autonomic computing grid computing...

Fulvio CapogrossoDistinguished EngineerServer Group, South Region, EMEA

e-business on demanda technology perspective

Open StandardsOpen Standards

Autonomic ComputingAutonomic Computing Grid ComputingGrid Computing

e-Utilitye-Utility

Web Services

fc

AgendaAgenda

ScenarioScenario

DefinitionsDefinitionsbusiness propositionbusiness propositiontechnical propositiontechnical propositionfinancial propositionfinancial proposition

The RoadmapThe Roadmap

Key Enabling TechnologiesKey Enabling Technologiesautonomic computingautonomic computinggrid computinggrid computingintegration technologiesintegration technologies



e-Utilitye-Utility

Web Services

fc

Today's e-business ScenarioToday's e-business ScenarioUnpredictable Demand

ComplexityComplexityManagementManagement

Business Business EfficiencyEfficiency

Business Business ContinuityContinuity

Operational Operational FlexibilityFlexibility

CostCost&&

ROIROI(Return-on-Investment)(Return-on-Investment)

TimeTimetoto

MarketMarket

Middleware

Storage

Edg

e S

erve

rs Transaction Servers

Data Servers

Web ApplicationServers

Directory & Security

Servers

Web Presentation

Servers

Customers

Suppliers

Partners

Employees

Difficult Economy Cycle

fc

Technology EvolutionTechnology Evolution

Internet & Connectivity Base Technology

Grid Computing Autonomic Computing

fc

Open StandardsOpen StandardsThe foundation for operational flexibility

WSDL

Web Services

Java

Linux

SOAP

XML

OGSA

fc

e-business on demand Summarye-business on demand Summary

Responsive Responsive

Variable cost Variable cost

Focused Focused

Resilient Resilient

Business attributesBusiness attributes IT AttributesIT AttributesOpenOpen

IntegratedIntegrated

VirtualizedVirtualized

AutonomicAutonomic

New EconomicsNew EconomicsComputing as a utilityComputing as a utility

Variable costsVariable costs

Variable ownershipVariable ownership

anan 'on-demand 'on-demand' Enterprise is defined by a set of :' Enterprise is defined by a set of :business attributesbusiness attributeskey characteristics of the operating environment key characteristics of the operating environment new environment economicsnew environment economics

'e-business on demand''e-business on demand' is the term is the term IBM has chosen to name a new IBM has chosen to name a new movement in business and computing, movement in business and computing, which is changing the way technology which is changing the way technology is deployed and used in businessis deployed and used in business



e-Utilitye-Utility

Web Services

fc

e-business on demande-business on demandCustomers

Focused

Resilient

Responsive

Suppliers

Application Providers

Service Providers

Network Providers

Technology Providers

Variable

Enterprise

fc

Sense and Respond Business ModelSense and Respond Business ModelAn adaptive management framework developed at the IBM An adaptive management framework developed at the IBM Advanced Business InstituteAdvanced Business Institute

Business on demand rather than business as plannedBusiness on demand rather than business as planned

Business strategy, structure and governance principles Business strategy, structure and governance principles designed for environments of unpredictable changesdesigned for environments of unpredictable changes

New business core competencies New business core competencies earlier understanding of what is happening now and in the near earlier understanding of what is happening now and in the near future future dynamically dispatch modular capabilities to respond dynamically dispatch modular capabilities to respond operational and management operational and management flexibilityflexibility

Sense

RespondKnowledge

fc

on demand ITon demand IT

The Path to 'on demand' IT The Path to 'on demand' IT

IT Infrastructure IT InfrastructureOptimizationOptimization

including variable cost commercial offeringsand e-Utility services

Business Process Business Process OptimizationOptimization

Business process integrationBusiness process re-ingeneering

GWLANSAN

NAS SAN

iSCSI Channel Attached

CPCPCPCP

Partitioning Clusters

Server Farms

Web Services

System Mgmt Database

ISVSolutions

Servers Storage Network

Customer Applications

Application Servers

OGSAMeta OS Services

��

��

��

��

��

��

��

��

��

��!�

��"��"��

��#�� $��

��"��

��"��$��

user requestsuser requests

'vertical' business processes'vertical' business processes

IntegrationIntegrationWeb ServicesWeb Services

Higher Levels of Higher Levels of VirtualizationVirtualization

Grid computing

�� LotusDB2

CRM ERP SCM

TivoliRational

PLM

��

fc

New 'on-demand' EconomicsNew 'on-demand' Economics

Cost

Ow

ners

hip

Fixed

OwnedAsset

Utility

Variable

Data center virtualization to build 'internal computing utilities' Data center virtualization to build 'internal computing utilities' higher resources utilizationopen standardspermanent and On/Off Capacity on Demand

External computing utilityExternal computing utilityto acquire computing and applications from service providers

'pay per use' 'pay per use'

More flexible, variable and economically attractive choices for buying and managing computing

fc

Autonomic Computing Grid Computing

Key Enabling TechnologiesKey Enabling Technologies

"Simple .. Open .. Standard .. Dynamic .. Virtual"

Web Services

Autonomic Computing Autonomic Computing www.ibm.com/autonomic

Is the term IBM uses for systems that have the ability to Is the term IBM uses for systems that have the ability to manage themselves and dynamically adapt to change in manage themselves and dynamically adapt to change in accordance with business policies and objectives accordance with business policies and objectives

'Faster, smaller, cheaper' and ....'Simpler'

fc

Autonomic Computing Autonomic Computing Autonomic Computing Reference ArchitectureAutonomic Computing Reference Architecture

4 major features 4 major features Self-configuring

Adapt automatically to dynamically changing environments

Self-healing Autonomic problem determination and resolution

Self-protecting Systems designed to protect themselves from unauthorized access

Self-optimizing Monitor and tune resources automatically

5 levels of implementation 5 levels of implementation maturitymaturity

BasicManagedPredictiveAdaptiveAutonomic

Delivered through multiple R&D projectsDelivered through multiple R&D projectsOpen standard basedLeveraging on high level features already existing in current high end server technology

Element

Monitor

Analyze

Sensors

Execute

Plan

Effectors

Knowledge

Sensors Effectors

Element

Monitor

AnalyzeAnalyze

Sensors

ExecuteExecute

PlanPlan

Effectors

Knowledge

Autonomic Manager

ManagedElement

fc

Autonomic Computing Delivery Autonomic Computing Delivery Core Enabling CapabilitiesCore Enabling Capabilities

Problem DeterminationCommon System AdministrationAutonomic MonitoringSolution InstallPolicy-based ManagementComplex AnalysisEnterprise Workload Management

Systems & Resources Management Systems & Resources Management TechnologiesTechnologies

Enterprise Workload Management

Project Symphony

Web Services


ISVSolutions



Application Servers


fc

The eWLM Environment The eWLM Environment

Quality of serviceQuality of service

Multiple heterogeneous serversMultiple heterogeneous servers

Complex transaction flowComplex transaction flow

Dynamic transaction topologyDynamic transaction topology

Middleware

Storage

Transaction Servers


Directory & Security Servers

Edg

e S

erve

rsWeb Pres.

ServersDatabaseServers

Income txnsIncome txns

fc

Enterprise Workload ManagerEnterprise Workload ManagerSystem resources management software for end-to-end autonomic System resources management software for end-to-end autonomic performance management of transaction workload executing on performance management of transaction workload executing on heterogeneous multi-tiered configurations heterogeneous multi-tiered configurations

Key elements: Key elements: policy based end-to-end active management to business priorities

customer defined performance goals for business importance classesdynamically discovery of workload topologyself-learning tuning algorithms from accumulated environment knowledge

Distributed system composed of two parts: a global eWLM manager and Distributed system composed of two parts: a global eWLM manager and multiple instances of local eWLM managersmultiple instances of local eWLM managers

Java application running on any Java supporting platformlocal eWLM for AIX, Solaris, Windows, zOS, OS/400, Linux, (HP-UX)

The global eWLM manager will provide knowledge based services to The global eWLM manager will provide knowledge based services to external users (ex. other system management applications)external users (ex. other system management applications)

Self-OptimizingSelf-Optimizing

fc

The eWLM Terminology The eWLM Terminology

eWLM management domaineWLM management domainSet of servers (OS instances) under the Set of servers (OS instances) under the eWLM controleWLM control

eWLM Management policyeWLM Management policyAdministrative definition of performance Administrative definition of performance expectations for the domainexpectations for the domainConcrete definition of the difference Concrete definition of the difference between good and bad performancebetween good and bad performance

eWLM Managed serverseWLM Managed serversLocal platform eWLM (LeWLM)Local platform eWLM (LeWLM)Control point for local activitiesControl point for local activities

eWLM management servereWLM management serverGlobal eWLM (GeWLM)Global eWLM (GeWLM)Control point for global activitiesControl point for global activities

G eWLM

L eWLM

L eWLM

L eWLM

L eWLM

L eWLM

L eWLM

L eWLM

L eWLM

fc

eWLM Policy

Local eWLM

Learn server role in end-to-end flows

Monitor resource consumption & delays

Identify service classes needing attention (e2e)

Identify potential resource adjustments

Project impact of adjustments

Initiate resource adjustment(s)

Application Instrumentation

eWLM Algorithms eWLM Algorithms

When local resources adjustments are not sufficient, to meet the end-to-end goals, the global eWLM makes further adjustments

Locally managed resourcesLocally managed resourcessystem resources (CPU, Mem, I/O)network resources (bandwidth,....)LPAR resources (physical, logical processors,..)VMWare resources (partitions,..)

Globally managed resourcesGlobally managed resourcesapplication traffic (load balancers)server instances (provisioners)

Domain Statistics

Server Instrumentation

fc

eWLM Platform InstrumentationeWLM Platform Instrumentation

eWLM Management Server

Middleware levelMiddleware levelExtensions to The Open Group ARM Standard, V3

Process registration, deregistrationWork request classification, start, and stopWork request execution thread bind, unbind

Extensions accepted and expected to be approved by 1Q2003

Operating system levelOperating system levelResource management & optimization layer Operating System

ARM Services

Resource Management & Optimization

eServer WLM(Java)

ARM APIs

ARM APIs

BusinessLogic

ARM APIs

Operating System

ARM Services


eServer WLM(Java)

ARM APIs

BusinessLogic

ARM APIs

Operating System

BusinessLogic

Application Processes(Middleware)


ARM ServicesLocal e WLM(Java)

ARM APIs

fc

correlator

Apache

BWLM BWLM

WebSphere

BWLM

local application

Class:Hot

Hop 0

correlatorClass: Hot

Hop 1

each work request has a unique 'transaction correlator'each work request has a unique 'transaction correlator'

application flows correlator along with each external requestapplication flows correlator along with each external request

arm_start_transaction API accepts an optional parent correlatorarm_start_transaction API accepts an optional parent correlator

correlator propagates service class, topology, reporting relationshipcorrelator propagates service class, topology, reporting relationship

Work Request CorrelationWork Request Correlation

fc

eWLM eWLM ManagementManagement Server DataServer DataService Classes

Application participation in end-to-end flow

Application Environments

0

0.5

1

1.5

2

5

10

15

20

25

Application IdentityProcess Relationships

Segment Resource Data

Resource ConsumptionResource Delays

Policy ObjectiveEnd-to-End Results

Policy ObjectiveEnd-to-End Results

Service ClassesApplication Environments

0

0.5

1

1.5

2

5

10

15

20

25

Application IdentitySummary Statistics

Participating Servers

Server IdentityState informationSummary statistics

Relational Topology

Local Local eWLMeWLM

GlobalGlobaleWLMeWLM

fc

Systems Management & AD Environment

Applications

Operating System(s)

Application Software

Hardware and Architecture(s)

Platforms

The IBM Software ApproachThe IBM Software Approach

TivoliSystems Management

Rational ��AD Environment

IBM and non-IBM Op. Systems,

Platforms

IBM eServer and non IBM hardware

�� LotusDB2

SCMPLMCRM ERP

fc

IBM Tivoli Software in the Solution LifecycleIBM Tivoli Software in the Solution Lifecycle

Monitor and Manage

Analyze and Report

Capture Requirements

RationalModel and Design

Rational

Deploy and Configure

Develop & Test

Rational��

Run Time Environment

�� DB2 Lotus ISVs

Solution Lifecycle

Monitoring, Configuration and Change Management

Correlation and Automation

Business Impact Management

fc

An An OrchestratedOrchestrated Environment Environment

Middleware

Storage

Transaction Servers



Edg

e S

erve

rs

Web Pres.Servers Database

Servers

RespondRespond

detect an increase in demand detect an increase in demand and deterioration in performance and deterioration in performance

reallocate existing reallocate existing resources according to resources according to business requirements business requirements

(provisioning)

automatically automatically

serversserversoperating operating systemssystemsmiddlewaremiddlewareapplicationsapplicationsnetwork devicesnetwork devices


SenseSense Monitor

Analyze

Execute

Plan

Knowledge

fc

Traditional ApproachTraditional Approach

Individual dedicated execution environments sized for Individual dedicated execution environments sized for estimated peak capacityestimated peak capacity

over-provisioningover-provisioninglow overall utilizationlow overall utilizationstatic allocations versus dynamic changing environments static allocations versus dynamic changing environments

Capacity PlanningCapacity Planning

Service Level ManagementService Level Management

Infrastructure ProvisioningInfrastructure Provisioning(manual)(manual)

��

��

��

��

��

��

��

��

��

��

��

��

��

monitoringmonitoring

fc

Orchestrated ProvisioningOrchestrated Provisioning

��

��

��

��

��

��

��

��

��

��

Capacity Planning

Service Level Management

Infrastructure Provisioning

resource pool��

��

��

��

��

��

��

��

Monitor

Analyze

Execute

Plan

Knowledge

monitoringmonitoring

fc

a family of modular on demand technologies and solutions for infrastructure resources orchestration with no human intervention

IBM Tivoli Intelligent (Thinkdynamics) OrchestratorIBM Tivoli Intelligent (Thinkdynamics) Orchestratorto manipulate the IT environment in real time according to to manipulate the IT environment in real time according to pre-defined business policiespre-defined business policiesincludes automatic provisioning of IT resources includes automatic provisioning of IT resources

IBM Tivoli Provisioning ManagerIBM Tivoli Provisioning Managerpart of the Orchestrator solutionpart of the Orchestrator solutionautomates the provisioning process of infrastructure deploymentautomates the provisioning process of infrastructure deploymentframework to collect and execute best practices for datacenter framework to collect and execute best practices for datacenter management management

IBM Web Infrastructure OrchestrationIBM Web Infrastructure Orchestrationpre-tested, pre-integrated solution for web serving environmentspre-tested, pre-integrated solution for web serving environmentssoftware, hardware, servicessoftware, hardware, servicesincludes pre-packaged intelligenceincludes pre-packaged intelligence

Project SymphonyProject Symphony

fc

� What it does:� Intelligently manages changes in web demands by

automatically repurposing new blade servers to meet service level agreements

� Key components of the “soft bundle” solution:� IBM Tivoli Intelligent ThinkDynamic Orchestrator�Orchestrator workflows and IBM Director scripts � IBM WebSphere� IBM DB2� IBM eServer BladeCenter� IBM FAStT900 Storage Server

� Capability:�Pre-tested, pre-integrated orchestration of web servers

on Blades with optional implementation services

IBM Web Infrastructure OrchestrationIBM Web Infrastructure Orchestration

GRID ComputingGRID Computingwww.ibm.com/gridwww.ibm.com/grid

Grid computing is a distributed computing model which supports the concept Grid computing is a distributed computing model which supports the concept of of virtual dynamic organizationsvirtual dynamic organizations ....... ....... by providing secure and coordinated by providing secure and coordinated access and sharing of heterogeneous and geographically distributed access and sharing of heterogeneous and geographically distributed resources (applications, data, processing power, storage capacity and other) resources (applications, data, processing power, storage capacity and other) ...... ...... over a network and across organizational boundaries, over a network and across organizational boundaries, ................ using a set using a set of open standards and protocols. of open standards and protocols.

fc

Resource VirtualizationResource Virtualization

The effect of applying hardware and/or software technology to a set The effect of applying hardware and/or software technology to a set of physical (real) resources to make them appear to the user as a set of physical (real) resources to make them appear to the user as a set of (virtual) resources with different (better) characteristicsof (virtual) resources with different (better) characteristics

a definition

Other virtualization technologies :Other virtualization technologies :virtual file systemsvirtual file systemsvirtual disk storage (storage tank)virtual disk storage (storage tank)virtual database (federated database)virtual database (federated database)

CP

Op. Sys.

Uniprocessor

CPCPCPCP

Virtualprocessors

(Partitioning)Multiprocessor

(SMP)

��

Op. Sys.

CPCPCPCP

ClusterGrids

Ex. processing servers virtualization

NAS AttachedStorage

SAN AttachedStorage

iSCSI Attached Storage

ChannelAttached Storage

GWLANSAN

Clustering Hardware

I/O Connections

DATA

fc

BiologicalBiologicalData and Data and

AlgorithmsAlgorithms

TextualTextual Data

DB2 Family

Sybase

Informix

SQL Server

Oracle

� ��

��

� ��

��

��

!�� "�� #� �# � � �$

�� #� � �� "� � ��

� � ��

TeradataMQ Series

ODBC

� � �

� % ��

��

Excel

& � � � '

��% �

& �� '…

Information IntegrationInformation Integration

fc

The Globus Project The Globus Project An initiative by the Argonne National Laboratory, the University of An initiative by the Argonne National Laboratory, the University of Southern California and the University of Chicago (1998) with the Southern California and the University of Chicago (1998) with the objective objective to develop fundamental technologies needed to build to develop fundamental technologies needed to build (computational) grids. (computational) grids.

many collaborators including IBM and Microsoftmany sponsors including government agencies, IBM, Microsoft, Ciscoaffiliated with many WW organizations such as the Global Grid Forum (GGF)

The Globus Project provides the The Globus Project provides the Globus ToolkitGlobus Toolkit, a collection of basic grid , a collection of basic grid services and APIs that can be used to build computational grids and grid services and APIs that can be used to build computational grids and grid based applications (open source (Globus Toolkit Public License (GTPL))based applications (open source (Globus Toolkit Public License (GTPL))

Globus Toolkit GT2.4IBM Grid Toolbox for GT2.0 for AIX and Linux

Open Grid Services Architecture (OGSA) Open Grid Services Architecture (OGSA) result of an R&D 2001 program with IBM participationdelivered by Globus as Globus Toolkit 3.0 (GT3)GRID Services specifications (as extensions of WEB Services)to be delivered with ALL IBM eServer platforms

fc

OGSA & Quality Of Service DeliveryOGSA & Quality Of Service Delivery

QoS is traditionally delivered at QoS is traditionally delivered at platform platform levellevel through vertical integration of through vertical integration of multiple technologiesmultiple technologies

"stovepipe solution towers"local platform dependent interfaces

At the At the IT Infrastructure levelIT Infrastructure level (multiple (multiple heterogeneous platforms) the heterogeneous platforms) the end-to-end QoS delivery is fracturedend-to-end QoS delivery is fractured

labor intensive re-integration of QoS delivery across multiple platformacross multiple implementationsacross multiple standards

Operations Management

Workload Management

Security Management

Availability Management

M.ware Services

Log & Trace Services

Loc. Process & Res. Mgmt

Clustering Services

Connectivity Mgmt

Physical Resources Mgmt

Platform

Middleware

Storage

Edg

e S

erve

rs

Transaction Servers

Data Servers


Directory & Security

Servers

Web Presentation

Servers

Customers

Suppliers

Partners

Employees

fc

Meta-Operating System functions for Meta-Operating System functions for seamless QoS delivery seamless QoS delivery

standard service definitions independent from the local (platform) implementationGRID Services

Interface

Bindings

Implementation

Database Abstraction ServicesDistributed Cluster Services

Domain Management Services

Distributed Log Services

Distributed Trace Services

Distributed Security Services

Policy and Repository ServicesResource Instrumentation Services

....................

Distributed Resource Management Across Heterogeneous Platforms

Common Base for Autonomic Management Common Base for Autonomic Management SolutionsSolutions

autonomic management interfaces exposed as GRID services

Supports the creation ofSupports the creation of virtual resources virtual resources as as stateful OGSA service instances stateful OGSA service instances

Open Grid Services Architecture (OGSA)Open Grid Services Architecture (OGSA)

fc

Web Services


ISVSolutions



Application Servers


OGSA Based Architecture FrameworkOGSA Based Architecture Framework

Common autonomic construct for all system elementsDistributed components and systems integrated as one virtual operating system

fc

To enable secure and coordinated access and sharing of To enable secure and coordinated access and sharing of heterogeneous and geographically distributed resources heterogeneous and geographically distributed resources (applications, data, processing power, storage capacity and other) (applications, data, processing power, storage capacity and other) across organizational boundariesacross organizational boundaries

Grid ComputingGrid Computing

fc

Grid ComputingGrid Computing

Grid Middleware (OGSA)

PhysicalPhysicalresourcesresources

Virtual Virtual environmentsenvironments

UsersUsers

Open Grid Services Architecture (OGSA)Open Grid Services Architecture (OGSA)

fc

SLAs protection by coordinating use of available resourcesSLAs protection by coordinating use of available resourceslocal grid, remote grid, provisioning

Multiple mixed workloads on the same GRID infrastructureMultiple mixed workloads on the same GRID infrastructure

Reactive mode (monitoring) or predictive mode (active forecasting) Reactive mode (monitoring) or predictive mode (active forecasting)

Middleware

Storage

Transaction Servers



Edg

e S

erve

rs

Web Pres.Servers Database

Servers

Virtualization layer


A WebSphere on demand Technology A WebSphere on demand Technology

fc

Resources Virtualization ValueResources Virtualization Value

Resource Virtualization can help to optimize the IT Resource Virtualization can help to optimize the IT infrastructure and:infrastructure and:

Reduce overall complexityReduce overall complexity

Standardize policies, procedures and operationsStandardize policies, procedures and operations

Drive adoption of open standards, end-to-endDrive adoption of open standards, end-to-end

Simplify systems managementSimplify systems management

Facilitate business continuity and disaster recoveryFacilitate business continuity and disaster recovery

Enhance resources utilization and optimize scalabilityEnhance resources utilization and optimize scalability

Optimize performance, availability, operational resiliencyOptimize performance, availability, operational resiliency

Improve total cost of ownership (TCO)Improve total cost of ownership (TCO)

Accelerate e-business integrationAccelerate e-business integration

Position IT infrastructure for the on-demand business modelPosition IT infrastructure for the on-demand business model

fc

Why LinuxWhy Linux

IT Infrastructure Optimization: Linux is a key playerIT Infrastructure Optimization: Linux is a key playerLinux lowers Total Cost of Ownership (TCO)Linux lowers Total Cost of Ownership (TCO)Linux allows application portabilityLinux allows application portabilityLinux enables higher levels of manageability (ex. backup and recovery)Linux enables higher levels of manageability (ex. backup and recovery)Linux provides operational flexibility (ex. test and temporary environments)Linux provides operational flexibility (ex. test and temporary environments)............................................

CIO/CEO talking pointsCIO/CEO talking pointsLinux is an Open Source operating systemLinux is an Open Source operating systemLinux can be tailored to meet individual business needsLinux can be tailored to meet individual business needsLinux is free or available at very low costLinux is free or available at very low costLinux runs on more hardware platforms than any operating system available Linux runs on more hardware platforms than any operating system available todaytoday

fc

IBM Commitment to LinuxIBM Commitment to Linux

IBM commitment:IBM commitment:has invested 1bn USD+ in Linux related developmenthas invested additional 300mill USD into Linux services development (implementation, consulting, education, ...)has created a new organization for Linux has made all IBM server platforms Linux-ready (iSeries, pSeries, xSeries & zSeries)has engaged closely with the Linux community to help Linux evolve (7,000 IBMers working in porting, research, services centers and development labs) has made IBM technologies available to Linux and open source communitieshas initiated Joint Customer Project for business and technical validationhas announced the IBM products Linux ready & established Linux distributions (more than 50 software products are available on Linux - Data Management, WebSphere Software, Lotus and Tivoli Software)Over 4,200 ISV applications have been ported to Linux has disclosed plan for next 12 months......has worked on more than 6,500 customer Linux projects in 2002120,000 competitive servers have been replaced by IBM Linux Servers

"IBM has made all of its server platforms Linux friendly, "IBM has made all of its server platforms Linux friendly, including zSeries, iSeries, pSeries and xSeries servers"including zSeries, iSeries, pSeries and xSeries servers"

fc

Responsive

Variable

Focused

Resilient

BusinessObjectives

e-business on demand Summarye-business on demand Summary

Business Efficiency

Operation Flexibility

Time to Market

Competitive Advantage

................

Value

Open

Integrated

Virtualized

Autonomic

IT AttributesInitiatives

Business Transformation

Utility Services

Operating Env. Application System

Fulvio CapogrossoDistinguished EngineerServer Group, South Region, EMEA

End Presentation

autonomic computing grid computing...

Documents