Virtualización & Grid para soluciones basadas en SOA

Presentación del tema: "Virtualización & Grid para soluciones basadas en SOA"— Transcripción de la presentación:

1 Virtualización & Grid para soluciones basadas en SOA
Octavio Bustos IBM Latinamerica Systems Tech. Advocate

2 Los elementos de SOA … un Servicio ? una Tarea repetitiva de negocios – ej., validar credito del cliente, crear una nueva cuenta. … Orientación a servicios? Una forma de integrar su Negocio como servicios interconectados y los resultados que ellos producen. … Arquitectura orientada a servicios (SOA)? Un estilo de arquitectura de IT que soporta la orientación a servicios … una aplicación compuesta ? Un grupo de servicios relacionados e integrados que soportan procesos de negocios construidos en SOA. Main Point: Services are repeatable business tasks. Business processes are a series of services snapped together like building blocks. SOA is an architectural style that makes this possible Let’s start by looking at some base-line definitions so we’re all talking about in the same terms. First of all, what is a service? <read definition> It’s important to stress that we’re talking about a part of a business process here. Don’t think about software or IT. Think about what your company does on a day to day basis and break those business processes up into repeatable business tasks or components. If you look at the graphic in the middle, this is the analogy of building blocks <do NOT use the word “Legos”> snapping together to build a structure. Services are the building blocks and they are snapped together into a business process. Second, what is Service Orientation? Building on our definition of a service, Service Orientation is a way of integrating your business as linked services and, more importantly, the outcomes that they bring. We’re still not talking about technology; we’re talking about a thought process and a business philosophy. What is SOA? It is quite simply the IT architectural style that supports the Service Orientation thought process makes it a reality. And finally, what is a Composite Application? <read definition> So composite applications are the actual running services that have been assembled and strung together to support the what your business does. SOA helps make building and adjusting composite applications fast and easy. **Main point** SOA makes it easy to snap together services into a business process just like snapping together building blocks into a structure.

3 Que desea para sus Servicios ?
Perspectiva de Infraestructura Service Service Que corra en cualquier parte… Que sea movil. Que sea flotante, Que pueda iniciarse y detenerse. … on demand Que pueda ser clonado ! Que pueda tener más de uno Que tenga capacidad de acceso a datos e información. Que este disponible. Que sea seguro. Service Composite Service Que corra en cualquier parte… Que sea movil. Que corra en Línea o como Job (batch) Que pueda ser agendado. Que tenga capacidad de acceso a datos e información. Que este disponible. Que sea segura. Composite Application

4 El reto La complejidad en la infraestructura de IT inhibe la agilidad, el rendimiento y la capacidad de aprovechar al máximo el valor de SOA. Las compañías necesitan adaptar su infraestructura de IT para ser exitosos en SOA. “Para 2008, 70% de las organizaciones de IT fallarán en la selección e implantación exitosa de una estragegia SOA al primer intento”. — 0.7 probability, Gartner, June 2005

5 Virtualización

6 Que es virtualización? Representación lógica de recursos sin las restricciones presentadas por las limitaciones físicas. Crear múltiples recursos virtuales sobre un mismo dispositivo físico. Extenderse más allá de la caja – ver y administrar recursos virutales como si fuese uno. Cambiar y ajustar dinámicamente por toda la infraestructura What is virtualization? It is a logical representation of resources not constrained by physical limitations. t’s a couple of things, and one way of looking at virtualization is taking something large, carving it up and making it look like many small things. Think of that as partitioning an Intel server using technologies like the VMware. Another way of looking at it is making many small things work in concert as one large thing. A good example of that is storage virtualization, where we can have multiple storage arrays from different vendors, work in concert as if they were one large single storage pool. And, finally, it’s being able to dynamically change and adjust these resources across the infrastructure. Think of that as being able to do what we could do in the mainframe for years, being able to extend that capability across a heterogeneous distributed enterprise. IBM Virtualization Engine A comprehensive platform to help virtualize the infrastructure

7 Valor de una infraestructura virtualizada
Mejora el TCO Reduce el costo de mantenimiento. Incrementa la utilización del activo. Conecta el desempeño de la infraestructura a los objetivos del negocio. Acceso a una infraestructura compartida. Aprovecha herramientas comunes a muchos servidores. Mejora la continuidad del negocio y la seguridad. Establece una base sólida para SOA We started out talking about what the motivators were for virtualization among our clients. And, again, the vast majority of those were TCO related, trying to drive up the asset utilization, decrease management costs, federal link infrastructure performance to the business goals and that’s pretty clear, I hope, from all the example I’ve been taking you through. What’s also clear, though, to us, from literally thousands of engagements is that TCO is not the only benefit of a virtualized infrastructure – that there are two other very, very important values. The second is this idea of access to a shared infrastructure, this ability to breakdown these islands of computing, these puddles of data, these stovepipe business processes and better integrate them into a single integrated virtual compute fabric, with a single virtual pool of information, and in so doing allowing clients to have common tools to manage across heterogeneous systems, allow them to increase their business resilience, their application availability and the security of the environment. Now, you might wonder how this has improved security. Many of our clients are finding that they have 10, 20 or even 30 copies of data across their infrastructure. Because their data is so fragmented, because their physical infrastructure is so fragmented, it’s impossible to identify where all those data exist. And, in many cases that data is redundant and in many cases there are copies of that data that should have been eliminated a long time ago, because it was back level. By having a virtualized environment, we have a single virtual pool of information and the access is integrated. You can implement a single access control mechanism for users. You can eliminate redundant or back-level copies of data, so you can improve security and, in fact – and very important to our clients today – you can improve their ability to respond to discovery motions and to know that as they respond to those discovery motions that they’re going to find the data that they expect to find and not a lot of back-level copies that they wished they didn’t have. We also believe that virtualization is the cornerstone for building the infrastructure necessary to implement a services-oriented architecture. So, for those of you who have customers who are embarking on SOA, we would say that virtualization is a key enabler for implementing SOA. And, finally, virtualization can increase flexibility and freedom on choice, particularly true in the storage world, which has long been characterized by monolithic vertical integration. Virtualization allows clients to much more freely mix and match technologies to have a single console for managing distributed heterogeneous resources, allowing them to lower the switching cost, more flexibility, more freedom of choice, masking complexity. Incrementa la flexibilidad Crea grupos de recursos del sistema. Mantiene libertad de elección y estándares abiertos. Simplifica al enmascarar la complejidad.

8 Simplificar el entorno de servidores y almacenamiento
Public Internet/ Intranet Clients Routers (Layer 3 Switches) Firewalls SSL Appliances Web Servers Layer 4-7 Switches Caching Appliances File/Print Servers Layer 2 Switches Web SCM Mgmt Gate FS WS ERP Gate ERP App Sec Storage Fibre Switches This is a greatly simplified view of a typical customer environment. Simplified because there is only one SAN, all the servers are neatly aligned, etc. Note that the administrator, which is where most of the cost of running the environment lies, is faced with multiple tools/consoles for managing this environment. The LAST thing the client wants is to add to this with yet another tool set as we introduce virtualization CLICK The animation shows the power of virtualization to greatly simplify the physical infrastructure. Thanks to Java and Linux, applications are no longer tethered to specific platforms – they can be rehosted on scale out (blades) or scale up (System i, System z, System p) servers. Important to note that we are showing NOT ONLY the consolidation of similar applications on to larger systems (ie multiple instances of a database or web server) but actually collapsing whole layers of the physical infrastructure – virtualizing the network as well as the servers. This is enabled by the fact that the virtualization technologies in our systems (derived from 40 years of mainframe experience) is mature, stable, secure enough to allow clients to confidently deploy mixed workloads onto a single system (ie batch + transactional; development + production; web facing + core data). This is not something that clients can do on Sun, HP, or x86. The net result is substantially higher utilization rates (because mixing workloads allows you to take advantage of the peaks/valleys of different applications), substantially improved performance (because networks are virtualized and communication is at memory speeds), significant reduction in infrastructure complexity, and a large reduction in administrative overhead (fewer things to manage, fewer consoles and tools). But this only addresses the server side. Click to next page to discuss the storage side of the infrastructure. Storage Area Network

9 Storage Virtualization
Facilitar el movimiento de Datos y aplicaciones Public Internet/ Intranet Clients Routers (Layer 3 Switches) Firewalls App Sec ERP FS 7 WS Mgmt ERP Web SCM Gate Gate Storage Area Network Storage Virtualization Storage Pool Beneficios Incrementa la disponibilidad de la aplicación Simplifica la migración de Datos Flexibilidad Clients are concerned about application availability and virtualization can dramatically improve availability by facilitating the movement of both applications and of data with minimal, and in many cases zero, downtime. Applications can be moved dynamically from one system to another – whether it be between blades in a bladeserver or between partitions on a single or multiple machines. Technology such as VMware’s V-motion can be used to demonstrate moving a running application from one virtual machine to another on a different physical machine without disruption of the application. One of the major contributor to application outages is data migration – either moving data from one system to another or expanding the amount of storage associated with an application. Through virtualization, the application can be insulated from the changes in the underlying physical infrastructure and dramatically improve availability. We can demonstrate moving a movie file from one array to another while the movie is playing without any disruption. This means that storage can be allocated to applications as it is needed and not over-provisioned ahead of time. This is a significant factor in the low utilization of today’s physical storage environments. Another benefit is that storage migrations can be accomplished much more quickly – often in hours as opposed to days or even weeks. This reduces the need for lease overlaps and reduces admin costs – clients often report that SVC pays for itself in a single data migration. Finally, this insulation and increased mobility means that clients have much greater flexibility and freedom of choice. The can more readily change from one vendor to another based on feature/function, price/performance, service/support as opposed to be locked into a single vendor.

10 Storage Virtualization
Incrementar la disponibilidad del Negocio Public Internet/ Intranet Clients Firewalls Routers (Layer 3 Switches) Web FS 7 WS Mgmt ERP SCM Gate App Sec ERP Storage Area Network Storage Virtualization Storage Pool SAN Storage Pool Storage Pool Beneficios Menores costos en recuperación de desastres. Reducir caídas planeadas Menor tiempo de recuperación In the wake of disasters such as Hurricane Katrina and 9/11, many clients are focused on improving the resilience of their infrastructure and implementing disaster recovery or business continuity plans. Virtualization can help clients address this in multiple ways. First, by simplifying the infrastructure, it makes it much simpler to mirror. By improving their overall TCO (through increased utilization, reduced hw/sw/admin costs, etc) clients find they can deploy their budgets and energies to addressing business continuity. Virtualization also allows clients to lower the cost of a DR site by allowing ‘mix and match’ of technologies. In the physical world, the norm is to require an exact mirror image of the production environment in the secondary site – even if the client only intends to use that site in the event of a disaster. So it makes no financial sense to implement the same quality of service in the secondary site (in terms of performance for example) as at the primary. (Note that some clients do run “hot – hot” – that is they run their production work at a secondary site one week a month perhaps. In this case they will require the same qualities of service at their secondary site.) Storage Virtualization allows clients to put lower cost storage (from different vendors, of different generations) in the secondary site. Virtualization (bother server and storage) can also reduce recovery time allowing clients to meet their recovery point objectives. Several of the customer examples that follow later in the presentation demonstrate this. Storage Virt

11 Plataforma de Virtualización
Workload Virtualization Information Virtualization Principios Clave Completa Abierto Heterogéneo Conocimiento compartido Standards and Open Interfaces Virtual Access and Management The IBM virtualization platform has been built keeping some key principles in mind: The platform must be comprehensive, and we have the most comprehensive platform in the marketplace. This includes technology that is part of the DNA of our server and storage systems, and management of those systems, either individually, or as an integrated pool. Beyond that, we also virtualize application workload and information. The platform must be open. On the next chart, we’ll discuss some of the standards that we’ve implemented in our products. Most enterprises use products from multiple vendors. Delivering products using open technologies allows our products to interoperate more smoothly across your infrastructure, giving you a high degree of flexibility. The platform must deliver function beyond IBM server and storage systems. It must be heterogeneous. The platform delivers technologies that leverage common skills across many different platforms, IBM and non-IBM. Resource Virtualizers Servers Storage Networks

12 Plataforma de Virtualización
Standards and Open Source Workload Virtualization Information Virtualization Servers Networks Storage Virtual Access and Management Resource Virtualizers Workload Management Usage and Accounting Resource Management Discovery and Mapping Hypervisors Containers Block Virtualization VIO VLAN Provisioning Orchestration Federation Cleansing Transformation Global File System Resource Pooling Scheduling Let’s look at the Virtualization Platform in more detail. Forming the base of the platform, you see servers, storage and networks. At an individual device level, these physical resources are virtualized in many ways. Hypervisors facilitate virtualization in servers and even some storage devices. Block virtualization is leveraged by our San Volume Controller, and also virtualize I/O and networks with technologies in our servers. Moving to the middle of the page, you see capabilities that not only benefit a single server, but the technology can also be applied to multiple systems, within a box or across an infrastructure. You can manage a workload that traverses several systems as a single workload. Using the same tools, you can management many different kinds of platforms. We have a tool that will measure usage across many elements in a physical or virtual infrastructure, and technology to discover resources across the environment, creating a topology of the relationships between those resources. Complimenting these system virtualization technologies are the capabilities to virtualize workloads. These include scheduling, provisioning, orchestration, and pooling. In addition, we provide many ways to virtualize information, including federation and the global file system.

13 IBM Virtualization Engine™ Platform
Workload Virtualization Information Virtualization Provisioning Manager Intelligent Orchestrator Workload Scheduler Load Leveler Business Grid Information Integrator Universal Database SAN File System General Parallel File System Servers Networks Storage Resource Virtualizers Standards and Open Source VMWare, Xen, Microsoft Virtual Server IBM Advanced Power Virtualization, z/VM, San Volume Controller VLAN, VIO Resource Dependency Service Enterprise Workload Manager Virtual Access and Management Usage & Acctg Manager Director TotalStorage Productivity Center In the last chart, we talked about the virtualization capabilities of the platform. Now let’s take a closer look at how some of those capabilities are delivered. Remember when we discussed one of the elements of virtualization, creating many small things out of one large? This is what is done when servers are partitioned. We use hypervisors in our POWER5 and mainframe systems to create partitions. In the x86 space, we partner with other companies that deliver virtualization products such as VMware, Microsoft Virtual Server, and Xen, a new emerging open source technology. SAN Volume Controller facilitates storage virtualization, benefiting either a homogeneous or heterogeneous storage infrastructure. Within many of our servers, we virtualize I/O connections as well as network connections. The number of connections is not constrained by the number of physical devices. The technologies here that deliver management across a virtualized infrastructure benefit either a single system, or link multiple systems together. Enterprise Workload Management delivers resource management and monitoring across a heterogeneous environments. IBM Usage and Accounting Manager can collect utilization data for elements across a virtualized infrastructure. IBM Director extends beyond the bounds of the original xSeries and BladeCenter environment where we first delivered it. IBM Director can now manage systems across our server line. Resource Dependency Service discovers servers, storage, networks and more across the infrastructure. Total Storage Productivity center provides a single point of control for data and storage. Complimenting system virtualization, products that provide workload virtualization include WebSphere Business Grid and several Tivoli products including: Provisioning Manager, Intelligent Orchestrator and Workload Scheduler. Products that virtualize information include: WebSphere Information Integrator and DB2.

14 Virtualización de Recursos Rompiendo las barreras de los recursos Físicos
LPARs para compartir recursos e incrementar la utilización. Aumenta la Disponibilidad con clusters. Virtualiza Redes e I/O para eliminar restricciones y simplificar la infraestructura. Virtualización de Almacenamiento para mayor utilización e incrementar la flexibilidad. Virtual Servers OS Device Driver OS Device Driver OS Device Driver OS Virtual I/O For decades, IBM has been using virtualization in the mainframe, which still holds the gold standard. Partitioning allows a customer to carve a large server into many small virtual servers, better utilizing resources and reducing cost. We’ve taken the technology that was born on the mainframe and are now exploiting it in our mid-range servers. For instance, in a POWER5 System p, a slice as small as 1/10th of a processor may be allocated to a virtual system. Virtual LANs and Virtual I/O break the constraints of physical resources. Network and I/O connections are no longer limited to the number of physical connections or adaptors. This allows the customer to add connections much more dynamically. Performance can also be improved by using “networks in a box”. When applications communicate between partitions, the network calls remain in the server reducing network latency. Clustering adds many options for high availability. Block virtualization, delivered by SAN Volume Controller, allows a customer to better utilize storage resources. Many heterogeneous storage devices can be combined behind an SVC. The SVC is also used to improve availability and offers business continuity. Cluster … SAN Hypervisor SAN Volume Controller Virt. Ethernet LAN Hypervisor Physical Server

15 Virtual I/O server (POWER5)
Optimizar Recursos en un solo servidor Incremente el uso del activo y la flexibilidad Hypervisor Ofrece: Mayor utilización de recursos. Mayor flexibilidad y capacidad de respuesta. Mejora la calidad en el servicio (QoS) en las cargas de trabajo. Estado del arte en virtualización. System z and z/VM Estándar para virtualización System p y System i Capacidades únicas incluyendo micro particiones y LPAR dinámicas. Virtualización de Redes y Virtual I/O, múltiples sistemas operativos en un servidor. Virtual I/O server (POWER5) Partition 1 . . . Partition N Virtual networks Hypervisor Virtual CPU Virtual Memory What you see here is a picture again of a hypervisor that we implement in our System z, System p and System i. This hypervisor helps enable the dynamic partitioning capability that offers fine-grained partitioning. System Z has set the gold standard. System p and System 5 offer a capability that is absolutely unique in the UNIX marketplace. CPU, I/O, and memory resources are shared among partitions, all managed by the hypervisor. The servers also have the capability to host multiple operating systems within one physical server. For example, a POWER5 server is capability of running AIX, i5/OS and Linux operating systems. A feature of System p, Advanced Power Virtualization, offers the capability to create partitions using as little as 1/10th of a processor. APV also delivers the Virtual I/O capability. VIO is an example of a server running in a partition which may be referred to as a utility partition. CPU resources I/O resources Memory resources Server Resources

16 IBM TotalStorage SAN Volume Controller
Linux UNIX Windows Virtual disks IBM EMC HDS HP Storage Pool Beneficios Ayuda a direccionar el explosivo crecimiento en almacenamiento Mejora la utilización del activo y el rendimiento. Migración de datos sin interrupciones. Crea capas de almacenamiento Mejora la continuidad del negocio.

17 TotalStorage Productivity Center
Administración y Acceso a recursos Virtuales Vea y administre muchos recursos virtuales como uno solo Simplifica la administración mediante el uso de herramientas comunes en toda la infraestructura. Visualiza y administra recursos físicos y virtuales. IBM Director TotalStorage Productivity Center

18 Software Distribution
What Do You Want to Do? BladeCenter Management Scheduler Manage from a single console Get Systems Up and Running Fast Real Time Diagnostics Troubleshoot problems quickly Software Distribution RDM Keep the system software up-to-date Deploy the OS and applications to multiple systems Monitoring Hardware Status Let me know if my servers are healthy VMM System Availability Maximize the utilization of my servers Capacity Manager Monitor the performance of my servers

19 Administración y Acceso a recursos Virtuales Vea y administre muchos recursos virtuales como uno solo Permite a la infraestructura responder a demandas del Negocio. Descubrimiento e identificación de recursos. Determina el costo de suministrar los servicios de IT. Enterprise Workload Manager Resource Dependency Service IBM Usage and Accounting Manager

20 ? Donde esta el Problema? Problemas de Tiempo de respuesta?
zSeries ? GOAL: QoS measures the actual performance of customers visiting your Web site. Speaking points: QoS measures the actual performance of customers visiting your Web site. The Quality of Service Monitor reports on the performance of your Web services. This is the piece that measures real customer experiences without installing any code in end-users’ environments. The QoS Monitor measures total round-trip time, back-end processing time, and client browser render time. It also supports statistical sampling on customer visits. QoS monitor can also generate alerts when constraints are violated. Monitors actual end user response times No client code required Alert on three response time metrics Round trip, backend service time and page render time Here’s how it works: The QoS monitor sits in front of your Web Server. When it sees a request, it starts a timer at TimeStamp 1. When the back end has processed the request and sends a response, the QoS Monitor puts TimeStamp 2 on it and attaches a small amount of code (260 bytes--5 lines) and sends it on to the client. The client starts a timer when the page starts to render and stops the timer when the page has completed loading and sends the results back to the QoS monitor. Based on time stamps, you now know back end processing time, client render time, and total round-trip time. Granted, the round-trip time isn’t exact because you didn’t measure the initial request time from the client, but the last leg back to the QoS monitor serves as a good proxy for how long the original request took. And it doesn’t matter that the server’s clock and the client clock aren’t in sync because the only significant measures here are the deltas. All of this information is reported to the Tivoli Internet Management Server. QoS raises an event if a constraint is violated. Real-time graphs of the 3 differentials are available through the management server console. Problemas de Tiempo de respuesta? Bajo rendimiento de la infraestructura? Problemas de aplicación ? Como determinarlo?

21 Un sitio para iniciar? La perspectiva del usuario final.
zSeries Entender el nivel de rendimiento y disponibilidad en el usuario Final. Ver tendencias en el tiempo de respuesta.

22 Acciones Autónomas de Administración – Balanceo de Cargas
zSeries EWLM Resource EWLM provee recomendaciones El balanceo de cargas modifica el comportamiento basado en las recomendaciones. Request Origins Load Balancer EWLM Resource EWLM Resource EWLM Domain Manager

23 Acciones Autónomas de Administración – LPAR
CPU resources Memory I/O Hypervisor Partition 1 Partition N . . . Server Resources Virtual I/O server (POWER5) Virtual networks Virtual CPU Virtual Memory Automáticamente asocia recursos de procesador a las demandas del servicio o aplicación. Optimización de múltiples servidores basada en objetivos de negocio predefinidos. Asignación de recursos basado en tiempo de respuesta e importancia del negocio. Decisiones de asignación basadas en vistas del servidor o grupo de servidores.

24 Virtualización de Cargas e Información SOA habilitado para GRID

25 Virtualización de Cargas e Información
El corazón de una infraestructura SOA habilitada para GRID Understand Cleanse Transform & Move Federate Parallel Processing Common Connectivity Information Server Data Content Discover, define, model, and govern information quality and structure Standardize, merge, and correct information Transform, enrich, place, publish and replicate information Virtualize access to disparate information Unified Service Deployment Unified Metadata Management Enterprise Service Bus Information Virtualization Framework virtualized information global name metadata “engine” management, choreography, etc. Content & Format Virtualization Transformation Federation Data Cleansing Data Analysis any sources, in any format Location & Performance Virtualization Caching Replication Data Movement & Placement Quality of Service replica cache access anywhere with quality of service query / status Job Resource Requirements (JSDL) Workflow Dependencies (BPEL) Scheduling Provisioning & Orchestration agreements exceptions resource pool add / delete Workload Management submit results Grid Middleware Workload and Information Virtualization Workload Virtualization Information Virtualization Autonomic

26 Donde desea estar ? Servicios Virtualizados y Dinámicamente Reconfigurables. Silo Services Composite Virtualized Dynamically Re-Configurable Componentized Integrated Modules Services Process Integration via Services Dynamic Application Assembly Components Objects Structured Analysis & Design Service Oriented Modeling Grammar Component Based Development Object Ad hoc IT Governance Emerging SOA Governance SOA and IT Governance Alignment Platform Specific Neutral Dynamic Sense & Respond Monolithic Architecture Emerging SOA Grid Enabled SOA Dynamically Re-Configurable Architecture Component Architecture Layered Architecture Function Business View Organizatio n Methods Applications Architecture Infrastructure Level 1 Level 4 Level 5 Level 6 Level 7 Level 3 Level 2 IBM’s SOA Maturity Model

27 Sinergia entre Virtualización/GRID y SOA
Virtualization and Grid Separa Aplicaciones de Servicios Separa aplicaciones y servicios de la infraestructura Application Development, Enablement & Tooling: “IBM Research work on X10 continue (PERCS)” and “IBM participation in Eclipse PTP” As part of the government funded “PERCS” (Productive, Easy-to-use, Reliable Computing System) project IBM continues to work on an experimental parallel programming language (based on Java) called X10. X10 offers many features and benefits which promise to make the semantics of parallel programming easier. Also, IBM Research led IBM’s involvement in the Eclipse parallel tools project (PTP) working group. This was important work to continue to evolve IBM’s parallel technologies and tools into working within a common framework (an IDE) based on Eclipse. PTP will be an important strategic counter-move to Microsoft’s threat with CCS (Compute Cluster Server). PTP Overview Designing and developing parallel programs (that is, multi-process, not just multi-threaded programs) is an inherently complex task. Developers must choose from the many parallel architectures and programming paradigms that are available, and face a plethora of tools that are required to execute, debug, and analyze parallel programs in these environments. Few, if any, of these tools provide any degree of integration, or indeed any commonality in their user interfaces at all. This further complicates the parallel developer's task, hampering software engineering practices, and ultimately reducing productivity. One consequence of this complexity is that best practice in parallel application development has not advanced to the same degree as more traditional programming methodologies. The result is that there is currently no open-source, industry-strength platform that provides a highly integrated environment specifically designed for parallel application development. In order to address this deficiency, the Eclipse Parallel Tools Platform Project (PTP) aims to extend the Eclipse framework to support a rich set of parallel programming languages and paradigms, and provide a core infrastructure for the integration of a wide variety of parallel tools. The PTP project will provide a comprehensive platform that addresses many of these problems facing parallel application developers today, including: no standard, portable, parallel IDE no existing open-source parallel debugger the proliferation of stand-alone parallel tools the inability of existing tools to interoperate the poor scalability and reliability of existing tools The Eclipse PTP will deliver a portable, scalable, standards-based parallel tools platform that will enable the integration of tools specifically suited for parallel computer architectures. These tools will provide functionality such as: parallel debugging performance analysis and profiling message traffic analysis data visualization problem solving environments model coupling automated parallelization data parallel languages By providing a common parallel tools framework, PTP will also encourage tool developers and vendors to adopt a highly integrated approach to their tool design. By decoupling the user interface requirements, tool developers will also be able to focus on their core tool functionality without having to waste resources on maintaining the infrastructure needed to support their tools. The result will benefit parallel application developers by ensuring that a wide range of highly integrated parallel tools will be available to meet their development needs. Procesos de Negocio y Agilidad en las aplicaciones Agilidad en la Infraestructura Aplicaciones Dinámicas Recursos Dinámicos

28 Resumen

29 IBM Virtualization Engine Platform Enhances Each IBM System
System i and iSeries Long heritage of virtualization- supports multiple OSs Strength is systems management – leverage VE console/Director with iSeries Navigator to bring Intel processor based servers into a more disciplined management environment Host RDS server & Director servers, EWLM server System p and pSeries Advanced POWER Virtualization puts it in a class of its own for UNIX virtualization EWLM coupled with APV offers dynamic resource management based on business goals Host RDS server & Director Server System x and BladeCenter BladeCenter – multiple OS support, virtual i/o, virtual networking Partner w/ VMware for fine-grained virtualization Leverage Director expertise to manage other platforms, host RDS serve System z9™ and zSeries Built on nearly four decades of IBM virtualization heritage and technology enhancement Positioned to become a business hub of the enterprise Host managers for EWLM, RDS, & Director TotalStorage Leverage SVC for higher availability and utilization – set up for ILM DS8000 – the most virtualized storage server

30 Virtualization Leadership
IBM can virtualize up to 80% of a client’s infrastructure Over 30,000 UNIX, mainframe, and System i customers exploiting systems-level virtualization System x customers deploy over 1,000 virtual servers a day IBM is the leading reseller of VMware 2,000 storage virtualization customers, +5 every day Over 3,400 Virtual Tape Systems supporting 1 Exabyte of data Over 500 grid implementations Hundreds of in-depth TCO studies IBM has had the privilege of working with thousands of clients over the past four decades to help them exploit our industry leading capabilities

31 ibm.com/systems/virtualization
Summary Develop a strategy … think holistically Start simple … start now Maintain flexibility … standards are key Select a partner with experience Improve TCO Access Through Shared Infrastructure Increase Flexibility ibm.com/systems/virtualization

32 IBM Virtualization Engine™ Platform
Workload Virtualization Information Virtualization Provisioning Manager Intelligent Orchestrator Workload Scheduler Load Leveler Business Grid Information Integrator Universal Database SAN File System General Parallel File System Servers Networks Storage Resource Virtualizers Standards and Open Source VMWare, Xen, Microsoft Virtual Server IBM Advanced Power Virtualization, z/VM, San Volume Controller VLAN, VIO Resource Dependency Service Enterprise Workload Manager Virtual Access and Management Usage & Acctg Manager Director TotalStorage Productivity Center