‘Run’’ as used
€/user environment €/cpu/hour, €/Gb, €/Gbps
Contract based Virtual machine
on user features based contract
subscribe develop & run
The solution already start & run No ownership
is a service but full control
Business Operations
user
App dev
Contract based
on APIs & tools
After the development
the solution is a service
17
18. A Continuum of Services
Operations App dev Business users
IaaS PaaS SaaS
Infrastructure Platform Application
as a Service as a Service as a Service
low high
Productivity
high low
Control
detailed (per hour…) macro (user/mois…)
Billing – Pay per use
difficult easy
Cost Predictability
18
19. Public vs. private cloud :
Access vs. Ownership
Provider of Public Public Cloud
Cloud
Public
Access
Private
Access
Private Cloud Virtual Private
Cloud
Private Third-Party
Ownership Ownership
19
20. Public Cloud
CAPEX
Outsourced
Operations OPEX Economy of scale
Very
Automated attractive
Operations prices
Cloud Mutualization
Computing
Immediate Massively
Availability elastic
Self-service Virtualized resources
resources
available
everywhere
20
21. Private Cloud
CAPEX
Outsourced
Operations OPEX Economy of Scale
Very
Automated attractive
Operations prices
Cloud Mutualization
Computing
Immediate Massively
Availability elastic
Self-service Virtualized resources
resources
available
everywhere
21
24. Why a Private Cloud?
Private Cloud The
Security Trusted
Public Clouds don’t meet
Cloud
Data
our requirements Localization
Confidentiality
A step to prepare
the use of a Public Cloud
Carefully Design
Virtualization • Interoperability
• Portability
Automatisation • Service Provisioning &
Monitoring
Self-provisioning • Governance / Billing to
prepare to an hybrid cloud
model
24
27. IT Economy
Standard Standard
for large Standard for the sector Differenciation
enterprises in the sector leaders
Decisions based on costs • Decisions based on Business Value
• Faster (Time to Value)
Standard Solutions
• Open for the next steps
Costs of projects
Build / Entry
Run
A Measurement Culture
Business Indicators
whenever possible Replacement / Exit
27
28. IT organizations must change
A Capacity oriented IT org
Data Centers X 000 m.d. for projects
Process
IT Standards Demand Management
Build & Run
A Business Services Business Services & KPIs
Less technology selection &
oriented IT org more business support
Focus will shift to Business Integrating, monitoring & supporting
Internal & external services
empowerment. Value through governance, orchestration
Encourage Innovation or & innovation
Expect to be bypassed ! Short Cycles – Dynamic service provider
29. Cloud Computing & Innovation
Culture of Preliminary Analysis
10-15 years ago, Limitation: Our capability to
elaborate ROI studies
experimenting was expensive Afraid of: Errors in ROI studies
Experimenting costs as much as More value for the time spent
More value for the euro spent
(or sometimes less than) Limitation: Our capability to manage
experimentations
the analysis Afraid of: Have lost an opportunity
to experiment
Culture of Observation allowing
fast learning and adjustments
29
30. IT role in business leverage of technology
Business’s use of
Business technology
Business self-
self- provisioned
provisioned Business
self-
provisioned
IT provisioning of
IT
technology for
provisioned
business
Business
self- Business
provisioned self-
provisioned
Business
self-
provisioned Innovation will x3 from 5% to 15%
Vendor mngt will x2 from 10% to 20%
Enterprise Arch will x2 from 10% to 20%
Source: Risk mngt will increase from 10% to 15%