Architecting Clouds

Cloud Computing: Infrastructure and Application
Architecture Considerations
Uli Hitzel, Cloud Architect (ASEAN) August 1st, 2012
Acknowledgements: this presentation leverages material from IBM cloud computing presentation decks, 360 degree architect session
(Sugandh
Mehta, Bernard
2012
IBM Corporation
Glasner, Harirajan Padmanabhan) and Architecting applications for the cloud @ IBM developer networks by Peter Bell (Systemsforge)
This handout is a modified version of the original

presentation deck.
developerWorks Live! Architecting Clouds
2012 IBM Corporation
About Uli
Architect (ASEAN)
projects at banks, telcos,
service providers
& in public sector
Technical Expertise
virtualization
automation
business process integration
web technologies
Session Agenda
Introduction
Infrastructure and Platforms
Cloud Applications
Cloud helps clients to
Transform
Increase agility and organizational

efficiency faster time to market
Improve
Value Chain
Create
Disruptors
Save costs - profit from flexibility in IT

spends (CAPEX/OPEX)
Innovators
Innovate and create new business
models additional revenue streams
Optimizers
Disrupt the market coming up with
services that only work on the cloud
model
Enhance
Extend
Invent
Customer Value Proposition
Cloud Computing Definition

Cloud computing is a new
consumption and delivery model
inspired by consumer
internet services.
Essential characteristics:
1. On-demand self-service
2. Broad network access
3. Location independent
resource pooling
4. Rapid elasticity
5. Measured Service
Usage
Tracking
Web 2.0
SOA
End User Focused

Virtualization
Service
Automation
Whats different with cloud computing?

Without cloud computing
Software
Hardware
Storage
Networking
With cloud computing
Software
Hardware
Storage
Networking
Software
Hardware
Storage
Networking
Virtualized resources Location

Automated service
independent
management
Salability & Elasticity
Standardized services Self-service
Shopping Mall B2B IaaS Cloud

Service provider owns the system
Business rents units to offer services to
their clients consumers
Shared resources (building, elevators,
parking, washrooms)
Shared Services (security, cleaners)
Country-wide, world-wide network of
shopping malls
better ROI, larger audience, more flexibility
& less risk for the shop owners
One Size Does not Fit All
What exactly do we want to do?

Workloads which ones? required
functionality and performance?
Service models
Deployment models
Organizational Motivations & Constraints
10
Workloads
11
Definition Cloud Service Models

Software
as a Service
Data
Data
Runtime
Runtime
Runtime
Middleware
Middleware
O/S
O/S
Data
Virtualization
Servers
Storage
Networking
Virtualization
Servers
Storage
Networking
Middleware
O/S
Virtualization
Servers
Vendor Manages in Cloud
Applications
Applications
Client Manages
Applications
Platform
as a Service
Client Manages
Infrastructure
as a Service
Storage
Networking
Customization; higher costs; slower time to value

Standardization; lower costs; faster time to value
12
Service Layers
Software
as a Service
(SaaS)
Platform
as a Service
(PaaS)
Collaboration
CRM / ERP / HR
Industry Applications
Financials
Web 2.0 Application Runtime
Development Tooling
Middleware
Database
Storage
Infrastructure
as a Service
(IaaS)
13
Java Runtime
Shared Pool
Virtual Machines
Images
Virtual Network
CRM , ERP Applications

Human Resources
Analytics
Financial Services
Mail, Web Conferencing
Applications
Middleware
Development tools
Java and Web 2.0 runtimes
Shared virtualized dynamic provisioning for:

Server functionality
Networking functionality
Data center functionality
Storage functionality
Cloud deployment models

Hybrid Cloud
Private Cloud
1
2
Enterprise
Data Center
Private
Cloud
3
Enterprise
Data Center
Managed
Private
Cloud
Enterprise
owned
Provider
operated
1.
2.
3.
4.
5.
14
Community Cloud
Enterprise
Enterprise A
Enterprise B
Hosted
Private
Cloud
Provider
owned and
operated
Enterprise C
Shared
Cloud
Services
Public Cloud
5
User
User
User
C
E
A User
User
B
D
Public
Cloud
Services
Customer managed private Cloud

Customer premise, provider operated private Cloud
Provider premise, provider operated private Cloud
Provider premise, provider managed, public Cloud
Provider premise, provider managed, provider applications, public Cloud
IBM Cloud Computing Reference Architecture

Cloud Service Provider
Cloud Service
Integration Tools
Common Cloud
Management Platform (CCMP)
Cloud Services
Service Creation
Tools
Process Integration
Application & Data
Integration
Identity Integration
Business-Processas-a-Service
3rd
Existing &
party
services, Partner
Ecosystems
tbd
Management
Federation
Software-as-a-Service
Spillover & Failover
tbd
Middleware
Infrastructure
Service Runtime
Development
Tools
Service Automation
Management
Service Offering
Management
Service Asset
Management
Platform-as-a-Service
Software
Development
Tools
Cloud
Enablement
Programming
Model
Resources
Image Creation
Tools
Subscription
Management
Middleware
Services
Service Operations
Management
Financial
Management
Image Lifecycle
Management
Infrastructure-as-a-Service
Application
Lifecycle Services
Service
Management
Development
Tools
Customer
Management
Service Management
Applications
Business
Support
Services
(BSS)
Service Quality
Management
tbd
Consumer Inhouse IT
Business
Processes
Operational
Support
Services
(OSS)
Server Virtualization
Storage Virtualization
Network Virtualization
Platform & Virtualization

Management
Developer & Tester

Collaboration
Development
Automation
Application Lifecycle
Traceability
Infrastructure
Server
Storage
Network
Facilities
Security, Resiliency, Performance & Consumability

Governance
15
Session Agenda
Introduction
Cloud Applications
Designing and delivering SaaS applications
Client case studies
16
Resource sharing considerations

Low
Low
High
Standardisation
Security risk
Cost/tenant
High
High
Low
Tenant
Tenant
Tenant
Tenant
Tenant
Tenant
Tenant
Tenant
Tenant
Application
App
App
App
App
App
App
App
App
Platform
Platform Platform
Platform
Operating System
Operating System
Operating System
Infrastructure
Infrastructure
Infrastructure
Infrastructure
Data center floor
Data center floor
Data center floor
Data center floor
Data center floor
II
III
IV
Physical-level
multi-tenancy
Hypervisor-level
multi-tenancy
Operating
System-level
multi-tenancy
Platform-level
multi-tenancy
Application- level
multi -tenancy
OS
Infrastructure
Infrastr.
Infrastr.
Traditional IT
Application
Platform
OS
Data center floor
Tenant
Platform Platform
OS
Platform Platform
Tenant
OS
OS
Legend:
Dedicated
Shared
17
Virtualization can be on any level

Cloud Service
Consumer
Existing & 3rd party

services, Partner
Ecosys tems
Clo ud
Service
Integ ration
To ols
Cloud Service
Creator
Common Cloud
Management Platform (CCMP)
Cloud Services
Business-Pro cessas-a-Serv ice
Software-as -a-Serv ice

Operational
Supp ort
Servic es
(OSS)
Bus iness
Sup port
Services
(BSS)
Service
Creation
Tools
Platform-as-a-Service
Consumer
In-hous e IT
CCMP
Infrastructure-as-a-Service
Inf rastruc ture
Security, Resiliency, Performance & Consu mability

Govern ance
UI
Application-level
virtualization
BSS
Platform-level virtualization
OSS
manages
OS-level virtualization
Hypervisor-/Infrastructure
level virtualization
Virtualization options
for Cloud service
implementations
18
Virtualization is a great start but it adds

complexity!
What is the utilization of your
environment? How do you
ensure adequate capacity?
How are your services

performing? How do
you isolate problems?
How do you
rapidly provision
services?
Server
Where are your

services running?
Storage
Network
Agility
Consolidate
How do you
manage image
sprawl?
Resources
Manage
Workloads
Automate
How do you
reduce cost of
service delivery?
Processes
Optimize
Delivery
How do you charge for

shared resources?
19
How do you
secure your
infrastructure &
protect data?
Journey to Cloud Computing
Cloud focuses on
eased service
consumption &
management
Consumption based
metering and dynamic
capacity optimization
Virtualization
underpins
Cloud
End-to-end real time

monitoring and optimization
Business service catalogs & self service
Service delivery automation
Virtualization management
The Cloud Computing Reference Architecture has best practises

for these aspects
Usage Metering and Accounting
Cloud Service
Consumer

Common Cloud
Management Platform
Cloud Services
Cloud Service
Flexible support of delivery models
Creator
Service
Manager
Business
Manager
Service Automation Management
Operational Support
Interpret and Execute Build-Existing
and &Management PlansOSS
Services
3 party
BPaaS
Orchestrate Management Componentry
services,
BSS Business Support

Services
rd
Partner
Ecosystems
Service Delivery Catalog
SaaS
PaaS
Consumer
In-house IT
Service Automation Management
Service
Request
Management
Change &
Configuration
Management
Image
Lifecycle
Management
Provisioning
Incident &
Problem
Management
IT Service
Level
Management
Monitoring &
Event
Management
IT Asset &
License
Management
Capacity &
Performance
Management
Platform & Virtualization Management
Customer
Account
Management
Service
Offering
Catalog
Service
Offering
Management
Contracts &
Agreement
Management
Service
Request
Management
Order
Management
Subscription
Management
Pricing
Entitlement
Management
Metering
Rating
Billing
Service Development Portal & API
Service Consumer Portal & API
Hybrid
Cloud
Integration
Service
Creation
Tools
Image Management
Clearing &
Settlement
Accounts
Payable
Accounts
Receivable
Design, build and manage images for cloud services

IaaS
Service Provider Portal & API
Virtualized Resource
Management
Deployment
Architect
Deploy cloud services on virtualized

resources
Manage virtual resources
Transition
Manager
Operations
Manager
Security &
Risk Manager
Customer
Care
Inf rastructure
Security, Resiliency, Performance & Consumability
Security
Governance
Design for Multi-Tenancy

Protect assets through Isolation, integrity, image- risk and compliance
management
21
Session Agenda
Introduction
Cloud Applications
22
Traditional vs cloud applications
Traditional applications:
Extend the life of these applications
Add new functionality more cheaply and quickly
Connect these applications to the cloud
Maintain the security features built into on premise systems
Cloud applications:
Connect to legacy applications
Connect to other cloud applications
Enhance capabilities at the lowest possible cost
23
Implementation Considerations (1 of 4)
Current Applications,
Current IT Assets, Other
SaaS Offerings
Business Drivers
Key Cloud Characteristics

Leveraged
Market and Business

opportunities
Competitive differentiators
Business Case
Business Model
Pricing Models
On-demand self-service
Broad network access
Resource Pooling
Rapid Elasticity
Measured Service
SaaS Offerings
Cloud Application
Integration
APIs needed
Customer and
Enterprise Backend Systems
Cloud applications supporting multiple SaaS offerings

SaaS offerings composed of several existing applications
Business requirements for each offering could be different
Integration with Backend systems
24
Implementation Considerations (2 of 4)
SaaS Offerings
Business Drivers

Leveraged
Market and Business

opportunities
Business Case
Business Model
Pricing Models
Resource Pooling
Rapid Elasticity
Measured Service
PaaS
SaaS Offerings
Cloud Application
IaaS
Integration
APIs needed
Customer and
choosing the appropriate underlying service layer: IaaS or PaaS

each layer has own set of business drivers and cloud characteristics
25
Implementation Considerations (3 of 4) consume external XaaS cloud services

SaaS Offerings
integration with external backends

Business Drivers

Leveraged
Market and Business

opportunities
Business Case
Business Model
Pricing Models
Resource Pooling
Rapid Elasticity
Measured Service
PaaS
SaaS Offerings
Cloud Application
IaaS
Integration
APIs needed
Customer and
Client Domain
Business partners and other
third-party provider domain
SaaS Offering
26
Integration APIs supported
PaaS Offering
IaaS Offering
Back-end Systems
Implementation Considerations (4 of 4) ROI, risks, cloud characteristics &

SaaS Offerings
business benefits are different for each

provider. Integrating external backends
Business Drivers

Leveraged
Market and Business

opportunities
Business Case
Business Model
Pricing Models
Resource Pooling
Rapid Elasticity
Measured Service
PaaS
SaaS Offerings
Cloud Application
IaaS
Integration
APIs needed
Customer and
Client Domain
Business partners and other
third-party provider domain
SaaS Offering
Business Benefits
Pricing Models
27
Integration APIs supported
PaaS Offering
IaaS Offering
Back-end Systems

Supported On-demand self-service
Resource pooling
Rapid Elasticity
Measured Service
Use Guidance from the Cloud Computing Reference Architecture

Users
Cloud Environment
Delivery
Cloud Computing Management Platform
Channels
Cloud External
Cloud Business
Cloud
Environment
Roles
Services
Applications
Cloud Managed Environment
Business(SaaS)
App Service ..n
Software Platform Services
Business App Service2
Saas
Administrator
CCMP External
Components
BSS Externalized
Components
Business Application
Service1
Note: SaaS Includes SaaS,
BPaaS, AaaS
SaaS
Consumer
Business Mgr
Cloud Resources
Infrastructure Services
Virtualized Infrastructure
CCMP Internal
Components
OSS Externalized
Components
Cloud Mgmt
Supporting Services
(CMSS)
Business
Support
Systems
CMSS
Components
Operational
Support
Systems
Consumer
of SaaS
Integration Layer
BAU Application Users
Other Enterprise Resources (BAU)

Data Stores
Legacy
Applications
Packages
Non Cloud
Systems
External Entities
Business Partner
Systems
28
Build Cloud Applications, not applications in the cloud !

Key Principles
Virtualize the Application

Stack
Componentize. Decouple &
design all components as a
Black Box
Design for scalability
NIST Definition of Cloud Computing

29
Design for scaling OUT !
Traditional way
add more RAM
use faster servers
expensive micro-optimization
complex caching
faster hard disks
Design for scalability
minimize mutable state
create asynchronous services
alternative data stores
automate deployment
design for failure
30
Design for Scalability: Design for failure
"Everything fails, all the time

Werner Vogels, CTO Amazon.com
find single point of failures
evaluate scenarios. What
levels of risk is acceptable?
failure tolerance
dont overinvest
31
Design for Scalability: Create components & asynchronous

services
Offload work from main application servers

Web 2.0
Break tasks into separate services, run by
different components
Scale independently
Use message queues for guaranteed delivery
32
Design for Scalability: Minimize Shared Mutable State

Variables shared across application
Multiple servers and processes trying to update the same

variables at the same time result in deadlocks, time-outs,
and failed transactions
minimize or eliminate those in webservers, application
and the database
specific considerations for filesystems, applications
and datastores
look at cluster filesystems, object stores, NoSQL /
CouchDB, MongoDB asynchronous fire & forget
updates
33
Design for Scalability: Automate Deployment
Virtual Application
Patterns
Highly automated deployments
using expert patterns
Business policy driven elasticity
Built for the cloud environment
Leverages elastic workload
management services
Virtual System
Patterns
Packaged for virtual
environments
Automated deployment of
middleware topologies
Traditional administration
and management model
Workload Platform
Services
Virtualized Middleware
Services
cloud applications
34
virtualized applications
OS Images for
Existing Software
Standard software installation
and configuration on OS
Images created through
extend/capture
Traditional administration and
management model
Virtualized Infrastructure
Services
existing applications
3 take aways
Cloud Computing is more than just technology.

Its a business transformation that can help saving
costs, increasing agility
The IBM Cloud Computing Reference Architecture
can help you design and implement cloud
environments based on best practises and
industry standards
Design applications specifically for the cloud,
design for scalability & failure
35
36
Trademarks and Disclaimer
37

Architecting Clouds

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Architecting Clouds

Uploaded by

Copyright:

Available Formats

Cloud Computing: Infrastructure and Application

This handout is a modified version of the original

developerWorks Live! Architecting Clouds

2012 IBM Corporation

developerWorks Live! Architecting Clouds

2012 IBM Corporation

developerWorks Live! Architecting Clouds

2012 IBM Corporation

Cloud helps clients to

Increase agility and organizational

Save costs - profit from flexibility in IT

Customer Value Proposition

developerWorks Live! Architecting Clouds

2012 IBM Corporation

Cloud Computing Definition

End User Focused

developerWorks Live! Architecting Clouds

2012 IBM Corporation

Whats different with cloud computing?

With cloud computing

Virtualized resources Location

2012 IBM Corporation

Shopping Mall B2B IaaS Cloud

developerWorks Live! Architecting Clouds

2012 IBM Corporation

One Size Does not Fit All

developerWorks Live! Architecting Clouds

2012 IBM Corporation

What exactly do we want to do?

developerWorks Live! Architecting Clouds

2012 IBM Corporation

developerWorks Live! Architecting Clouds

2012 IBM Corporation

Definition Cloud Service Models

Vendor Manages in Cloud

Vendor Manages in Cloud

Vendor Manages in Cloud

Customization; higher costs; slower time to value

developerWorks Live! Architecting Clouds

2012 IBM Corporation

Web 2.0 Application Runtime

CRM , ERP Applications

Shared virtualized dynamic provisioning for:

developerWorks Live! Architecting Clouds

2012 IBM Corporation

Cloud deployment models

Customer managed private Cloud

developerWorks Live! Architecting Clouds

2012 IBM Corporation

IBM Cloud Computing Reference Architecture

Spillover & Failover

Platform & Virtualization

Developer & Tester

Security, Resiliency, Performance & Consumability

developerWorks Live! Architecting Clouds

2012 IBM Corporation

developerWorks Live! Architecting Clouds

2012 IBM Corporation

Resource sharing considerations

Data center floor

Data center floor

Data center floor

Data center floor

Data center floor