1

CLOUD COMPUTING VIA zENTERPRISEa hybrid management model

Jim Porell, Distinguished Engineer, Deputy CTO, Federal Sales jporell@us.ibm.com

© 2011 IBM Corporation

2

Government Focus on Cloud Computing• Administration CIO Vivek Kundra has been promoting cloud since he

came to office

• OMB Directives to “go to the cloud first…”– If not, then explain why

• Several agencies want to be cloud service providers– Dept of Interior, National Business Center– USDA National Information Technology Center– Others

• GSA push to let contracts for Infrastructure-as-a-Service (IaaS), Platform-as-a-Service (PaaS), and Software-as-a-Service)– GSA awarded contract to Google for internal cloud based e-mail

• Dept. of Interior awarded contract to MSFT (protested by Google) for cloud e-mail

• Many agencies are wary of “public” and “community” cloud offerings with security being the single greatest concern

3

Summary: Start today – save tomorrow• Deployment goals:

– Take out cost– Reduce risk– Improve Security and Resilience– Meet or exceed service level needs– Provide investment protection for the future

• Identify immediate ROI potential. For example:– Database consolidation– Smart Analytics via data sharing– Development environment– Infrastructure as a service (Linux instances, Storage)

• Designate and execute pilot projects to validate/quantify ROI benefit– Joint agency/IBM effort

• Provide results back within three months

4

4

ReducedCost=

IT org…leverages virtualization, standardization, and automation to DoJ ITfree up operational budget for new investments

Virtualization +Standardization Automation+

Cloud Computing on System z builds onthe industry’s leading virtualization technology

Cloud Economics

OptimizedBusiness

=…allowing Business units to optimize new investments for direct

business benefits

Agility + Business & IT Alignment

ServiceFlexibility+

5

Six Steps to Cloud implementation

Create IT roadmap

Implement Cloud services

Establish architecture

Define business value

Determine Cloud delivery modelAssess workload

66

EmployeeBenefits

BusinessTravel

IndustryProcesses Procurement

Business Process as a ServiceBusiness Process as a Service

Cloud Service Models – opportunities to savePrioritize possibilities for our organizations

Virtualized, Shared, Dynamically ProvisionedVirtualized, Shared, Dynamically Provisioned

Servers Networking Data CenterFabric Storage

Infrastructure as a ServiceInfrastructure as a Service

IBM System zSolution Edition forCloud Computing

Financials Collaboration Analytics CRM, ERP, HR

Software as a ServiceSoftware as a Service

IBM Smart AnalyticsCloud for System z

Database Middleware

Platform as a ServicePlatform as a Service

ApplicationRuntimes

DevelopmentTooling IBM WebSphere

CloudBurst Appliance

EXAMPLES:

77

Number of Tenants

Size

of T

enan

ts

Large tenantsMedium tenantsLong tail of small tenants

Medium Tenants Small TenantsLarge Tenants

Isolation: DatabasesShared: Hardware

Isolation: TablesShared: Database

Isolation: RowsShared: Tables

MT ApplicationMT ApplicationMT Application ornon-MT application

Example: Database-as-a-Service with DB2 on z/OS or Oracle on Linux• Multi-tenancy: multiple companies or users using the same software with a level of

isolation– Tenants are companies or users that would have historically installed and used a

single instance of software solely for their own use – Multi-tenancy allows companies/users to use the same software with a level of isolation

• Multi-tenancy can further reduce hardware and maintenance costs of DBaaS• Analogous to users running various applications on the same operating system

– The point is to share management and hardware costs among a number of tenants– Tenants, like the distinct users on an operating system require a level isolation

8

Time Business Units IT org BenefitNew Application developedRequires Oracle DB

Repeat for each new application

Days Determines size of serverMakes recommendation

Weeks to Months

Dept justifies purchaseDept does P.O.Orders ServersPays SW license fees

Orders ServersPays SW license fees

Either org can order servers, but IT would manage them.

Week(s) Servers receivedSoftware installed/testedDatabase loaded

Floor space shrinksEnergy & cooling costsLicense fees growMore FTEs req’d

Days Dept tests configDeploys applic in product’n

DB in productionSends bill for Ops service

Pays IT Ops Bill Who pays when new data center req’d?

Current X86 or UNIX DB server deployment

9

Time Business Units IT Org BenefitOne time set up

Oracle competency centerConsolidation on System z

New Application developedRequires Oracle DB

Repeat for each new application

Days Determines size of serverTells dept the cost

Days Dept justifies purchaseRequests service

Day(s) DB Server provisionedIncremental fees, if anyOracle instance cloned/createdDatabase loaded

Same floor space & energyLicense fees managed

Days Dept tests configDeploys applic in product’n

DB in productionSends bill for Ops service

Pays IT Ops Bill Reduce costsReduce risk and timeImprove security/resilience

System z Linux DB “Cloud” deploymentChange the economics and simplify for users

1010

Consolidation on System z

$0.0

$0.5

$1.0

$1.5

$2.0

New X86 z10 z196

Mil

lio

ns

HW maint

HW

SW S&S

SW OTC

10

Linux on z196 Lower acquisition costs of hardware and software vs distributed servers* Less than $1.00 per day per virtual server (TCA)* Reduce floor space by up to 90% compared to distributed servers* Reduce energy consumption by up to 80% compared to distributed servers*

Lower acquisition costs of hardware and software vs distributed servers* Less than $1.00 per day per virtual server (TCA)* Reduce floor space by up to 90% compared to distributed servers* Reduce energy consumption by up to 80% compared to distributed servers*

The Most Efficient Platform for Large Scale Consolidation

74% less thanNehalem74% less thanNehalem

39% less than Nehalem39% less than Nehalem

Distributed server comparison is based on IBM cost modeling of Linux on zEnterprise vs. alternative distributed servers. Given there are multiple factors in this analysis such as utilization rates, application type, local pricing, etc., savings may vary by user.

*

New x86 z10 z196

Consolidate 40 Oracle server cores onto 2 Linux cores on zEnterprise Consolidate 40 Oracle server cores onto 2 Linux cores on zEnterprise

1111

Smart Analytics Cloud

A private cloud within the enterprise A solution for delivering business intelligence to the entire organization

Creates… That delivers …

Phase 1: Create awareness of, a strategy for, and a governance foundation for BI across the organization

Phase 2: Preparation for the Smart Analytics Cloud

IT org and IBM Services

Phase 3: Install the basecloud, integrate into thecorporate enterprise, and test the cloud use cases

Phase 4: Educate the enterprise for on-going success with the Smart Analytics Cloud

The solution components…

Learn more at: http://www.ibm.com/systems/z/solutions/cloud/smart.html

Example: Smart Analytics Cloud for System z

Open, enterprise-class BI platform

Web OfficeSearchMobile

Cognos 8 BIA broad range of BI capabilities

IBM System zCentralize, Virtualize, Simplify the BI infrastructure

IBM Hardware

IBM Software

1212

Near Real time Analytics canspeed results, reduce cost, reduce risk

R I I NS TC E L

MA INFRAME

ClaimsPOSCredit/Debit

DecisionSupport

FilterExtract

Move

PII input

DB

tmp

tmp

resultresult

result

Traditional Operations

z196

ISAO

DecisionSupport

Transform

zClaimsPOSCredit/Debit

DB

CognosOnLinux

Business Problem–Data warehouse can detect trends, but not necessarily prevent fraud or upgrade transactions in real time because data is copied in bulk or batch mode

Insight instead of Hindsight–Data is copied in nanoseconds instead of hours or days–Opens up opportunities for real time analytics

–Preventing fraud–Making business analytic decisions faster

–Improved performance and lowers cost–Uses blade-based specialty processors, storage for warehouse workloads–Boosts overall query performance 5x – 10x–Customers could see a 40% reduction in storage utilization–Supports in-memory column store for parallel star schema queries –Uses column-based compression to minimize storage needs–Unchanged interfaces to DB2 for z/OS and thus no changes to the BI/DW applications–Provides capability to perform both transactional (OLTP) and warehousing (OLAP) type of queries in the same database management system

1313

Cloud computing services from IT org can deliver measurable results and address IT infrastructure challenges

IT attributes From To

Server/storage utilization 10-20% 70-90%

Test provisioning Weeks Minutes

Change management Months Days/hours

Release management Weeks Minutes

Metering and billing Fixed cost Variable cost

Service catalog ordering Months Days/hours

Service access Administered Self service

Payback period for new services Years Months

SOURCE: Based on IBM and client experience.

Virt

ualiz

atio

n

Jour

ney

to c

loud

Stan

dard

izat

ion

Aut

omat

ion

1414

Cost Difference Expectations

Unclassified / FOUO

The chart to the left demonstrates real world customer experiences in cost and time differences between a System z10 and an x86 server deployment.

Actual costs and differences will vary, but this helps to demonstrate some of the possible savings, as well as higher costs with System z.

It’s the aggregate cost value that favors System z, at both the initial acquisition (TCA) and multi-year operation (TCO)

Not all deployments will favor System z, but for this case, it is expected to see similar value

1515

Higher Utilization Shared everything architecture Superior workload

management Host large and small

workloads

More Efficient Data Center Less power and cooling Less floor space Fewer parts to monitor

Increased Productivity Efficient, rapid provisioning Superior life-cycle

management Fast, easy technology

refresh

Greater Reliability, Availability Hardware / hypervisor

redundancy Decades of RAS innovation Capacity and Backup on

Demand

Why deploy clouds on larger, scale-up servers like System z?

16

Summary: Start today – save tomorrow• Deployment goals:

– Take out cost– Reduce risk– Improve Security and Resilience– Meet or exceed service level needs– Provide investment protection for the future

• Identify immediate ROI potential. For example:– Database consolidation– Smart Analytics via data sharing– Development environment– Infrastructure as a service (Linux instances, Storage)

• Designate and execute pilot projects to validate/quantify ROI benefit– Joint agency/IBM effort

• Provide results back within three months

17

BACKUP FOLLOWS

18

6 Components of CloudSupports z/VM and Linux

Subscribe to Service Request a service “Sign“ Contract

Offer Service Register Services and

Resources Add to Service Catalog

Service Creation Scope of Service SLAs Topologies, Best Practices

Management Templates

Deploy Service Request Driven Provisioning Management Agents and Best Practices Application / Service On Boarding Self-service interface

Manage Operation of Service Visualize all aggregated

information about situations and affected services

Control operations and changes

Event handling Automate activities to

execute changes Include charge-back

Terminate Service Controlled Clean-up

Cloud Service Lifecycle Management

1919

z/OS z/OS z/VM z/VM

BusinessApps

DataServing

LinuxVirtualServers

LinuxVirtualServers

Shared Everything Infrastructure(CPU, Memory, Network, Adapters, Crypto, Devices)

(1) Based on zEnterprise comparison to virtualized x86 alternative(2) Based on three-year acquisition costs for large-scale, enterprise-class workloads(3) Based on life-cycle management testing of large-scale virtual server environment conducted by IBM

VirtualServers

VirtualServers

zEnterprise-ManagedInfrastructure(Hardware, Hypervisors,Virtual Servers)

Hypervisor Hypervisor

Consolidate even more with zEnterprise IFLs: up to 60%faster at 33% lower price

Increase energy savings asyou scale, up to 75% (1)

Spend up to 70% less onacquisition costs (2) and boost staff productivity by up to70% (3) compared tovirtualized x86 alternatives

Incorporate IBM Powerand System x technologiesfor unparalleled levels of workload optimization

Manage and govern theintegrated environmentto deliver superior business results at a lower cost

Hypervisor

VirtualServers

High Speed Private NetworkHigh Speed Private Network

IBM zEnterprise for IT Optimization, Consolidation, Cloud ComputingEven Greater Savings, Operational Simplification, and System Reliability

20

20

Putting zEnterprise System to the taskUse the smarter solution to improve your application design

Sys

tem

z H

ard

wa

re M

an

ag

em

en

t C

on

sole

(H

MC

)w

ith U

nifie

d R

esou

rce

Man

ager

zBX

Select IBM Blades

Blade HW Resources

Optimizers

IBM

Sm

art

An

aly

tics

Opt

imiz

er

z HW Resources

z/OS

Support Element

LinuxonSystem z

z/VM

Private High Speed Data Network IEDN

Private Management Network INMNUnified Resource Manager

Private data network (IEDN)

1 All statements regarding IBM future direction and intent are subject to change or withdrawal without notice, and represents goals and objectives only.

Customer Network Customer Network

System z Host

Linux on System x 1

AIX on POWER7

Dat

aP

ow

er 1

Fu

ture

Off

eri

ng

Fu

ture

Off

eri

ng

Blade Virtualization Blade Virtualization

System z PR/SM

z/TPF

z/VSE

LinuxonSystem z

2121

HMCHMC

Hypervisors Energy

Networks

Energy Management▀ Monitoring and trend reporting of CPU

energy efficiency.▀ Ability to query maximum potential power.▀ Static power savings.

Performance

VirtualServers

Operations

Operational Controls▀ Auto-discovery and configuration

support for new resources. ▀ Cross platform hardware problem

detection, reporting and call home.▀ Physical hardware configuration,

backup and restore.▀ Delivery of system activity using new

user.

Hypervisor Management▀ Integrated deployment and

configuration of hypervisors▀ Hypervisors (except z/VM) shipped and

serviced as firmware.▀ Management of ISO images.▀ Creation of virtual networks.▀ Manage and control communication

between virtual server operating systems and the hypervisor.

Network Management▀ Management of virtual networks including access control

▀ Single view of virtualization across platforms.▀ Ability to deploy multiple, cross-platform virtual

servers within minutes▀ Management of virtual networks including

access control

Virtual Server Lifecycle Management

▀ Wizard-driven management of resources in accordance with specified business service level objectives

▀ HMC provides a single consolidated and consistent view of resources

▀ Monitor resource use within the context of a business workload

▀ Define workloads and associated performance policies

Workload Awareness and Platform Performance Management

zEnterprise hardware management and platform management …

Key▀ Manage suite▀ Automate suite

Save time, cost and simplify asset management

Decrease problem determination and resolution time for cross-platform resources

Improve and simplify cross-platform availability procedures

Enable broader and more granular view of resource consumption

Factory installed and configured network

Improved network security with lower latency, less complexity, no encryption/decryption

Allow critical workloads to receive resources and priority based on goal-oriented policies established by business requirements

Smart business adjustments based on workload insight

Provide deep insight into how IT resources are being used

Gain flexibility, consistency and uniformity of virtualization

Provide the business with faster time to market

Simplified network management for applications

… value made possible by the Unified Resource Manager

Simplified installation of hypervisors

Gain significant time to market with improved speed of deployment

Simplified installation of hypervisors

Gain significant time to market with improved speed of deployment

Simplified energy management

Energy cost savings