IT Assignment
Chapter 2: Why Does IT Behave the Way it Does? Bill Flemming
MAKING SENSE OF IT BUSINESS MANAGEMENT
With all the money spent on IT system management, with all the products purchased to improve IT performance, and with all the consulting dollars spent, why has the average IT organization progressed only to the point where IT management proudly boasts of proactive engagements in preventing system failures?
In the 2007 edition of this book, SAS Institute quoted Gartner statistics that surveyed the degree to which different industries developed in Gartner's IT Infrastructure and Operations Maturity Model. Gartner data from 2005 showed that only nine percent of IT operations had reached the "Service" level of maturity. Nine percent! Further, 51 percent wanted to be at the Service level by 2006. What were the corresponding numbers for results released in 2008? Eight percent had made it to Service and twelve percent hoped to be there by 2012.[1] Granted, the Gartner IT Maturity Model changed to a more sophisticated version, but with a sustained industry-wide lack of progress toward achieving IT infrastructure and operations "maturity," either the paradigm is wrong, or IT leadership and their C-Suite peers haven't yet put all of the pieces in place to achieve the goal. This chapter examines the key elements of IT maturity, what the industry provides in terms of products and thought leadership, remaining gaps, and how the best practice CIO can address these issues.
The best place to start is a neglected but essential leading indicator: higher education for IT professionals. This first became obvious after speaking about IT opportunities with a newly minted Computer Science graduate student. As we spoke about system and IT business management jobs, it became clear to me that this new grad from a high-priced school had no idea what I was talking about. This lack of integrated IT system and business management curricula isn't just restricted to young graduates. Some of the answers to our neglected business management education lie in how IT professionals are educated and trained.
Colleges and universities educate and train engineers within Computer Science curriculums. Engineering expertise is required to keep the lights on and the systems functioning, but the comprehensive, strategic management of information technology is not an engineering exercise, particularly the challenging business alignment portion. Strategic IT management is a business exercise that requires a more comprehensively educated breed of manager. When hiring IT professionals out of college, most businesses hire Computer Science majors, MBAs, or Computer Engineers. Those IT engineers eventually progress into management. The curriculum for an undergraduate Computer Science major at Carnegie Mellon appears in Exhibit 2.1.
Where are the courses in Capacity Management? Service Level Management? Costing? Business Alignment? Computer Science and Computer Engineering programs train technologists but not business managers; MBA programs produce business managers but not IT technologists. If higher education is not producing IT business managers, then the answers have to come from elsewhere. To get to the answers a historical perspective is necessary.
Exhibit 2.1: Carnegie Mellon Computer Science Curriculum
Source: www.cmu.edu/academics/interdisciplinary-programs.shtml .
When I first entered the industry in the mid-1980s, stable, efficient mainframes were the predominant business computer. They were housed in raised-floor, environmentally controlled, glass-walled rooms typically called the Glass House. While machines were large, the business computing world was small by today's standards. The programming languages—COBOL, FORTRAN, and PL/1—were procedural and well-suited for the business applications of the era. For the most part, business applications were devoted to payroll, receivables, payables, general ledger, transaction management, batch reporting, and database management. Business applications crunched numbers quickly and accurately. The business connection was obvious. Businesses without the applications did the number management by hand—a pencil-and-paper bound method that was slow, expensive, and prone to error.
While the Glass House was centrally controlled, monolithic, and inflexible, it was stable, reliable, and almost always available. End users at the time complained about the control and inflexibility of IT, but that inflexibility and control safeguarded system management. IBM's Multiple Virtual Storage (MVS) operating system and Systems Network Architecture (SNA) network environment defined and controlled access. User transactions were entered into 3270 terminals written for Customer Interface Control System (CICS). Users claimed they were hard to use, and training costs were high. Virtual Telecommunications Access Method (VTAM) controlled peripheral access. Everyone adhered to the standards set by this MVS/SNA world. The quickest route to vendor disgrace was to offer a product incompatible with Glass House standards, and when product incompatibility brought down a production system, the vendor seldom got a second chance. Stability and reliability were paramount values for mainframe performance.
Before long, MVS could not supply all the necessary system management tools.[2] While MVS focused on control, other system management vendors such as Computer Associates and Legent marketed products for job scheduling, tape backup, report production and distribution, network management, storage management, console management, and capacity management/chargeback systems. During this era, most businesses managed for high transaction volume during business hours and tight batch windows in the evenings. System managers were concerned with hardware tuning and ensuring the availability of online environments. They were well paid and no one in the business understood them, but MVS was the operating system that ran the world. It was a perfect system management world, but not a perfect end-user world.
The hermetically sealed Glass House that no one on the outside understood was not spending time reacting to outages. MVA and SNA were built to provide an environment that set standards, and those standards ensured high availability. Network managers knew what the access traffic was going to be and where it was coming from. Unlike today, networks were not open to the outside world. Capacity Management concentrated on CPU utilization and storage needs in a controlled fashion. The Glass House was charged with providing stable environments for transactions, accounting systems, and reports. In a sense, they were completely aligned with higher business priorities. But all good things must come to an end. The personal computer and UNIX made their debuts, rapidly reordering IT system management priorities.
Desktop and Distributed Technology Explode
In the early 1990s, several technologies incubating since the 1970s had matured and gained rapid acceptance, but each of these innovative technologies depended on the other co-evolving technologies to achieve that acceptance. For example, the Internet depended on other enabling platforms, and the enabling technology that led the charge and changed all the rules was the personal computer (PC) when it debuted in the 1980s.
The PC broke rules and standards formulated for the Glass House on a number of levels.[3] IBM made business decisions about the PC that had the unintended consequence of changing the mainframe system management world forever. When IBM decided to enter the PC business they adopted a nonstandard IBM approach. To save time getting a product to market, IBM outsourced the product components, including the operating system DOS to Microsoft and the CPU to Intel. Uncharacteristically, IBM also decided to only license DOS and the CPU chips, allowing Microsoft and Intel to sell the technology to competitors in any manner they chose. Based on the business model and early marketing, IBM probably never intended or foresaw that the personal computer would become a personal Business computer. Early PC marketing showed a family gathered around a DOS green screen PC with a child at the keyboard as everyone beamed at the camera.[4] The PC was positioned as flexible and easy to use in the home, classroom, and office. The PC was slow to gain traction until Microsoft and Intel began to sell to other companies.
Without the restraining influence of the mainframe world and an attractive platform to exploit, the time frame of 1992 to 1995 unleashed the greatest era of technology innovation the computer industry had ever witnessed. Windows 3.1, Office, LANS Web browsers, the Internet "superhighway," ISPs, and client server computing all hit the market. When Microsoft introduced Windows 3.1 and Intel developed more powerful chips, PC sales gained rapid traction. That traction generated the demand for easy-to-use IT applications such as Windows. In short-order Silicon Valley, and the industry addressed that demand, albeit in a manner that created huge problems for IT system management.
In April of 1992, Microsoft released Windows 3.1. It was an instant success with three million copies sold in two months.[5] Windows 3.1 wasn't just a desktop release. It went far beyond personal productivity. Windows 3.11 for workgroups was an expansion of 3.1, which added many network capabilities for network connectivity, peer-to-peer support, and client/server applications. The Windows NT release soon dominated the LAN server market.
Shortly after Windows 3.1, Office 92 hit the market, which established the PC as a platform for professionals in most businesses. By extension, Microsoft had set new standards for Graphical User Interface use and business application interface standards. By comparison, those 3270 green screen applications began to look very tired and cumbersome.
As the PCs arrived on more and more business desktops, Bell Labs introduced another technology that took hold in the marketplace and became a key distributing component enabler,[6] a new operating environment that made it easier for developers to create software, particularly with non-procedural applications. UNIX was born, and it was open sourced—a new environment for small, powerful, cheaper machines that could be easily networked. No system management, baked in, none available.
A further technology with roots in the 1970s emerged as a key distributed computing enabler. Ethernet won international approval by the International Organization for Standards as the standard for LANS in 1989.[7] This acceptance made Ethernet the leader in LAN technology and spurred LAN growth in business offices connecting WINTEL machines to client/server applications and email.
The move away from mainframes toward the world of graphical desktops and client server applications hit critical mass. The ability to develop multitier applications, such as clients using Windows, database servers, and business processing across LANs on cheaper, faster UNIX servers exploded. Applications moved off the mainframe and sometimes out of the IT organization altogether. Distributed computing succeeded in breaking the Glass House monopoly of a carefully crafted environment that kept computing stable and available.
In 1993, Marc Andreessen introduced Mosaic, an Internet browser developed at the University of Illinois, and everyone with a computer had access to the World Wide Web.[8] Mr. Andreessen left to found Netscape where he found stiff competition from Microsoft and Internet Explorer. Along with browsers, search engines entered the marketplace in 1994 with WEBCRAWLER, Lycos, and a pair of 1995 births, YAHOO and Altavista. The twin technologies, browsers and search engines helped fuel the Internet explosion and later ecommerce. The growth in the Internet alone was more than enough for system managers to handle. Exhibit 2.2 shows the growth in Internet host sites from October 1991 to July 2006, where 439,286,364 sites existed.
The result for IT system management was a huge vacuum, and vendors scrambled to fill the void. Applications and the infrastructure that housed them were scattered everywhere outside the Glass House. Stability, availability, and response time became IT organization headaches, and firefighting was the new normal. Gaining visibility into the "health" of the distributed environment was not yet possible. Until 1991, Legent, a leading system management vendor only offered products for the mainframe, when the company acquired Spectrum for its XCOM 6.2 product to connect disparate systems through to mainframes to share data. From there, Legent and Computer Associates, another leading systems management vendor, spent millions and millions of dollars on R&D to bring the disciplines of the Glass House out on to the wide open spaces of the new Wild West of distributed computing. "Simply put, we're accelerating CA's move into client/server," said Charles Wang, Computer Associate's CEO of the acquisition.[9]
Exhibit 2.2: Growth of the Web
Source: Robert H. Zakon, "Hobbes' Internet Timeline v8.2," January 1, 2010, www.zakon.org/robert/internet/timeline/ .
The lack of stability in the distributed world was aggravated by another problem. Software vendors rushed products to market with both infrastructure incompatibility problems and a general lack of testing. Both situations lead to further system instability because immature products created outages that sometimes brought distributed production systems down. Without adequate tools to manage those kinds of problems, IT lapsed from the stability and predictability of the Glass House to the chaotic problems still evident in some enterprises today.
One of the first products that addressed the problem of distributed visibility to enter the system management marketplace and gain traction was brought by a UNIX hardware vendor, Hewlett Packard (HP). HP used agents to monitor the health and availability of a device. The simplest implementation was to send an agent alert to a console when the device was unavailable. As the product and monitoring industry grew, the agents read and collected a variety of metrics from system logs. Agent technology reporting alerts back to a central console become the standard for distributed operations management. The system management disciplines of backup, report management, job control, and capacity management grew more slowly because the IT organization had to first solve the acute pain of system instability—still a problem today.
Y2K and the Growth of the Enterprise Applications
As the 1990s ended, two new, self-feeding phenomena entered the fray. Programs written (either hardware or business applications) with only two digits representing the year field would potentially be rendered inoperative on January 1, 2000. Software would not be able to tell whether the year was 1900 or 2000, and some predicted catastrophic consequences if the programs and hardware weren't retrofitted or replaced with new Y2K compliant products.
Faced with enormous retrofitting costs with little upside in terms of additional functionality, many companies elected to buy new enterprise applications and new hardware. The new enterprise applications pushed IT organizations deeper into the management of business processes because the enterprise applications leveraged the new distributive technology rather than running on mainframes. Related business processes included e-commerce, enterprise resource planning, supply chain management, and customer relationship management with applications were very different from traditional accounting-oriented mainframe applications.
The physical implementation of the enterprise applications also brought a technology complexity that most IT organizations lacked the system management experience acumen to handle. Many of the "clients" were browser-based utilizing Intranet and Internet networks.[10] The applications crossed enterprise IT boundaries in the sense that business processes could be open to other businesses and consumers. Servers part of more complex architecture were placed into pools to provide specialized functions rather than simply working as a part of multi-tier application architecture. Data was far more complex, and there was more of it every day.
Applications of innovative technologies promoted new uses, more complexity, more demands for service, and wider distribution as employees worked both inside the enterprise and at home. Enterprises now primarily promote ongoing growth and innovation, finding more and more intensive uses for IT-enabled technology, along with the requisite demands for less cost and more business alignment. And system management? Best practice CIOs work to lead their IT organizations out of fire-fighting mode toward practices that increasingly enable enterprise strategies as educated, informed business partners. Enterprise IT organizations face competitive system management outsourcing pressures from managed service providers. Wherever system management responsibilities reside, the best practice CIO must remain current with new system management processes and tools entering the arena. The system management industry still lacks a coherent view of what is needed to address the demands placed on the CIO and IT organization. The following sections build just such a perspective for systems management.
The System Management Challenge
After the spine-tingling technology transformation of the 1990s and the enterprise-wide adoption of ever more complex applications, it is fair to say that the business demand for new IT enablement continues to intensify. It is also fair to say that system management as a profession still lags behind the technology transformation of the 1990s and business demands for application of the technology. It is interesting to note that the elements of technology transformation in the early 1990s took place approximately twenty years from their inception in the 1970s. While the PC would have undoubtedly been a successful product independently, other products such as UNIX, Ethernet, Internet, browsers, Client/Server, and even proprietary software applications would not have succeeded as spectacularly as they did without concurrent platform development.
From the perspective of a twenty-year cycle for system management starting from 1992, 2012 should mark the start of a system management golden era. The industry has many perspectives to consider and many products ready to mature, exploit, and integrate with other products to initiate a new era of system management. Viewed as an industry, IT system management is composed of the industry analyst community, major system management/hardware vendors, niche players, and thought leaders from IT itself.
The industry analyst component is the vanguard of twenty-first century system management. While many firms work in this market space, SAS settled on two thought leadership representatives that have very different approaches and only join ideas at the business alignment juncture. McKinsey concentrates on managing IT as a business and promoting IT practices as strategic business enablers. Gartner emphasizes IT maturity models and approaches IT business management considerations from system management toward business alignment. McKinsey is steering non-IT business managers toward IT business alignment and Gartner is steering IT engineers through the engineering maturity processes to achieve IT business alignment.
During the Glass House era, business applications were primarily devoted to accounting and online financial transactions. The business value was more obvious, and the technology was far narrower in application compared to the enterprise applications in play today. With the broader, enterprise-wide applications, the need for strategic intent is now even more acute. McKinsey therefore emphasizes the articulation of strategic themes. In the article, "Innovations in IT Management," McKinsey states that IT generates value on two complementary levels: (1) core assets of hardware, software, and processes, and (2) value in use.[11] "Value in use" refers to business applications that are optimized to yield maximum investment value.
Optimizing investment value here means that the enterprise establishes a value on the IT organization and its resources through a series of metrics that determine the economic value of the IT investment to the business. While metrics are not hard and fast, important measures determine the cost-to-revenue ratio, strategic value, and competitive edge. McKinsey also advocates measuring operational value by putting key performance indicators (KPIs) on the operational level of the business, such as "on-time delivery."
In a second article, "Managing IT in a Downturn: Beyond Cost Cutting," McKinsey emphasizes delivering increased value to the business as opposed to merely cutting cost.[12] As the IT organization and its resources increase in value and integrate more deeply into business processes, it becomes harder to cut costs and easier to increase value. After IT has engaged in efforts to streamline application portfolios, reduce infrastructure costs, and outsourcing, what remains is to create greater business impact through better management of sales and pricing, sourcing and production, support processes, and performance management (PM).
McKinsey continues their strategic focus via metrics in a third article, "Assessing Innovation Metrics."[13] While most enterprises value innovation, most don't measure it. Those enterprises that actually measure innovation generally depend on eight metrics to make their assessment. Innovation metrics provide strategic direction for innovation activities, guide allocation resources, and improve innovation performance. Interestingly, McKinsey noted that few enterprises tracked the relationship between innovation and shareholder value. Companies track revenue growth, customer satisfaction, and percentage of sales, but less than one-third track the relationship between innovation spending and shareholder value. Only the best companies pursue and measure innovations as a portfolio, and track the entire innovation process as inputs.
The Gartner Maturity Models approach to IT system management thought leadership begins from the polar opposite of McKinsey. Gartner maturity models aim directly at where system management has consistently felt the most pain from the early 1990s through to today: the inability to keep IT infrastructure up and running. Because of infrastructure redundancy and other safeguards, entire production systems seldom go dark. Rather than fight system-wide outages, the IT organization fights a continuous stream of small outages. Fighting small or large outages is expensive, time consuming, and a drain on services. Gartner reports that fifty-one percent of IT budgets are spent on system management and support of product applications.[14]
The Gartner IT Infrastructure and Operations (I&O) Maturity Model is a matrix with people, processes, technology, and business management on the left axis, with the maturity steps called Survival, Awareness, Committed, Proactive, Service Aligned, and Business Partnership across the top of the matrix. Survival, of course, is absence of any formal strategy or functions, an extremely chaotic, immature environment. Awareness indicates a level of insight that system management improvements are both possible and necessary. The IT organization is aware of needs, has some basic tools in place, but faces a big job for system management maturity. Committed requires an investment in both tools and processes. Real progress is made with the Proactive stage, where the IT organization implements processes, standards, domain management, and project management. The Service Aligned IT organization begins to implement customer and business services and management, which is essential to attaining a business partnership. IT organizations at the Business Partnership stage of maturity focus on business processes, business optimization, and business contribution metrics. Business Partnership practices overlap with the McKinsey perspective practices. The Business Partnership stage cannot be attained without an infrastructure foundation that is as well managed, stable, and available as the mainframes were in the Glass House.
How mature has the IT industry become? Exhibit 2.3 shows the results. I&O maturity levels will differ across industries, enterprise size, and business strategies, but the exhibit is an estimate of I&O maturity at each level with a prediction of progress by year-end 2012.[15] Gartner states that sustainable maturity develops over years, but CIOs and their IT organizations face rapid return on investment (ROI) expectations of four to six months. The answers to another Gartner survey that polled CIOs in 2006 for the reasons behind their lack of progress are still as revealing and relevant:
· Lack of senior management support
· Lack of practical implementation guidelines
· Lack of time to develop a thoughtful approach
· Lack of hierarchical reporting structure
· Lack of effective organizational communication[16]
System management vendors have always been on the fault line between industry advances that are a step or two ahead of enterprise adoption and the fight to keep infrastructure up and running. The system management tool market is large and lucrative because the need for these products and services is obvious, the pain is acute, and the large ROI is easy to justify. The vendors aim their products and services at the heart of the most critical distributed computing model problems: lack of infrastructure stability and operational firefighting. With homogeneous environments that were widely distributed and extraordinarily complex, monitoring and managing within that model was a massive challenge. As the industry matured beyond firefighting and chaos, vendors expanded their respective legacy products and acquired technology to round out their offerings.[17] Their subsequent offerings followed the Gartner maturity models building from infrastructure management toward business alignment and management, as opposed to the McKinsey perspective that linked IT practices with enterprise strategic intent.
If the history of business computing since 1992 tells us anything, the lesson would be that the problem to be solved is bigger than any single vendor or any single sector in the industry. The large system management vendors bring strong solutions in certain areas while other portions of the solutions seem to be isolated (Portfolio Management), an afterthought (Capacity Management), or nonexistent (Financial Management). The next plateau for the system management vendors coincides with the next rung on the Gartner scale, Service Management, and is marketed to a segment of IT that is complex and labor intensive.
Vendors have their own particular slant on service management. Whether marketed as Service-Oriented Architecture (SOA) or Business Service Management (BSM), the heart of the offering is a service catalog of IT offerings. Simply put, offerings are assembled to create a business application. The underpinning of the application is a contract that specifies the parameters of the offering, usually in terms of availability/response time/throughput (A/R/T). Such contracts may have chargeback provisions to be paid to IT and penalties paid to the end user if IT fails to meet service levels. Vendors are expanding products to provide varying amounts of automated provisioning with a set of common goals: minimize service costs, minimize developing and building service/business processes, reuse services, and of course, keep IT customers happy.
Niche vendors tend to be smaller startup companies plugging holes in the system management market that the megasystem management vendors address poorly, if at all. Service management is the core offering of most niche vendors, along with the customer's ability to transparently calculate the cost of service. Most of these offerings also have a dashboard of the service and financial results, and many of the niche vendors align with Information Technology Infrastructure Library (ITIL) consultancies.
A review of the products offered in this space reveals overlap with the larger vendor service management products. Service catalogs and service-level management are standard. Niche vendors differentiate their services most markedly in terms of the costing/chargeback engines. Cost of service and transparency are usually key components of financial management, but not for SAS. Nor is reporting results in a dashboard PM. If costing is limited to service while ignoring capacity management and other important IT costs, the niche is a bit too narrow. The tool is rendered to the tactical, rather than strategic, level.
This discussion now turns to the very individuals responsible for IT system management and who perform the work day in and day out. Giving advice is much easier than implementing advice. SAS gathers IT management perspectives from conferences, publications, and one-on-one interactions. Everyone in IT management seems to agree on one idea: The Business is First. Always. Most IT managers want "actionable" metrics. Others avidly look for ROI Management. The front runners are taking a strong, disciplined approach to cost cutting and financial management.
In 2007, Intel presented "The Road to Enterprise SOA" at America's SAP Users Group (ASUG), which discusses how the enterprise followed a maturity formula to attain SOA.[18] The project was driven by a SAP upgrade and a need for master data management. This situation is emblematic of current system management: grappling with the complexity of the enterprise applications and gaining a solid system management foundation. Intel's desired state was service portfolio planning, which would lead to cost savings in service re-use. The steps on the path were infrastructure consolidation, virtualization, instrumentation, service taxonomy, and value dials—a great start, but a very engineering-centric approach that would dovetail in the future with business services planning driven by service portfolios, and enable Intel to roll out applications faster and cheaper. The process began in 2001, and by 2007 was paying substantial dividends.
At the P100 conference in Orlando, Florida in January 2009, two major themes were cost-cutting and financial management. A panel of CIOs participated in a discussion devoted solely to cost-cutting. Their attitudes reflected a bit of McKinsey and a bit of Gartner. Panelists cited big savings of up to 33 percent in managed services and hoped to gain the same with virtualization. They stressed strengthening their infrastructure. Cutting costs was only part of their management emphasis. Demonstrating business benefits and using metrics were equally important, and participants expressed a desire to do a better job. Participants generally wanted to look and plan further into the future, some as far as four years. Industry analyst perspectives on the benefits of both maturity and business alignment were well represented.
Madge Meyers of State Street Bank presented a vision that clearly stood out. She also re-articulated the idea that an optimized infrastructure reduces costs. (In the spirit of disclosure, State Street Bank is a SAS customer.) In the domain of financial discipline, State Street is a clear front-runner. Under the rubric of governance, State Street established a model of end-to-end business cases and ROI management. This model produces self-funding enhancements, cost transparency, and charge back by usage.
In the broader context of SAS client and sales engagements, calls, and conversations, the typical IT organization is still ratcheting their way up to a level of proactive management, where they manage daily threshold infrastructure exceptions that could affect availability or service and implement an effective long-term capacity management forecasting system. The inability to forecast capacity across the enterprise is crippling to many IT organizations. This kind of effort is most always managed by engineers, and the proactive/capacity projects are rarely visible outside of the originating group, much less tied to business strategy.
Four IT Business Management Domains
Distributed computing sent shock waves through IT and system management from which IT has yet to fully recover. Industry analysts, system management vendors, niche vendors, and IT management all have perspectives on partial solutions, but none of the stakeholders espouses a complete model or the means to integrate all the perspectives. The essential management pieces are obviously missing and must come from new management approaches. This section explores the four IT business domains of capacity optimization, service level management, financial management, and business alignment as they apply to the current IT industry environment. SAS established these four domains based on ITIL version 3, our own thought leadership, and industry analyst feedback to form a hybrid model with invigorated focus and emphasis. The enhancement of these four domains will help lead to the transformation of IT system and business management into a stable, available, and business-aligned model. This section discusses the four domains in terms of the current state and emerging opportunities to expand the management of each IT business domain, to fulfill its promise and enable the other domains to fulfill their strategic promise.
Current State: When discussing capacity management, people often misunderstand the term capacity. As an essential IT management domain, capacity in this context means maintaining enough infrastructure to meet business computing requirements where infrastructure means network bandwidth, connections, server capacity, data storage space, even power and cooling. Capacity management does not refer to human resources, office space, desks, or parking places. From the SAS perspective, capacity optimization is a desired state for capacity management in terms of demand management and availability management. IT organizations have long been hampered in this area by their inability to create an enterprise view of servers, networks, and end-to-end enterprise business applications performance and metrics. A direct result of the rapid acceptance of distributed computing in the early 1990s, a heterogeneous, extremely complex infrastructure is now spread out across the globe, and managed and monitored by heterogeneous systems tools. Large companies generate vast amounts of performance and utilization metrics data, which they are unable to consolidate into a single enterprise view. Without enterprise views, it is nearly impossible to forecast capacity and business needs or to synchronize capacity with service views. The end result is that capacity management is usually done poorly—machine-by-machine or location-by-location—or not done at all.
A broader problem emerges when capacity management is performed server-by-server, location-by-location, or through educated guesswork. Inadequate capacity management acts to prevent business alignment by improperly sizing applications, burdening service management, and destroying financial projections. Capacity managers either buy too much infrastructure or undersize capacity, only to make a panic buy later. Most often, capacity managers either overbuy pooled infrastructure or buy oversized infrastructure to host a single application. Capacity managers work under the premise that oversized infrastructure will deliver the necessary performance and stability while preventing service disruptions during production times. Buying infrastructure sized for cost efficiency can lead to disruptions if managers can't assimilate the views needed to manage for performance.
Opportunities: Beyond the inherent promise to provide adequate infrastructure to run the business, capacity management presents business management opportunities in terms of both cost management and alignment to the other three business domains. Armed with end-to-end views of servers, capacity managers can provide a quick and large ROI by eliminating excess server capacity through server consolidation. Server consolidation requires both a current utilization baseline and a time series forecast of utilization extending out a year.
Consolidation is a solid maturity approach utilized by forwardthinking businesses. In addition, many IT organizations look to virtualization as a cost reduction initiative, and are correct in looking to virtualization for cost gains. Best practice capacity managers properly size virtualization allocations to maximize the cost management opportunity (i.e., not too big nor too small) by studying the allocation utilization rates and forecasting their growth. Not involving capacity managers in virtualization projects is a missed opportunity. After determining the optimal capacity, managing to that level wrings out the costs of idle capacity and also eliminates panic infrastructure buys. Buying in panic mode never results in a smooth, cost-efficient implementation, and enterprises often come dangerously close to impacting service levels. A recurring theme in this section is the interconnectedness of the processes across these four management domains. Capacity management should not be performed in the isolation of an engineering silo. To achieve greater levels of IT maturity, capacity managers must cross-pollinate and be cross-pollinated by performance measurement information flowing into and out of the four other IT business management domains.
For example, capacity managers need financial information beyond budget allocations for capacity buys. Determining the cost of capacity not only promotes more effective capacity management but also provides foundation costs for service management and service management contracts. The cost of capacity includes the cost of utilized capacity, reserve capacity, and unused capacity. Attacking unused capacity without distressing service levels is a combined capacity, service, and financial management exercise. Capacity managers must also determine capacity in nontraditional ways. Predicting the exhaust rates of standardized services and the growth of business services create other ripe opportunities.
Current State: From thirty-five thousand feet above the landscape where fine details are obscured, the various approaches to service management appear more alike than different. The differences sometimes appear to be more of a marketing exercise. From that altitude, service managers appear to build catalogs of IT services (Premium Web Service, Bronze Network Access, etc.) and assemble those components into business service contracts that specify service parameters in terms of availability, response time, throughput, and service hours. System management and niche vendors sell this approach, which appears to also follow the engineering focus of Gartner IT Infrastructure and Operational Maturity Model. Only a small minority of IT organizations has achieved this level of maturity. Best practice service managers remain vigilant for several issues while implementing this large and very necessary IT business management domain.
Service managers ignorant of the true cost of service are driving in a thunderstorm without windshield wipers. It's hard to see, and the risk of collision is unacceptably high. For example, service managers run into cost mismanagement issues when they enter into service contracts with business users who cost more than specified by the contract. Some niche vendors already address a portion of this issue, but service managers generally fail to make the connection between IT operational budgets and the planning process for new and ongoing services, as well as the connection between capacity management and service level management.
Opportunities: The capacity management section discussed the value of forecasting for optimizing infrastructure over time. Forecasting provides equal value for service management. Service reporting and financial measures are inherently reactive. Forecasting service level performance, future capacity needs, and cost of service growth augment service management practices. In addition to forecasting, a management structure that contains value measures is another essential opportunity for more mature service management. Reporting cost and service results are inadequate without knowledge of the degree to which the IT organization succeeded in terms of key enterprise strategic objectives and the metrics that track them. Meeting business goals and objectives are the ultimate measure of IT value. Forecasting performance is engineering; applying intelligence to enabling and meeting business strategy is a new level of business maturity for IT management.
Current State: We don't need to discuss IT financial management from thirty-five thousand feet. So neglected is the subject that I doubt it would even be visible from that altitude. Most CIOs and their IT organizations need to get down on their hands and knees at weed level to see the primary problems. Most IT organizations manage their budgets in spreadsheets. Spreadsheets are easy for individuals to use, inexpensive, and most everyone already has one. But as widely distributed spreadsheets quickly lose their effectiveness, they become very expensive. Difficult to consolidate into department views and then into an enterprise view, widely distributed spreadsheets result in inaccuracies, and a large percentage of expense planning is lost. The wide use of spreadsheets for IT budgeting can be traced to adoption of corporate budget systems that are not tailored for IT organizations and resource management. In addition, IT also budgets in another management area where the spreadsheet inaccuracy is even higher: planning and managing the portfolio of new and existing business service projects. Large corporations have hundreds and hundreds of such projects trapped in spreadsheets that they ruefully call "the swamp." With the utmost difficulty and many complete failures, financial managers attempt to reconcile the hundreds of spreadsheets into an accurate, consolidated view, and then reconcile the consolidated portfolio view to the operational budgets.
Trapped within this swamp are the answers to essential IT management questions with enterprise-wide strategic implications: On whom are we spending, and what are we spending it on? Was the spending justified? Optimized? How are we prioritizing IT support, service, and spend? How do we plan future IT resources and services? How do we minimize unused IT resources? How does this information inform overall decision making for IT, business units, and the bottom line?
Opportunity: The opportunity is to create an IT financial management system for a service-oriented IT organization. IT financial managers may continue to use a spreadsheet for individual operational and portfolio project planning, but they put all the data in one foundation to preserve its integrity. Financial managers then combine financial data with capacity and service data, which provide financial intelligence to the other IT business management domains in a usable format for optimizing their own strategic management decisions.
Capacity managers need to know the cost of capacity, including unused capacity, and cost of the support processes necessary to manage the infrastructure. Service managers need to know the cost of service, unused service, and who consumed the service, including relevant support processes for both the standard service catalog components and each business service.
Current State: True business alignment, or in Gartner terms, business partnership, is still an illusion as borne out by Gartner statistics. They report that less than one percent of IT organizations achieve their Business Alignment stage of maturity—a very small number in any survey sample that includes best practices. Two factors account for the lack of business alignment in the IT industry. First, vendors do not supply IT organizations with the support they need for developing portfolio management and system management tools that span metrics collected by various enterprise and IT monitoring and management systems. IT vendors focus on building maturity from the infrastructure management up—a purely engineering focus. What tools exist for financial and business management are neither integrated nor applied by IT managers seeking to solve their business challenges. Second, despite IT analysts pointing the industry toward aligning visions and techniques, the most important alignment achievement must include participation by the business intelligence community for new IT strategies with supporting applications.
Opportunities: Because true examples of business alignment are rare and few people in the industry have actually seen a single example, the opportunity for business alignment is far greater than most CIOs realize. This IT business management domain opportunity means that as a key strategic enabler for most enterprises, CIOs would manage the IT organization like a business with a business. CIOs and their IT organizations will a create value axis for every IT product and service from the performance metrics in each of the four IT business management domains. Business objectives that IT products and services must enable will be traced back to overall business strategy through these metrics. Enabling these strategic business objectives will carry a negotiated price tag to build and support after implementation. That negotiated price tag must fit within the ROI calculation of each business objective. CIOs will determine support costs through business user volume and service estimates, which IT will translate into service and capacity levels. Once built and implemented, the applications and their related IT services will be measured, forecasted, and optimized from business objectives, service results, costs, and business strategy.
The IT transparency illustrated here requires a different set of tools and processes that have yet to be broadly discussed in the IT market. While the rest of the enterprise is either already using or receptive to strategic performance management (SPM), many IT organizations have yet to reach this stage of business maturity. The utilization metrics increases in value when associated with strategy, initiatives, goals, and objectives that are mapped to other IT management domains.
PUTTING THE PIECES TOGETHER
While IT has made strides toward monitoring and managing infrastructure on the machine level, management is often done with multiple tools in multiple locations. In order to enable the four IT management domains with the required infrastructure metrics, best practice CIOs work to consolidate the metrics trapped in isolated tools and design new IT infrastructure metrics data management tools to access, integrate, aggregate, analyze, and manage large quantities of IT resource performance data from hardware, operating system software, networks, Web servers, databases, and applications.[19]
Step One: IT Infrastructure Metrics Data Management
IT resource performance metrics are generated by the logging mechanisms inherent to IT resources or are created by the Enterprise Systems Management tools used in managing IT infrastructures. Everything needed to analyze IT resource performance data from multiple sources for capacity planning and forecasting, consumption metrics for financial management, service-level performance measures, seasonality analysis, and enterprise IT performance summaries should be included in the initiative.
As demonstrated in Exhibit 2.4, SAS collects resource data, normalizes across platforms, and publishes the data for use by multiple users across the IT management domains. Of course, data management must scale to enterprise demands. When faced with multiple-system management tools, locations, or perhaps no tools at all, inherent to the SAS IT Resource Management server is the ability to natively extract information from many industry standard operating systems and systems management tools. Native support is the predefined ability to translate source data into a SAS representation for storage in the performance database. Data is retrieved from its native data source and brought into the IT resource data warehouse, where IT metrics are represented exactly as they are extracted. In the detailed level of the performance database, it is likely that IT data metrics are available for analysis and reporting on a per-polling cycle or per-event basis, as would be necessary for daily proactive management of a stable system. As detailed level data is retained, it is reduced—statistically summarized into aggregation levels that require incrementally less storage, enabling longer term storage as would be required for capacity management time series analysis. Any part of this data warehouse definition can be modified, added to, or deleted from at any time.
The staging transformation invoked by SAS IT Resource Management adapters extracts the raw IT resource performance data, performs any calculations and conversions that are required by that adapter, and loads (stages) the resulting data into tables in the IT data mart. Staging jobs can be run interactively or scheduled to run in batch mode, depending on the enterprise requirements. An aggregation transformation specifies how data is to be transformed and stored so that it can provide analysis and report-ready IT resource performance data.
For example, an aggregation transformation provides specifications for filtering data, calculating statistics, performing rolling accumulations, and ranking and grouping (classifying) the data according to user specifications. For any given adapter, SAS IT Resource Management generates transformations that create information maps referencing the data needed to create and view reports.
Exhibit 2.4: Enterprise Systems Management Tools for Managing IT Infrastructures
Exhibit 2.5: Data Reporting Tool Accessibility
Simple summary aggregation tables and the information maps for those tables are the primary inputs for creating IT resource performance and capacity planning reports such as CPU utilization, threshold analysis, and peak period analysis. SAS offers a collection of easy-to-use query and reporting interfaces for different types of users and recipients (e.g., capacity planners, IT infrastructure analysts, IT operations managers, senior IT management for business alignment, financial managers, and service-level management). Whatever the IT management domain, data must be accessible to a wide variety of reporting tools and visualization techniques as represented in Exhibit 2.5.
In summary, IT data management must aggregate data into enterprise and domain management views, regardless of the source of the data. Data must be accessible to other tools as necessary as well as reporting and visualization tools. IT data management is the foundation that the capacity, service, financial, and business alignment management domains are built upon.
Step Two: Capacity Management for Risk, Expense, and Quality
Lack of senior management support, practical implementation guidelines, time to develop a thoughtful approach, hierarchical reporting structure, and effective organizational communication—of these five major barriers to IT maturity, the lack of time to develop a thoughtful approach and the lack of effective organizational communication are more symptoms than root causes. If CIOs and their IT organizations found the time to develop a coherent articulation of enterprise-relevant IT business thought expressed in terms of business measures and ways IT resources enabled those business measures, IT/business communications would mature quickly.
As a business within a business, mature CIOs expect enterprise business users to provide the IT organization with business plans based on specific technology request and commensurate measures that demonstrate the ways that the technology resources provide and enable their strategic objectives. Best practice CIOs expect those business plans to address the four IT business management domains to facilitate the development of appropriate metrics and ongoing communications about performance results and related IT service forecasts. In short, thoughtfully developed business plans that promote strategic communications must include business measurable goals and objectives for the alignment domain, volume estimates for the capacity domain, service requirements for the service management domain, and the price the technology user is willing to pay for capital investments and ongoing services expenses for the financial management domain. Adequate capacity enables technology business plans and service levels. Well-managed capacity promotes financial management. Capacity management has other benefits as well. Here is what Martha Hays and Margaret Churchill have to say:
Critical IT systems are managed by monitoring day-to-day activities to ensure they are performing within stated utilization and performance thresholds. Responding to system alarms causes the operations team to react to the problem and implement an immediate change. These changes may cause unforeseen consequences on other parts of the environment, resulting in a ripple or cascade effect across the enterprise.
A capacity management process will instill a discipline of preparing for the future and planning for appropriate changes well before the system alarms go off. This allows the IT organization to predict when thresholds will be reached and prescribe the right changes. By doing so, better decisions can be made about server consolidation, hardware procurement, and service level management.[20]
The discipline of capacity management reduces risk and expense in a far-reaching manner. SAS places capacity as the first IT management domain to emphasize that mature capacity management (1) addresses instability problems endemic since the advent of distributed computing; and (2) pulls IT much closer to the service management, financial management, and business alignment domains. Hayes and Churchill expand the implications for the CIO and IT organization, where the figures they reference appear in Exhibit 2.6:
Forecasts provide advance notice that an outage or other type of problem is likely to occur. Once the predictions have been made, what is the best way to prevent the problem from actually occurring? Modeling delivers this capability through what-if scenarios that identify the impact of planned or unplanned changes to the system in areas such as workload levels, workload patterns, server configurations, network infrastructure, and storage arrays.
Exhibit 2.6: Hayes and Churchill on Capacity Management
Source: Martha Hays and Margaret Churchill, "Paper 6151 Forecasting + Modeling: A Partnership to Predict and Prevent Capacity Bottlenecks," Computer Measurement Group Presentation, 2006, Volume 1, (Computer Measurement Group), www.daschmelzer.com/cmg2006/PDFs/033.pdf .
IT firefighting is expensive, time-consuming, and distracts the organization from spending time on innovative solutions that add value back to the business. However, IT fire prevention is seldom practiced. The combination of forecasting and modeling, as part of a capacity management process, provides the insight needed into future IT events, and effectively prevents the IT fires.
Indeed. One begins to wonder how much of the IT organization's budget has been spent firefighting in the last seventeen years. And more:
The combination of these two methods also extends capacity management from a silo or server-centric view, to a broader end-to-end view. This allows predictions to be made about utilization, response time, workload growth, workload changes and infrastructure modifications.
For example, today's systems may be running within acceptable thresholds, and a linear trend shows that there is enough capacity on the servers to sustain a 5 percent fixed workload growth over the next three months. However, this trend does not account for a spike in demand that is expected for the upcoming holiday season. By using a time-series forecast, we can predict that the servers will bottleneck at the beginning of the peak shopping season. A model is built from the forecasted data and is used to evaluate several configuration changes that could support the increased workload. This analysis allows us to plan for the changes needed to the system to support the seasonal demand. With ample time before the peak season starts, we have the time to properly procure and plan for the changes, avoiding emergency procedures and expensive procurement. We have maintained user service and utilization levels—and have prevented an IT fire.
How does IT move from firefighting to a thoughtfully determined, well managed capacity that provides for service levels and financial management?
Capacity management evolves once an organization has implemented a PM system that focuses on monitoring current systems and reacting to alarms that result from exceeding utilization thresholds or service levels.
Since there must be an actual historical basis for the predictions, the data captured from performance monitors is required for the capacity planning process. Even if systems monitoring and event management are taking place, the resulting data must be stored for analysis and reporting purposes.
A great deal of capacity management information can be uncovered from basic analysis of the system and event data. These include the following:
· Which systems are experiencing outages or exceeding utilization thresholds?
· Is there a pattern based on time, day, month, etc.?
· Are the events consistent with changes in workload?
· Is there correlation between multiple events?
· When an issue occurs, what is the impact on the other systems?
For further analysis, data can be captured from the system monitors and stored in a centralized performance data repository. Since information from heterogeneous systems may have been collected at different time intervals and in different time zones, an extraction, transformation, and load (ETL) is used to "homogenize" the data. Once done, reports can be created that provide detailed historical analysis. These reports are useful to show a correlation of past events and past system responses to those events.
Once the data is captured, forecasting and modeling can be done. ITIL, however, does not differentiate between the two, as depicted in Figure 1. In practice, the two methods are different even though they use some of the same terminology.
Capacity builds on IT performance metrics. After performing a basic analysis of the infrastructure and establishing baselines, the role of forecasting becomes increasingly important. The CIO and IT organization perform forecasting on the baseline to predict normal business growth and from business plans that pass along volume predictions to capacity and financial managers.
A common report generated from historical data is a linear trend. This is the quickest and simplest report and is supported by spreadsheet software. Trend lines are based on an average of three or more historical data points extended to some future point. Trend lines are appropriate when future system behavior is expected to be at the same rate as the historical data with no seasonality.
In cases of seasonality or expected workload changes, trends will provide erroneous results. This is depicted in Figure 2. The forecast would indicate that there is sufficient capacity to support the system. However, when a new Web site comes online, the Web server is unable to handle the traffic. As a result, the trend provided an incorrect result since it was not able to predict the impact of this workload change.
Business decisions based on this trend will result in a false confidence that the system will continue to operate properly. As a result, basic linear trending is not recommended for making predictions of complex systems. For systems that experience pattern changes due to seasonality or planned business promotions, robust forecasting like time series with seasonality, trend, and event correlation is recommended.
Another problem with the linear trend is that it can't be used for systems that are close to experiencing a bottleneck. These systems will start queuing and/or consuming additional system overhead. This will cause the utilizations to skew, which will not be identified through a linear trend. Notice how a prediction of seasonal behavior from the chart above would have resulted in underestimating the capacity needed.
Time Series Forecasting analyzes time series variables and forecasts future values by extrapolating trends and patterns in the past values of the series or by extrapolating the effect of other variables on the series. With the use of sophisticated statistical software, a forecasting model can be developed and customized to best predict your time series. In the IT environment, these time series are easily obtained from your performance measurement data which has been stored and summarized in a capacity database (CDB). By choosing the proper time intervals (day, week or month) and variables as input, your historical data can lead you to a justifiable forecast of capacity requirements. Forecasting is accomplished through the following:
· Storing and summarizing the data
· Analyzing the data to determine seasonality and other patterns
· Selecting an appropriate forecasting method
· Generating a forecast, which includes expected business projections and events.
In addition to data management, analysis and forecasting tools are essential to capacity managers. Forecasting the complex enterprise applications is a set of intricate tasks, not the least of which is picking the proper statistical to fit the forecast problem. One approach is to hire statistical talent. Another is to use statistical software that analyzes the data and the problem and selects the best statistical approaches for you. After integrating statistical management approaches, capacity managers expand their roles by sending capacity data to financial managers and receive costing information in return. The next step is service management, because service managers would be ill prepared without capacity service level forecasts.
Step Three: Standard Services as a Foundation to Business Services
Stepping into service management requires the leveraging of the previously built IT data management foundation. Since service contracts are the underpinning of the services that the IT organization provides to enterprise business users via their business plans, cost of service is crucial to contract negotiation. Determining how much each business is able to pay for the IT business service is one building block in determining the value IT provides to business users.
Building an array of standard services based on the technology silos that all the Computer Science majors manage is a natural starting point. Network engineers manage networks as a service. The IT organization independently manages Web servers as a standard service. Standard services are those IT components that enable the business plans submitted by the business users. Standard services can be turned into "branded" services by attaching differing levels of value to each service, such as premium, gold, and bronze service levels, with price and performance gradations for each level. In terms of IT management, each standard service requires an internal contract that specifies the level of service, the cost we wish to manage to, and a capacity plan. Monitoring the results is essential.
With a stable set of standard services, service managers match business plan requirements in terms of availability, response times, and throughput to the catalog of services and determine the price when the business plan passes through capacity to ensure that the volume estimates in the plan can be supported. Agreement on services and costs constitute the elements of a service contract. Planning each contract is an exercise in capacity, service, and financial management.
Financial Management: A New Era
Financial management specific to the IT organization has received less thought and support than either capacity or service management. Very few IT organizations have found effective solutions for financial management challenges. Most IT organizations find that even the budgeting applications are ill suited for IT management. Tools for planning the capital and expense of business user new project portfolios are nonexistent. The result: IT manages budgets and projects in a swamp of spreadsheets. There are several reasons for the dependency on spreadsheets, but primary among them is the General Ledger (GL). The GL has other important purposes in the custodial role of finance that center on external reporting and internal controls. It has its own highly controlled environment with many rules governed by external reporting and tax requirements. Since the GL is optimized for external and tax reporting purposes, it is a poor tool for internal analysis and reporting, especially for the service-focused business of IT management.
Assuming that the best practice CIO manages the IT organization as an internal company separate from those business units consuming IT services, the resulting perspective clearly emphasizes the four IT management domains as the basis of the IT Value Axis: If IT services are being purchased, what does the customer expect from IT? What would IT as a supplier provide? The customer would expect:
· Consistent service delivery as specified in contract
· Consistent quality
· Competitive pricing
As an external supplier, IT would provide:
· Adequate capacity to assure consistent service delivery as required by contract
· Robust processes to provide consistent service
· Invoices with prices and supporting documentation of services provided
In such a business relationship, the price that the customer is willing to pay is based on perceived business value and a benchmark of prices and services offered by other IT suppliers. Customers are willing to pay for what their business services consume but not for waste or excess capacity. Excesses are not allowed to inflate the price. If the price is too high, the customer finds other alternatives. It is critical to note that the cost to provide the service does not establish the price for a profit-oriented IT provider. Independently, market conditions set prices, as it does for the rest of the business. The profit motive drives the supplier's desire and willingness to provide the service. Profit equals revenue minus cost. For a high enough price, an IT service provider purchases whatever capacity it needs to make a profit. When prices are too low, IT service providers redeploy resources to a more profitable market and potentially stop offering the service. If profitability is low, IT service providers are highly motivated to improve the efficiency of their operations. If profitability is too low, the service provider leaves the business.
IT financial management must address the costing of services, processes, and business services while bridging that data into a project portfolio planning environment that anticipates project capital costs (essentially new capacity and development costs), projects the ongoing expense of the capitalized costs, and delivers a projected financial impact on existing capacity for capacity managers. Key ingredients of the model depicted in Exhibit 2.7 are:
· Standard financial and statistical forecast models used for planning
· Fixed and variable qualifications available in both cost and plan model
· Planned capital expenditures and commitments used to manage depreciation
· Interactive scenarios used for simulations
· IT planning process that can be integrated with enterprise planning as a separate loop
· Department ownership of the loop
· Comparisons of operational and financial results available for monitoring and management
As depicted in Exhibit 2.7, the planning process allows planners to retain the front end use of Excel but without the consolidation headaches associated with free-floating spreadsheets. Planning data is housed within the SAS foundation but displayed in Excel. Project managers plan individual projects by assembling the anticipated standard service components. Within the spreadsheet are the costs for the services, the existing capacity, the unit costs, entry for the new volumes, and the impact on capacity. Conditional highlighting displays when capacity is impacted. Projects are consolidated for project directors and capacity managers as shown in Exhibit 2.8.
Exhibit 2.7: IT Financial Management Process
Central to the financial management system is the conversion of operational budgeting into the costing of services, processes, and capacity. The tool used for this conversion is Activity-Based Management (ABM). An IT ABM analysis starts by identifying the resources deployed in the IT organizations. Typically, these are recorded in the GL for each IT department—usually the only relationship between ABM and the GL. The SAS ABM model begins with GL costs by department and by account as its first view, retaining the original GL view to provide a firm common starting point and for correlation with external financial reporting. However, it then also provides a parallel view of resources that better reflects operational requirements and realities.
Exhibit 2.8: Project Consolidation for Directors and Capacity Managers
Regrettably, the level of GL data detail is both greater and lesser than the optimum set required for ABM. Where there is too much detail, such as accounts required for tax reporting, these details can be aggregated. Where there is too little detail, such as for asset depreciation, a well-designed ABM model uses its capabilities to split these costs to a more appropriate and useful level to support decision making. The reorganization is accomplished in the transition between the GL view and the new operations view.
This operationally focused view reorganizes resources and their costs to align them with operational resources—it captures resources in terms of teams of people, types of equipment, and other more operationally natural categories. This operational view is designed for use and easy understanding by operating personnel. At the same time, all costs are easily traced and reconciled to traditional financial views, which aggregate into a robust and comprehensive resource view that is also verifiable. This provides IT managers with a view of what the resources that they deploy actually cost in terms that they easily understand and use for decision making purposes. This view is also the foundation for subsequent views of activities and services. The overall model cost flow and management views are shown in Exhibit 2.9.
Exhibit 2.9: IT Cost Flow Model
IT components are tracked separately because they constitute a significant cost and play an essential part of IT business services. For these components, the cost recorded in the GL is usually aggregated to a level not useful to operations management. For this reason, component information is obtained from asset registers or other sources to identify the current costs of components at a sufficiently granular level to support service level agreements and capacity forecasting models.
After establishing an operational view of resources, the next step is to assign these costs to the work that these people do. This model uses existing ITIL process definitions as activity definitions. Using the ITIL process definitions provides a sound, common basis for measuring predictable, repeatable IT processes. ITIL processes are a necessary component in the IT maturity process. Without them, it would be difficult to attain a proactive, service-oriented, value-driven IT organization. Resources are traced department by department to the ITIL processes representing the work performed by people from each department. At this level of detail within the departments, this work is considered in terms of activities. Later, using reporting tools, costs are easily reported at the organization-wide ITIL process level. After these costs have been assigned, the costs of ITIL processes become available. Since the ITIL costs are tied to the operations resource and financial resource views, people can observe more powerful relationship costs with breakdowns by department and by types of resources either by the operational view or the financial view of cost.
The next step in ABM deployment enables tracing costs to the products and services provided by the IT organization. Following the business plans, service level agreements contain the value definition for business applications as well as service level requirements. In addition to documented SLAs, there may be more general standard services offered as a cost savings or legacy services not yet formally documented with customers. Even without formal documentation, IT should ascertain cost and service performance for internal management purposes. ABM methodology requires a cause and effect tracing of costs. Consumption metrics often provide the best basis for this assignment. Standard services may also be traced to specific services, where they are used as components of a broader business service offering.
At this point, cost analysis of services becomes available for service cost trend and service unit cost trend. With the relationships to ITIL processes, components, and people resources already established, these services can be analyzed by ITIL process and/or the resources consumed. Conversely, resources can be analyzed in terms of the services that ultimately consume them.
In the final step of ABM deployment, services are assigned to customers based on usage consumption metrics. Cost metrics are now available by customer. As these are added to the relationships already calculated in the model, a rich analysis base becomes available to help users and providers understand the operations and relationships to operating results. Since the symphonies that IT plays are business applications covered by SLAs, the results can be used for both value reporting to customers and also for internal IT optimization. In summary:
· Activity-Based Modeling is used do define, manage, and change the cost model and scenarios.
· The cost model contains operational and project data to derive costs, notably time per position to produce project costs.
· Costs assignment rules are defined on operational drivers.
· The cost model is populated, integrating financial data from financial systems and operational data from operational systems.
· Web reporting and business reporting are also used.
A sample of an online analytical processing (OLAP) cube extracted from ABM and posted to a dashboard appears in Exhibit 2.10. The subject is the costs associated with a branded standard service—Premium Web Servers. The cost transparency shows three main cost categories: the business services (projects) consuming the service, overhead reserve costs, and excess capacity costs. The next level of cost categories breaks down the individual hardware components of the service and the ITIL processes consumed by the branded, standard service. The data is essential for financial management.
Exhibit 2.10: ABM-Extracted OLAP Cube Posted to Dashboard
Step Four: Strategic Performance Management
PM isn't necessarily the last step in achieving IT management maturity and business alignment. PM can and should be applied throughout the maturity process. Let's review the key issues. CIOs identified the lack of organizational communication and the lack of time for thoughtful planning as barriers to IT maturity. Educational institutions generally prepare either IT engineers or business managers but not IT business managers. The distributed computing technology adopted in the early 1990s that neglected to include system management as a part of the business model put IT in a twenty-year cycle of attempts to overcome that neglect. Enterprise applications are complex, expensive, deeply embedded into business processes, and therefore demand a new business/IT relationship that is still evolving. And lastly, no one, not the system management vendors, industry analysts, niche vendors, or IT management, has a complete answer to the best practices at this stage of IT maturity. We do know that CIOs and their IT organizations have not comprehensively focused on the IT business management domains or applied business analytics to those IT management domains. In other words, CIOs are still learning to use business tools with IT tools for a complete picture.
Exhibit 2.11: Strategic Integration
PM has three discrete measure areas to communicate throughout the IT organization and to enterprise business partners: (1) business strategy, goals, and objectives; (2) IT internal engineering and management strategy, goals, and objectives; and (3) the cross-section of the business and IT strategy, goals, and objectives (see Exhibit 2.11). Suffice to say that PM requires a foundation as do the other IT management domains. PM is not a matter of placing metrics in a spreadsheet. PM provides the deliberate linkage between the management centers of business enablement ROI, business objectives, IT service contracts, capacity management, forecasting, and IT service cost transparency. Gary Cokins writes that many application failures are tied to the lack of a strategic view and the lack of a forward view. Let's tie the pieces together with examples.
PM metrics drill down from broadly stated metrics to lower levels of detail following the path of organizational intent through to group performance. Metrics should contain, at a minimum, a target to manage toward the actual achievement and a performance calculation that expresses how close the actual achievement came to the target. In Exhibit 2.12, PM is managing the performance achievement for three essential IT domains that link IT to the business: business objective measurements, IT service level performance, and the cost of the service.
Exhibit 2.12: Executive Dashboard PM View
Exhibit 2.13: Marketing Automation Drill-Down
For an executive view, this dashboard is probably sufficient. For other levels of the business it is not. Tracking the business performance requires three levels displayed in four examples of detail. The second level measures business objectives that justified the investment in Marketing Automation. While there is more business detail in subsequent drill-downs, the dashboard icons link directly to the original business case objectives for the Marketing Automation investment decision (see Exhibit 2.13).
The drill-down on the cost performance provides more insight by delivering an OLAP cube summary of the cost of providing the service of a business process in the Marketing Automation. Financial transparency detailing the costs of providing the service is attained by presenting standard service consumption by infrastructure component and the IT system management processes that were expended in supporting the infrastructure components in the service (see Exhibit 2.14).
Measuring business objectives, service contracts, and the cost of service is a major step forward for IT maturity. It is, however, only part of the maturity picture. IT internal engineering and management must be brought into the PM picture. Let's start with the business of capacity management. Capacity has several management objectives that look for cost savings and service improvement. Capacity managers must manage to keep excess capacity as low as possible through consolidation and virtualization and forecast as accurately as possible (see Exhibit 2.15).
Exhibit 2.14: Standard Service Consumption
Exhibit 2.15: Capacity Management Objectives
Exhibit 2.16: Excess Capacity Visualization with the Financial OLAP Cube
Excess capacity is a cost that IT must absorb, and the IT organization must manage excess as close to zero as possible. In Exhibit 2.16, the financial OLAP cube depicts the excess capacity costs by device and the IT system management services engaged in managing the excess capacity. Excess capacity isn't simply idle components, but components that are utilized to some degree but underutilized beyond the overhead capacity built in for headroom.
Service management must manage the twin centers of standard service and business services to a level of cost and service performance determined as a part of the IT organization's business management and planning. Standard services are managed to internal IT service contracts and are measured as a part of good IT business management. If the standard service is not performing to specifications, then the business services won't perform to contract specifications either (see Exhibit 2.17). Excess capacity can also be managed in the business-facing side of IT: in the standard services that comprise the business services. IT "sells" the standard services. Excess capacity here is similar to excess inventory. In this OLAP example, this service has far too much excess capacity (see Exhibit 2.18).
Exhibit 2.17: Shared Services Performance Dashboard
Has the application of traditional PM supplied all the information and tools necessary to make intelligent decisions? Do we have all the tools to judge which of our management domains impacts success or failure the greatest? Have we the tools to test the impact of proposed changes to our initiatives and forecast the results? In other words, can we apply advanced analytics to PM and move PM planning and interpretation from educated guesswork to confidence interval based analytics?
We can. SAS calls it Intelligent Scorecarding.
CHANGING THE WAY IT BEHAVES
The CIO and IT organization spend significant time, effort, and money designing and implementing strong, proactive programs to build effective capacity, service, financial, and alignment practices. You convinced your business users to provide business objectives along with their requests for IT investments. From those business objectives IT has created service, cost, and capacity contracts and spawned a broader set of internal management objectives. Now IT and its business partners are measuring alignment metrics up and down the value axis. In spite of all the planning, the measurement results indicate you are a little short of expectations. The programs didn't produce the results everyone had bet their bonuses on.
What do you do now? IT has been steered in a new direction, and while results are adequate, improvements must be made. Do you stay the course, adjust the course, or is a new, improved course needed?
If it were possible … to know which assumptions and initiatives were correct and which were wrong … to know which measures had the greatest impact on success, which assumptions failed, and to what degree … to know how much better we might understand the ways that IT was doing the right things, measuring the right things, and selecting measures reflect the correct strategy in the proper portions … to know when changes needed to be made (as they inevitably do) … could we then predict the impact of those changes before the changes are made?
This chapter is not just about SPM. It is about SPM and Intelligent Scorecarding for IT, which gives IT an additional tool, honed specifically for IT, to help steer IT management and initiatives in the direction that provides the most value for the business. Poorly conceived and ill-applied SPM undermines the achievement of optimized results, or even worse, becomes the wrong tool addressing the wrong problem. Deliberate and judiciously applied SPM is absolutely necessary for effective and efficient strategy execution.
IT strategy that addresses system management issues while knitting an effective business relationship with IT customers has proven to be an elusive goal since the end of mainframe dominance in the early 1990s. To understand the ways that SPM and Intelligent Scorecarding applications address this strategic IT management goal, we have to rummage through several subject areas and cross-pollinate germane material throughout four main themes:
1. IT and SPM
2. Management domains in the PM context
3. Analytical performance management
4. IT analytical PM
Addressing the Challenges of IT and Strategic Performance Management
SPM for IT is more than just generating and reporting a mishmash of stovepipe metrics. Reporting historical performance while ignoring correlation or forecasting is not a better way forward. SPM is not magic. It is part art, part science. SPM has its own lessons learned from how businesses applied it to handle business problems. SPM and analytical performance management (APM) do not have much of a history in IT management applications, so we have to extrapolate and garner insight from its use in other areas of the business. The art of SPM eventually leads us to the realization that the IT organization's business users are trying to solve very similar business issues as IT managers. The users just have a little more problem-solving experience in this area.
In its simplest form, SPM without Intelligent Scorecarding provides the linkages to ROI, objectives, and IT enablement, which contains a simple view of the business/IT value axis. The value axis is broad enough and layered enough to not only contain the business/IT value axis but also reflect an internal IT management model that covers the four essential IT management domains of business alignment, capacity management, service management, and financial management discussed earlier in this chapter. The value axis is important because most application failures lack strategic perspective—such application failures look to correct current issues and fail to address future conditions. While the IT organization's business partners must justify the acquisition and purpose of strategic business applications, IT is not excused from understanding and contributing toward that strategic perspective, which is the crux of business/IT value axis alignment.
Accordingly, the CIO's contribution to the value axis starts when IT first begins to process new business application goals and objectives. The best practice CIO uses the following steps to translate the application into empirical values that can be implemented and measured:
· Business application goals and objectives
· Determining the price to be paid for IT enablement and ongoing service
· System projections of volume, data storage, and other criteria
· Service level requirements
· Impact on capacity: infrastructure, capital, and expense
· Measurement system
Sounds easy enough, unless your IT organization hasn't matured into a service provider or learned to manage IT like a business. In that case, the steps on the value axis won't be attainable in the short term. IT organizations all too often manage by engineering expertise that translates into technology stovepipes. Managing by engineering and technology stacks is an effective and efficient approach given that we educate and train engineers by discipline rather than educating and employing generalists. What is not provided, especially as engineers are promoted to management, is an enterprise management view of IT organization responsibilities—a management view of IT not taught in most management programs. Earlier sections articulated the perspectives researched by two key members of the analyst community, McKinsey and Associates from a business perspective, and Gartner taking an engineering approach. While these two perspectives start from different ends of the maturity process, both promote the conclusion that IT should be run like a business. The following discussion fills in the space between these two perspectives.
SPM provides more than just linkages between ROI, objectives, and IT enablement. The benefit is far, far greater, and the job demands more than mere linkages. Engineering disciplines in IT are typically awash in internal key performance indicators (KPIs) generated by monitoring and management tools that don't necessarily address the obstacle that stymies most IT organizations from maturating into business-oriented enterprise service providers: firefighting infrastructure failures. Neglected, or missing altogether, are the business-oriented KPIs from service management, financial management, and business/application strategy.
Although many CIOs and their IT organizations have very little applied experience with SPM, the business side of the house generally has similar communication and management holes. SPM synchronizes business and IT improvements to create value. Since misapplication or under-application of SPM can lead to undesired outcomes, Gary Cokins lists three questions that enterprise and IT leadership should use SPM for focusing value creation and synchrony:
1. What products or service lines should we offer or not?
2. What markets and types of customers should we serve or not?
3. How are we going to win and keep winning?
When business is good, choices are easier; when business is bad, choices are harder. Many executives have found themselves on new ground that makes decision making more challenging than in the past. Enterprises have discovered that many of the products they have in the marketplace are basically commodities that offer little difference in terms of quality or functionality. Consumers choosing between cell phone A or cell phone B or flat screen 1 or flat screen 2 have little to lose (or gain) between brand choices or where to make the purchase. One transaction seems as good as the other.
Compounding the commoditization of products and services, enterprise and IT executives find themselves managing organizations that are complex and constantly changing. Their strategies often fail due to the lack of communication. Interest consequently spiked in SPM because businesses faced eight major chronic problems that needed resolution:
1. Failure to execute strategy
2. Unfulfilled return on ROI promises from transactional systems
3. Escalation in accountability (consequences) for results
4. Need for quick trade-off decision analysis
5. Mistrust of the managerial accounting systems
6. Poor customer value management
7. Dysfunctional supply chain management
8. Broken budgeting process
How were these issues addressed through SPM? The most typical activity was the reporting of standard information at standard frequencies. A simple, widely used example is reporting sales by geographic region each quarter.[21] Executives see how many flat screens 1 and 2 were sold, when and where, on a quarterly and regional basis. Adding just a bit more sophistication, managers and executives can drill down into the data and see trends, patterns, or anomalies that need attention. They see that flat screen 1's numbers were trending up in the Chicago area in August, but most likely they wouldn't make a connection with the Cubs stretch run for the pennant, know why flat screen 1 was chosen over flat screen 2, or know which flat screen was the most profitable model of the two. Additional sophistication adds executive alerts for responsible managers when parameters approach established preset performance levels. The Cubs were eliminated (again), sales nose-dived, and the alert was sent.
Since the IT organization's maturity has been surveyed, measured, and critiqued, it is fair to ask about the maturity of SPM application in businesses. More Davenport statistics:[22]
· Integrated SPM across the entire organization: 37%
· SPM throughout the enterprise, but not integrated: 32%
· SPM in some areas: 24%
· Nothing: 7%
It seems that almost all enterprises use PM and reporting to some extent. Most reporting, however, is not built on the foundation of a balanced scorecard or strategy maps. In other words, enterprise strategy is generally not embedded in the reporting system. Without the underlying relationships of major enterprise strategy initiatives, deliberate to the extent that investments measurably impact enterprise objectives, the result is often a loose knit system of KPIs, for ill or good, displayed on an executive dashboard that does not demand the attention of the decision maker.
A comparison of the eight major chronic problems that drove enterprise executives toward SPM and the problems faced by CIOs and their IT organizations today reveals striking similarities. A deep and persistent topic for the CIO over the last fifteen years has been the failure of IT to execute strategy on two principal levels: system management and customer value management. A key part of IT's customer value management breakdown is centered on the "unfulfilled return on ROI promises from transactional systems." Even as unfulfilled ROI shows up as a major driving force on the business side of the ledger, enterprise executives point the finger at IT costs rather than service delivery and optimizing IT system value by those very same systems. The keys to costing too much? (1) Mistrust of the managerial accounting systems; (2) a broken budgeting process; and (3) dysfunctional supply chain management in the form of disconnected engineering silos.
Few CIOs question that their IT organizations face daunting cost and PM challenges. System management problems that affect business performance and drain cash continue to plague enterprise IT. As any firefighter will tell you, it is cheaper and far more convenient to prevent fires than to put them out and repair the damage. When any organization finds itself reflexively reacting to one unforeseen event after another, the employees find it very difficult to muster the energy and resources to move strategic imperatives forward. The toll exacted in the case of such an IT organization is a group of people, no matter how talented or well-led, who are unable to overcome the system management challenges wrought by the rapid technology changes of the 1990s and Y2K-spawned enterprise applications that have burrowed so deeply into enterprise business processes (and yes, those same applications that failed to produce a ROI).
As the CIO and the IT organization seek to solve system management and business alignment issues, APM is the approach that can provide the tools to form, test, and forecast strategy that aligns enterprise strategy with IT enablers. Earlier sections identified four key IT management domains, which are also the building blocks of IT management maturity and the basis for SPM Intelligent Scorecarding.
Best practice CIOs recognize the importance of business management opportunities and how those opportunities should be strategically managed and measured, both to enable business end users to gain ROI for their IT investments, but also to enable strategic internal IT resource management. Intelligent SPM doesn't necessarily use capacity management as a starting point, but capacity management is the starting point for IT management maturity. Intelligent Scorecarding addresses the challenge faced by best practice CIOs to balance the right capacity at the right cost, to provide strategically enabling IT services, and to measure the impact on enterprise business objectives. Best practice CIOs and their IT organizations develop capacity management from an engineering silo into partnership with other IT enterprise business management domains and cement system management maturity with strategic enterprise business priorities. For example, after determining the optimal capacity level, actively managing the capacity to that level wrings out the excess capacity costs of idle capacity and also eliminates panic infrastructure buys. Determining the cost of capacity not only effectively manages capacity but also provides foundation costs for service management and service management contracts. The cost of capacity includes the cost of utilized capacity, reserve capacity, and unused capacity. Capacity managers must also determine capacity beyond the traditional manner. Predicting the exhaust rates of standardized services and the growth of business services remains a ripe opportunity for the best practice CIO.
Coupled with cost, service management is the most visible and measurable portion of the IT organization. All IT resource users have to be able to articulate what the IT organization should provide, how much is it going to cost, and why it is needed. Thus far in the marketplace, service reporting and financial measures have been reactive in nature, as has been most IT-related enterprise PM reporting. Forecasting service-level performance, future capacity needs, and cost of service growth moves service management from looking in the rearview mirror to using a GPS. Best practice CIOs may not avoid all the problems, but they can manage them. Forecasting provides value for service management. Best practice CIOs use one other opportunity for service management: the IT-related strategy management structure must contain enterprise business value measures. Reporting cost and service results are inadequate without knowing progress toward enterprise business strategy objectives. Meeting business goals and objectives are the ultimate measure of the ways that IT enables enterprise strategic value. Applying intelligence to enabling and meeting business strategy is a new breed of IT management.
When capacity management has additional tools that include the cost of capacity (including unused capacity), the cost of the support processes, and the strategy metrics to manage toward enterprise strategic objectives, capacity emerges from the cocoon of an engineering silo as an IT strategic enabler. When service management considerations include the cost of service, unused service, who consumes the service, relevant support processes for both the standard service catalog components and each business service, along with strategy metrics, the financial domain becomes the pinnacle of the new IT management model of business alignment.
Because many CIOs and IT organizations have little experience in applying either SPM or Intelligent Scorecarding to business alignment, best practice examples are rare. Most examples of business alignment center on portfolio management, which has been disconnected from true financial management and SPM as just another silo. The opportunity presented by business alignment is far greater than typically imagined. As one of the key strategic enablers of most enterprises, IT would be managed like a business while performing as a part of the greater enterprise. IT products in the form of applications that enable business processes enjoy the same benefits as other enterprise-wide management tools. Enterprise leadership creates a value axis from the performance metrics from each of the IT management domains. Each IT organization responsibility must have business objectives that trace back to overall enterprise business strategy, enabled by the CIO. This enablement must have a pre-established price tag before implementation of any new IT products, projects, or services. That price tag must fit within the ROI calculation of the enterprise business objectives, where the support costs are determined by the volume and service estimates articulated by the IT resource users, which the CIO and the IT organization translates into service and capacity levels. Once built and implemented, IT applications and related services can be measured, forecasted, and optimized from business objectives, service results, costs, and business strategy.
Intelligent Scorecarding and Analytical Performance Management
Thomas Davenport articulated four key points that he considered the Holy Grail of APM:
1. In an ideal world, consider or control for all possible variables that might have a substantial effect on financial performance (customer relationships, employee attitudes and behaviors, level of innovations, value of brand equity, and the others) for one overall equation that described the relative contributions.
2. No longer would organizations report metrics simply because they are familiar or because a standard balanced scorecard format suggests them.
3. Business strategies (those in a strategy map) would be testable.
4. Firms would also be able to predict the impact of increases or decreases in non-financial performance.[23]
Davenport suggests that enterprises statistically test relationships in business strategies through PM reporting: where that reporting would be more creative, relevant, and valuable than the sales stats for flat screen TVs and other conventional performance reporting. Businesses would know what worked and what failed, and modify their strategies accordingly. Davenport suggests applying statistical analyses to variables in PM, with demonstrates dramatic results. Carrying his work further, SAS shows how to perform the analysis.
Let's examine some Davenport examples. Each example tested two variables, one financial and one nonfinancial:
· Hilton: Five percent improvement in customer retention results in a 1.1 percent increase in annual revenues at a typical property.
· Harrah's: For each 1 percent growth in its share of customer gaming budgets, its share price increases by $1.10.
· Best Buy: Discovers that for every tenth of a point on a five-point scale increase in employee engagement at a particular store, operating income rises $100,000.
· Victoria's Secret: Finds that raising its average conversion rate by 1 percent brings more than $35 million in sales and $15 million in operating profit.
In a couple of more sophisticated examples where two variable analyses were performed with controls to focus the results:
· Toronto Dominion Bank: Controlled for customer service-financial performance of its branches, the bank finds that customer service equals 19 percent of the variation in branch profitability and further finds that service improvement only affects profitability in the middle of service rankings; incentives are aimed at the middle.
· Store24: Creates a balanced scorecard and a strategy map for a program titled "ban boredom" on the assumption that entertained customers buy more; the program does not work; it lowers profitability even when controlled for demographics and income levels, but the program works where employee skill levels are high.
As discussed in Chapter 1, enterprises regard the relationship between customer satisfaction and loyalty, employee satisfaction, and product capability as key indicators in driving financial success. Most businesses measure them to some degree. Other nonfinancial variables are less common. For example, a tech services company found that the average time it takes to close a case is a strong predictor of gross margins. An oil refiner found uptime closely correlated to profits.
Intelligent Scorecarding takes the best of APM and embeds itself in the strategy map of a scorecard. Most businesses have avoided the rigor of a balanced scorecard or the complexity of the strategy map that evolves from a balanced scorecard. Balanced scorecards impose a structure that might be too confining for IT. The first edition of CIO Best Practices discussed an IT balanced scorecard. Seeking a way to guide the implementation of IT maturity, SAS subsequently labored on that scorecard, complete with a strategy map, for another year or more. While strategy maps have value, SAS has found that the approach does have its flaws. A strategy map is a map, nothing more. If one is planning a car trip and uses an atlas to map the route, travelers can rest assured that the map is accurate but cannot be assured that the chosen route is the best route. These maps are not designed to identify traffic patterns, bottlenecks, construction zones, or factor in weather forecasts. Neither are strategy maps. No history to look back on. No forecasting to rely on. No indication which part of the plan is more important than any other.
Intelligent Scorecarding removes the tediousness of designing a scorecard, balanced or otherwise. CIOs and their C-Suite peers can design visually, drawing relationships as the path forward is more clearly articulated. As executives design strategic IT measurements, the tool applies statistical techniques to measure results and predict the impact of changes. By building the statistical measures and forecasts of the relationships into the strategy map, Intelligent Scorecarding overcomes a major shortcoming of strategy mapping, which gives equal weight to all relationships, whether financial or nonfinancial. In addition, Intelligent Scorecarding removes uncertainty associated with making changes in strategy by forecasting the impact of proposed changes with confidence intervals.
Designing Intelligent Scorecards can be done either by working from the top down or by building from the bottom up. Working from the top down is the easier and surer method. In honing this tool for IT, SAS works from the top, beginning with the four IT management domains of capacity, service, financial, and business alignment. Each domain is a "perspective" in the strategy map. As in any strategy map, thought must be given to the relationships between the management domains. The trick is to settle on the order of influence between domains.
|
|
|
NOTES
1. Donna Scott, Jay E. Pultz, Ed Holub, Thomas J. Bittman, and Paul McGuckin, "Introducing the Gartner IT Infrastructure and Operations Maturity Model," Gartner ID Number: G00147962, October 2007, confluence.arizona.edu/…/introducing_the_gartner_it_iInfrastructure+and_Operations_1479621.pdf.
2. Paul McCann and David Migliore, "What is MVS?" November 3, 2005, http://searchdatacenter.techtarget.com/sDefinition/0,,sid80_gci212618,00.html.
3. Tom Hormby, "What a Legacy: The Origin of the IBM PC," August 11, 2006, http://lowendmac.com/orchard/06/ibm-pc-5150-origin.html.
4. IBM, "IBM Personal Computer," Brochure 1982.
5. Google search for "Microsoft Windows 3.1 Sales History," www.google.com/search?hl=en&tbo=p&tbs=tl%3A1&q=microsoft+windows+3.1+sales+history&aq=f&aql=&aqi=&oq=.
6. Michael Hauben and Rhonda Hauben, "On the Early History and Impact of Unix Tools to Build the Tools for a New Millennium," Netizens: On the History and Impact of Usenet and the Internet (Wiley-IEEE Computer Society Press, 1997), www.columbia.edu/~rh120/ch001j.c11.
7. Brevard User's Group, "History of the Ethernet," LAN Networking Networks Packet Xerox, http://bugclub.org/beginners/history/EthernetHistory.html.
8. Marcus Kazmierczak, "History of the Internet," September, 24, 1997, http://mkaz.com/ebeab/history/.
9. "Legent Corporation—Company History," www.fundinguniverse.com/company-histories/Legent-Corporation-Company-History.html.
10. Steven Chan, blogs.oracle.com/images/Architecture%20Diagram%20R12.png.
11. Michael Bloch and Andres Hoyos-Gomez, "How CIOs Should Think about Business Value," McKinsey Quarterly, March 2009, www.mckinseyquarterly.com/How_CIOs_should_think_about_business_value_2307.
12. James M. Kaplan, Roger P. Roberts and Johnson Sikes, "Managing IT in a Downturn: Beyond Cost Cutting," McKinsey Quarterly, September 2008, www.mckinseyquarterly.com/Managing_IT_in_a_downturn_Beyond_cost_cutting_2196.
13. Vanessa Chan, Chris Musso and Venkatesh Shankar, "Assessing Innovation Metrics: McKinsey Global Survey Results," McKinsey Quarterly, November 2008, www.mckinseyquarterly.com/McKinsey_Global_Survey_Results_Assessing_innovation_metrics_2243?pagenum=5.
14. Michael Smith and Kurt Potter, "IT Spending and Staffing Report, 2009," Gartner ID Number: G00164940, January 27, 2009, www.gartner.com/DisplayDocument?doc_cd=164940.
15. See note 1.
16. Gartner, Incorporated, April 4, 2006/ID G00138514.
17. Jean-Pierre Garbani and Peter O'Neill, "The Megavendors in IT Management Software," Forrester Research Incorporated, May 21, 2008, www.forrester.com/rb/Research/megavendors_in_it_management_software/q/id/43904/t/2.
18. Gregg Wyant, Russ Heinsen, "Intel and the Road to Enterprise SOA," 2007 ASUG Annual Conference, Session 1601, presented by Intel Corporation.
19. Joseph Hatcher, "Overview of SAS ITRM Resource Management 3.1.1," Introduction to SAS® IT Resource Management 3.1.1, (Cary, NC: SAS Institute Inc., 2007), support.sas.com/documentation/onlinedoc/itsv/intro311.pdf.
20. Martha Hays and Margaret Churchill, "Paper 6151 Forecasting + Modeling: A Partnership to Predict and Prevent Capacity Bottlenecks," Computer Measurement Group Presentation, 2006, Volume 1, (Computer Measurement Group), http://direct.bl.uk/bld/PlaceOrder.do?UIN=203791356&ETOC=RN&from=searchengine. All subsequent Hays/Churchill quotations and exhibits come from this presentation.
21. Thomas H. Davenport, "The Rise of Analytical Performance Management," SAS Institute White Paper, www.sas.com/resources/whitepaper/wp_5596.pdf.
22. Ibid.
23. Ibid.
34 | Bill Flemming