2 Expert systems to power a smarter planet. With intelligence being infused into the systems, processes and physical objects we interact with every day, our planet is becoming smarter before our eyes. Less visible, but equally important, are the leaps in intelligence being achieved by the computing systems at the heart of this progress. This is about far more than processing speed; it s about intelligent architectures that allow us to build expertise directly into our computing systems expertise that can support today s cloud services and big data requirements. Case in point: a new class of computers called expert integrated systems, designed with scale- in architecture and built- in expertise. Combined, these ideas will radically simplify enterprise computing. support twice the peak application load of previous- generation IBM systems. We can now get systems up and running in as little as four hours and, using patterns, we ve been able to cut months off of application deployment times. Results like this are encouraging, but even more encouraging is that over 100 software vendors like SAP and Infor are already using PureSystems open design to develop their own unique patterns. We expect the library of patterns to grow as our clients begin to explore the possibilities of this approach. Let s build a smarter planet. ibm.com/expertise Expert integrated systems scale in, driving performance more intelligently not by adding processors, but by tightly integrating the processor with the memory, storage, networking, virtualization, management and middleware layers above it. And unlike today s converged IT solutions, these systems don t just integrate IT components, they integrate expertise as embodied by something we at IBM call patterns. Patterns distill the collective knowledge related to a specific field or task. They balance and coordinate the complex array of resources required to deliver vital capabilities ranging from sophisticated services like business analytics and cloud environments to bedrock IT functions like application deployment and database provisioning. Recently, IBM introduced the world s first family of systems with integrated expertise, IBM PureSystems. Early testing is promising. The scale- in architecture of PureSystems can
3 Expert Integrated Systems: Changing the Experience & Economics of IT Table of Contents Introduction IT Burden Enter Expert Integrated Systems The Pattern Principle Exploring the Benefits of Expert Integrated Systems Conclusion Introduction It s never been more challenging to work in IT. The continually shifting demands of business and other organizational environments demand a quicker and more effective response from those charged with matching technology to enterprise goals. At the same time, IT faces increased pressures to reduce costs, as markets worldwide force ever- greater efficiencies in enterprise endeavors. Ironically, while responding to enterprise needs, IT has become ensnared by complexity. Options meant to help organizations meet their objectives have become obstacles in the way of IT s ability to match technology to business goals. Instead of actually solving problems for the businesses they serve, IT professionals spend most of their time and funds on the day- to- day procurement, configuration, and maintenance of various systems intended to solve those problems. Fifteen years ago, the largest IT budget item was server hardware and the associated infrastructure that went with distributed computing. Over time, however, the emphasis shifted. Today, IDC research shows that as much as 70 percent of IT costs are going to server management or administration (1). This emphasis on operational tasks is taking away from IT s ability to innovate and solve problems. It is time IT becomes the innovator of business solutions once again. Building on a trend toward the use of converged infrastructures and integrated systems in IT, expert integrated systems have emerged that add built- in patterns of expertise to simplify key tasks across all IT resources throughout
4 application lifecycles. The result is a simplified experience that improves IT economics and resource usage. In this e- book, we take an in- depth look at these expert integrated systems what they are, how they work, and how they have the potential to help CIOs achieve dramatic savings while restoring IT s role as business innovator. IT Burden Over the last decade, costs associated with managing and administering distributed systems have doubled, according to IDC. By 2013, the firm predicts that nearly 70 percent of all IT spending will fall in this area, far outstripping capital spending on servers as well as on power and cooling.(1) Figure 1: The majority of worldwide IT operating costs involve server management and administration.(1) One reason for this increase in operating costs is virtualization. On one hand, virtualization has revolutionized IT by enabling enterprises to move from over- provisioning hardware and software in order to accommodate computing demand to a model where an infrastructure of compute, networking, and storage elements can be allocated to workloads as needed. This is part of the reason that IDC predicts there will be less spending on new servers in 2013 than there was in (1) However, it has become a challenge to set up and maintain virtualized environments. Not only does virtualization require increased spending on compute, network I/O, and storage; it also calls for more extensive integration, configuration,
5 testing, and troubleshooting. The organic complexities this situation introduces will only increase as the number of virtual servers proliferates. up in a commissioned study conducted by Forrester Consulting on behalf of IBM, which obtained details from over 500 IT professionals worldwide about the time spent constructing large- scale applications. Figure 2: As virtualization has grown, so have the costs of managing the environment that supports it. (2) It isn t just virtualization that consumes IT resources, however. For both physical and virtual elements of the infrastructure, IT spends hundreds of days working on new projects as well as upgrades. This includes buying, installing, and maintaining hardware, middleware, open- source software, and applications of all sorts. For example, in a recent IBM survey, respondents reported that new IT projects require 11 months to complete and upgrades require nearly eight months. 3 Similar findings turned Figure 3: A recent survey of IT professionals indicates how time- consuming setup, configuration, and testing of new systems can be. Given these kinds of resource requirements, it s no surprise that according to a commissioned survey conducted by
6 Forrester, 25 percent of new IT projects in the US alone deploy over budget, and 34 percent deploy behind schedule.4 What I m looking for is to get the industry best- practice right out of the box, because that will speed our time to market so I can spend time being innovative, said Andrew Smith, VP, Corporate IT Application Hosting Service, McKesson, during a roundtable discussion in April. Figure 4: Enormous resource requirements cause substantial delays and budget overruns for IT. Perhaps the worst news of all, though, is that resource drain is causing IT to fall short of meeting its own cost- cutting goals. IBM research shows that roughly one in five datacenters is operating at the most efficient levels. Yet if these datacenters were able to improve efficiency, up to 50 percent of the IT budget would be available for innovation and new solutions. What s required, many experts agree, is simplification of the constraints that present systems put on IT. Enter Expert Integrated Systems Clearly, IT needs a new approach to computing that effectively simplifies the complexities that are currently stifling
7 innovation. The fundamental concept involves making various hardware and software components work together as one system that can automate, optimize, deploy, and manage IT tasks without the hands- on intervention that s eating up years of datacenter time. This is the approach of expert integrated systems, which are characterized by the following elements: Built- in expertise. Expert integrated systems capture and automate the collective knowledge of experts, culled from a wide variety of sources. This knowledge is implemented in patterns of expertise that literally automate what experts do at every level of the system, dramatically shortening deployment time and allocating system and application resources for optimal performance, security, and reliability. Integration by design. By integrating and lab- tuning hardware and software in a ready- to- go, workload- optimized system, expert integrated systems deliver one point of management and one management console for every element of the environment. These characteristics make expert integrated systems suitable for use in cloud services. Simplified user experience. Expert integrated systems can be ordered, unpacked, plugged in, and managed as one system with one interface. This eliminates the need for IT staff and lines of business to juggle multiple components in a complicated lifecycle. Expert integrated systems are a step beyond the concepts implemented in converged systems, in which applications and core infrastructure software are abstracted by virtualization and made distinct from processing, storage, and networking. According to IDC, there are three kinds of converged systems: Converged infrastructure systems comprise devices with systems and network- level resource control and automation software and some combination of server, storage, and network hardware sold as special SKUs and designed for general- purpose, distributed workloads. The IBM PureFlex TM System is an example. Converged platform systems comprise converged infrastructure systems, as defined previously, but include additional pre- integrated packaged software and customized system engineering optimized to enable application development software, databases, and testing and integration tools (e.g., business intelligence [BI], content management, database, and application server middleware). The IBM PureApplication TM System is an example. Converged certified systems are similar to converged infrastructure and converged platform systems, as defined previously, but the actual pre- integration, delivery, and customer support activities are provided by a third party other than the original equipment manufacturer (OEM). (5)
8 Expert integrated systems go further than these converged systems, integrating the compute, storage, networking, and platform middleware for both physical and virtual resources, and adding something more: built- in expertise that enables a vastly simplified user experience, making it possible for IT to focus its efforts on innovation instead of maintenance and management. According to IBM, there are two types of expert integrated systems: Platform systems. These systems include full application platforms with integrated expertise, which are designed and tuned to specific tasks, such as transactional Web and database applications. Infrastructure systems. These expert integrated systems offer control over servers, storage, networking, virtualization, and management along with integrated patterns of virtual and hardware resources. click Click on image to enlarge Figure 5: Expert integrated systems offer the means to free professional's time by automating routine tasks allowing then to focus on innovation and growth. IBM PureSystems TM The IBM PureSystems family, IBM's family of expert integrated systems, are defined as systems that are integrated by design, include built- in expertise and deliver a simplified experience. PureSystems offerings include: IBM PureFlex TM System is a comprehensive, flexible cloud infrastructure system with integrated expertise for IT organizations that want the simplicity of integrated and optimized servers, storage, networking, virtualization and management but prefer to install, configure and manage the operating system, middleware, database, runtime and application software as they do today. IBM PureApplication TM System is a comprehensive cloud application platform system with integrated expertise for IT organizations that not only want the simplicity of preintegrated infrastructure but also want integrated middleware, database and runtime environments combined with fully integrated platform monitoring and management capabilities. These systems are specifically designed for web and enterprise applications and data workloads. Three IBM PureData TM System models were introduced in October PureData System models are comprehensive data platform systems with integrated infrastructure and middleware optimized exclusively for specific Big Data needs: IBM PureData System for Transactions: optimized exclusively for transactional data workloads. IBM PureData System for Analytics: optimized exclusively for specific Big Data needs. IBM PureData System for Operational Analytics: optimized exclusively for operational analytic data workloads.
9 The Pattern Principle Patterns of expertise are one of the features that raise expert integrated systems above the level of converged systems and make them much more than an appliance or a box of parts with a recipe. Instead of presenting IT with systems that can be used in different ways to solve different problems, patterns capture and present proven best- practices and expertise gained from decades of IBM experience in optimizing the deployment and management of datacenters, software infrastructures, and applications around the world. These patterns are captured, lab tested, and optimized into a deployable form. best practices for dynamic scalability, high availability, security, and other features. The IBM Transactional Database and Data Mart Patterns: These patterns provide capabilities essential to provisioning and managing enterprise data services. They can be used to create DB2 databases optimized to support transactional (OLTP) and operational analytics. Below are two screen shots from a Web application pattern in IBM s PureApplication System: Click on image to enlarge Software patterns can contain the pre- defined architecture of an application. They can comprise a set of directions for installation, configuration, and management of components such as Web servers, databases, caching servers, LDAP servers, and system processes. Patterns also deliver repeatable deployment of an architecture for full lifecycle management. Examples of software patterns supported by the PureApplication System: The IBM Web Application Pattern: The pattern eliminates the need for high- value personnel to set up the parameters of Web- based applications. It codifies Figure 6: Patterns in IBM s PureApplication System demonstrate embedded expertise.
10 CASE STUDY: SugarCRM Leading CRM application provider SugarCRM has an application pattern for the IBM PureApplication System, substantially cutting deployment time from days to minutes and reducing virtually all configuration and deployment human errors. Best known as a SaaS vendor, SugarCRM supports flexibility of deployment, providing an on- premises private cloud option with PureApplication System. The SugarCRM pattern automatically provisions the entire SugarCRM stack, configuring and optimizing the application layer and platform middleware for its underlying infrastructure. Deploying SugarCRM via a pattern reduces setup time by up to 98 percent. The option of private cloud deployment demonstrates the readiness of the PureSystems family to interact with and enable cloud services. Notably, SugarCRM isn t Java- based being it is written in PHP. The adaptation demonstrates PureSystems open platform for delivering patterns of expertise, which also can be created for Ruby, C++, and custom applications. We won t have to spend as much time getting [users] up and running, SugarCRM CEO Larry Augustin told The New York Times. This is is going in exactly the right direction. Who Makes Patterns of Expertise? Patterns of expertise for IBM PureSystems come from a range of sources. IBM offers patterns of expertise gleaned from decades of client engagements, best- practices, research and development, and global experience with datacenter optimizations. What was once a book or written document has become a software product, complete with binaries and ready for deployment. Patterns are also created by third- parties. Via the IBM PureSystems Centre, IT professionals have access to solutions and applications from IBM and a range of IBM partners both for initial installation and for maintenance and upgrades. These include solutions for vertical sectors banking, education, financial markets, government, healthcare, insurance, telecommunications and energy and utilities. IT also can create its own patterns of expertise through a series of tools that capture custom virtual images, virtual application patterns, and existing expertise for operating system and middleware installation. IT also can combine pre- defined components to build new image packages and create a standard VM format to fit enterprise needs.
11 Exploring the Benefits of Expert Integrated Systems Expert integrated systems deliver value throughout the IT lifecycle. They offer substantial advances in efficiency, simplicity, agility, and control of system resources. For instance, expert integrated systems present enterprises with a broader, more open ecosystem of capabilities and solutions. Thanks to integral expertise, there is no delay or shortfall if an enterprise cannot produce the IT skills required for a particular project. Once selected, new applications can be up and running within hours, reducing deployment times that formerly extended for months. Because they are integrated by design, expert integrated systems optimize application resources. They consolidate hundreds of Web application workloads and drive up to 90 percent resource utilization to improve asset return fitting twice the number of applications per square foot of datacenter space compared with traditional deployments. (6) Expert integrated systems lower ongoing risks and costs by automating provisioning, scalability, and intelligent management. Resources adapt elastically to workload spikes, obviating the need for IT to purchase more software, hardware, or bandwidth to meet temporary conditions. And by ensuring adherence to enterprise security and architecture policies automatically, time spent on compliance is reduced. While every enterprise will capitalize on the benefits of these improvements in its own fashion, the all- around efficiency and speed provided by expert integrated systems frees up resources for innovation and allows IT to focus on responding to business needs more quickly. Conclusion IT faces a turning point. In most enterprises worldwide, the complexity of technology and the sheer resources required to procure, deploy, and maintain solutions are interfering with IT s ability to deliver solutions. The old world of IT was complex equals powerful, said Peter Hinssen, CEO of the Across Technology consultancy, in a roundtable discussion during the PureSystems introduction. We re now getting into an era where simplicity is powerful. Expert integrated systems offer a way for IT to free itself from the burden of inefficient systems, not only by integrating essential elements and making systems easier to use but by the infusion of patterns that allow IT to set up applications, datacenters, software infrastructures, and cloud services that deliver business value instantly. We re trying to get closer to the business, to be more innovative, said Hinssen. At the same time we want to do what we do centrally in the easiest, most cost- efficient, most economy- of- scale way that we can possibly imagine. Expert integrated systems offer IT the way to not only achieve these goals but to exceed them. They are nothing less than a
12 new category of computing designed to help organizations move forward. For Further Reference A Pattern Example: Business Intelligence To understand the functions of an expert integrated system, it s helpful to examine how it changes the current IT experience. Suppose an IT department is called on to deliver a mulit- tier Buisiness Intelligence (BI) platform. This means IT must install and configure multiple databases, an application server tier, report servers, and HTTP servers, wire it all together, and tune it for performance, After that, constant maintenance is required, taking into account all of the various interdependencies, including security patches and software upgrades. Perhaps different platforms must be accommodated, as well as distinct use cases and even UIs. Now, with a BI Pattern, all of this becomes automated via a single encapsulated entity that can be maintained, deployed and replicated much more rapidly. Besides moving development and deployment time from weeks to minutes, the pattern also provides for built- in elasticity so that when the demand for the BI solution spikes, the system can automatically adjust resources levels to meet the changing demand - and it can do this without the need for a redesign of the deployment or even the need for intervention from the IT department. The system simply detects the spike, adjusts resource levels based on business set policies, and when the demand shrinks, the system will adjust accordingly. With patterns of expertise you can now provide something like a complex BI solution much more rapidly and reduced ongoing operational expense, freeing IT to focus on providing new innovative solutions to the business.
13 Footnotes 1. IDC, Game Changing Virtual Technology: Major Shifts and Innovations that will Forever Change your IT Business, doc #DR2011_GS4_MB, March IDC White Paper, sponsored by IBM, Expert Integrated Systems: A Next- Generation Computing Platform, April IBM Market Insights 2011 Business Benchmarking Time- to- Value Study 4. A Q commissioned study conducted by Forrester Consulting on behalf of IBM 5. IDC's Worldwide Converged Systems Taxonomy, 2012, IDC #235293, June Based on IBM testing and documented in IBM System x Virtualization Server Consolidation sizing methodology. IBM Flex System x240 supports 2.7X more Peak Utilization Virtual Machines (VMs) than previous generation BladeCenter HS22V (81 VMs vs. 29 VMs). PureFlex System provides better scaled performance than HS22V due to higher total memory capacity of Flex System x240 Compute Node. To run the same application load as one rack of PureFlex System, customers would require up to 3 racks of BladeCenter HS22V. Statement is made conservatively at 2x improvement in density.compared utilizing virtualized applications between PureFlex System and previous generation BladeCenter HS22V. Specific client environments and results may vary.