Business and IT ( Improving Data Governance )

25

CHAPTER 2

Information Systems, IT Architecture, Data Governance, and Cloud Computing

L E A R N I N G O B J E C T I V E S

2.1 Name the six components of an information system and match the various types of information systems to the type of support needed by business operations and decision-makers.

2.2 Describe an IT infrastructure, an IT architecture, and an enterprisewide architecture (EA) and compare and contrast their roles in guiding IT growth and sustaining long-term performance.

2.3 Explain the business benefits of information management and understand the importance of data governance and master data management in providing trusted data that is available when and where needed to support sustainability.

2.4 Understand the concepts of data centers and cloud computing and understand how they add value in an organization.

2.5 Describe the different types of cloud services and the various forms of virtualization and understand how they add value in an organization.

C H A P T E R O U T L I N E

Case 2.1 Opening Case: Detoxing Location-Based Advertising Data at MEDIATA

2.1 IS Concepts and Classifications

2.2 IT Infrastructure, IT Architecture, and Enterprise Architecture

2.3 Information Management and Data Governance

2.4 Data Centers and Cloud Computing

2.5 Cloud Services and Virtualization

Case 2.2 Business Case: Data Chaos Creates Risk

Case 2.3 Video Case: Cloud Computing at Coca-Cola Is Changing Everything

26 C H A P T E R 2 Information Systems, IT Architecture, Data Governance, and Cloud Computing

Introduction One of the most popular business strategies for achieving success is the development of a competitive advantage. Competitive advantage exists when a company has superior resources and capabilities than its competitors that allow it to achieve either a lower cost structure or a differentiated product. For long-term business success, companies strive to develop sustainable competitive advantages, or competitive advantages that cannot be easily copied by the competition (Porter, 1998). To stay ahead, corporate leaders must constantly seek new ways to grow their business in the face of rapid technology changes, increasingly empowered consumers and employees, and ongoing changes in government regulation. Effective ways to thrive over the long term are to launch new business models and strategies or devise new ways to outperform competitors. Because these new business models, strategies, and per- formance capabilities will frequently be the result of advances in technology, the company’s ability to leverage technological innovation over time will depend on its approach to enter- prise IT architecture, information management, and data governance. The enterprisewide IT architecture, or simply the enterprise architecture (EA), guides the evolution, expansion, and integration of information systems (ISs), digital technology, and business processes. This guid- ance enables companies to more effectively leverage their IT capability to achieve maximum competitive advantage and growth over the long term. Information management guides the acquisition, custodianship, and distribution of corporate data and involves the management of data systems, technology, processes, and corporate strategy. Data governance, or informa- tion governance, controls enterprise data through formal policies and procedures. One goal of data governance is to provide employees and business partners with high-quality data they can trust and access on demand.

Bad decisions can result from the analysis of inaccurate data, which is widely referred to as dirty data, and lead to increased costs, decreased revenue, and legal, reputational, and performance-related consequences. For example, if data is collected and analyzed based on inaccurate information because advertising was conducted in the wrong location for the wrong audience, marketing campaigns can become highly skewed and ineffective. Com- panies must then begin costly repairs to their datasets to correct the problems caused by dirty data. This creates a drop in customer satisfaction and a misuse of resources in a firm. One example of an organization taking strides to clean the dirty data collected through inac- curate marketing is the data management platform, MEDIATA, which runs bidding systems and ad location services for firms looking to run ads on websites (see Table  2.1). Let’s see how they did this.

Dirty data are data of such poor quality that they cannot be trusted or relied upon for decisions.

Introduction 27

Case 2.1 Opening Case

DIRTY DATA AHEAD

C ou

rt es

y of

B ill

y R

ay

Detoxing Location-Based Advertising Data at MEDIATA

Company Overview MEDIATA uses its audience and media delivery platform to deliver thousands of successful online advertising campaigns across Australia, Hong Kong, and New Zealand. Known as a “programmatic solution specialist,” the MEDIATA platform is truly cutting-edge. It runs bidding systems and ad location services for companies that are looking to run ads on websites and provides its clients with high-impact, fully man- aged, 100% transparent advertising campaigns that produce results. MEDIATA is committed to shaking up the online advertising industry and is evolving into a fast-growing international business. MEDIATA clients include Qantas, LG, Virgin Money, Konica Minolta, Optus, Carls- berg, Honda, ACCOR Hotels, Air New Zealand, Heinz, Woolworths, Citi- bank, and JP Morgan.

The Problem MEDIATA uses IP address data to locate customers and ad effectiveness. Unfortunately, as much as 80% of ad inventories come with an incor- rect location and MEDIATA realized that this “dirty data” was adversely affecting their business. Location-based advertising provides organi- zations and companies alike with massive benefits. Target customers can be reached easily and effectively through marketing campaigns tailored specifically for them. For example, utility companies and internet service providers usually have certain areas or regions that they service. Using location-based targeting (see Figure  2.1), these companies can target television, newspaper, and online display ads to attract new customers. Another benefit includes the reduced waste of running marketing campaigns in unprofitable areas. Firms can choose precisely where their advertisements are displayed without wasting resources on customer segments that will not respond because of location or preference discrepancies.

Advanced data analytics in location-based advertising also allows companies like MEDIATA to reach customers where and when they are in decision-making mode using programmatic bidding algorithms and ad inventories. Browser-based ads use these algorithms to predict which customer segments will click on certain ads at certain times of the day. Automated bidding then ensues, with the ad spot on the page going to the highest bidder (Cailean,  2016). However, the data must be accurate to be useful and MEDIATA realized that their data could be much better than it was. Given the importance of this technology to advertisers and digital advertising agencies, there are overwhelming issues to overcome.

The issues stem from outdated methods of locating Internet users through IP addresses. These old systems do not pinpoint where exactly traffic is coming from, rather they give advertising agencies broad geo- graphic regions to work with, and the ads go to random coordinates within the regions. Since the value of these activities comes from having accurate targeting, the inaccuracies of the antiquated systems severely impact profitability. As targeting regions shrink, information becomes more valuable and accurate, but even small inaccuracies dilute the value of demographic information applied to an audience.

The Solution In 2016, MEDIATA established a data governance program in which it partnered with Skyhook, a U.S. global location software company to

TA B L E 2 . 1 Opening Case Overview

Company MEDIATA was launched as Valued Interactive Media (VIM) in 2009. Rebranded in 2013 as MEDIATA

Industry Communications; Advertising

Product Lines Wide range of programmatic solutions and products to provide practical solutions for digital marketing campaigns to deliver successful online advertising campaigns to organizations across Australia, Hong Kong, and New Zealand

Digital Technology Information management and data governance to increase trust and accessibility of data to facilitate a company’s vision

Business Vision Shake up the online advertising industry. Improve transparency and foster greater cooperation between partners

Website www.mediataplatform.com

28 C H A P T E R 2 Information Systems, IT Architecture, Data Governance, and Cloud Computing

FIGURE 2.1 Location-based advertising.

2.1 IS Concepts and Classification Before we being to explore the value of information systems (ISs) to an organization, it’s use- ful to understand what an IS is, what it does, and what types of ISs are typically found at differ- ent levels of an organization.

In addition to supporting decision-making, coordination, and control in an organization, ISs also help managers and workers analyze problems, visualize complex sets of data, and cre- ate new products. ISs collect (input) and manipulate data (process), and generate and dis- tribute reports (output) and based on the data-specific IT services, such as processing customer orders and generating payroll, are delivered to the organization. Finally, the ISs save (store) the data for future use. In addition to the four functions of IPOS, an information needs feedback from its users and other stakeholders to help improve future systems as demonstrated in Figure 2.2.

The following example demonstrates how the components of the IPOS work together: To access a website, Amanda opens an Internet browser using the keyboard and enters a Web address into the browser (input). The system then uses that information to find the correct web- site (processing) and the content of the desired site is displayed in the Web browser (output). Next, Amanda bookmarks the desired website in the Web browser for future use (storage). The system then records the time that it took to produce the output to compare actual versus expected performance (feedback).

Information systems (ISs) is a combination of information technology and people’s activities using the technology to support business processes, operations, management, and decision- making at different levels of the organization.

IPOS is the cycle of inputting, processing, outputting, and storing information in an information system.

improve the effectiveness of MEDIATA’s user profile data by more pre- cisely locating IP addresses resolving MEDIATA’s challenges related to dirty data. Skyhook’s Context Accelerator Hyperlocal IP uses big data analytics to provide over 1 billion IP addresses to advertising platforms and cleaned MEDIATA’s dirty data to pinpoint customers within 100 meters, thus increasing ad effectiveness for its clients. Hyperlocal IP achieves this by using big data analytics to provide over 1 billion IP addresses to advertising platforms.

Now, every time a device like a cell phone or laptop requests a location, the on-device software scans for Wi-Fi, GPS, or cell tower data. Combining all of these data points allows Skyhook to provide extremely accurate coordinates and pass this information along to MEDIATA to use.

While this approach still is not perfect, it allows MEDIATA’s adver- tisements to become closer than ever to their target customers. A nine-month study conducted after implementing Skyhook showed that MEDIATA saw a 20% increase in marketing campaign effectiveness.

Creating and employing this data governance system allowed MEDIATA to clean its datasets and create new, effective methods to reach target audiences.

Questions 1. What business challenges did MEDIATA face because of its

dirty data?

2. What is the function of location-based advertising? 3. Why is it important to maintain accurate location data? 4. How did Skyhook and data governance enable MEDIATA to

achieve its vision?

5. What benefits did MEDIATA achieve as a result of implementing data governance?

Sources: Compiled from Cailean (2016), Schneider (2014), and Schneider (2015).

IS Concepts and Classification 29

Components of an IS A computerized IS consists of six interacting components. Regardless of type and where and by whom they are used within an organization, the components of an IS must be carefully man- aged to provide maximum benefit to an organization (see Figure 2.3).

PROCESSING Programs

Equipments Storage

FEEDBACK Error Report

Performance Metrics

Hard Drive Server USB

INPUT Data

Information Knowledge Instructions

STORAGE OUTPUT Reports Graphics

Calculations

FIGURE 2.2 IPOS cycle.

People

DATAD

Procedures

Network Softwarek So

Hardware

FIGURE 2.3 Components of an IS.

1. Hardware Any physical device used in a computerized IS. Examples include central pro- cessing unit (CPU), sound card, video card, network card, hard drive, display, keyboard, motherboard, processor, power supply, modem, mouse, and printer.

2. Software A set of machine-readable instructions (code) that makes up a computer application that directs a computer’s processor to perform specific operations. Computer software is nontangible, contrasted with system hardware, which is the physical compo- nent of an IS. Examples include Internet browser, operating system (OS), Microsoft Office, Skype, and so on.

3. People Any person involved in using an IS. Examples include programmers, operators help desk, and end-users.

4. Procedures Documentation containing directions on how to use the other components of an IS. Examples include operational manual and user manual.

5. Network A combination of lines, wires, and physical devices connected to each other to create a telecommunications network. In computer networks, networked computing

30 C H A P T E R 2 Information Systems, IT Architecture, Data Governance, and Cloud Computing

devices exchange data with each other using a data link. The connections between nodes are established using either cable media or wireless media. Networks can be internal or external. If they are available only internally within an organization, they are called “intranets.” If they are available externally, they are called “internets.” The best-known example of a computer network is the World Wide Web.

6. Data Raw or unorganized facts and figures (such as invoices, orders, payments, customer details, product numbers, product prices) that describe conditions, ideas, or objects.

Data, Information, Knowledge, and Wisdom As you can see in Figure 2.3, data is the central component of any information system. Without data, an IS would have no purpose and companies would unable to conduct business. Gener- ally speaking, ISs process data into meaningful information that produces corporate knowl- edge and ultimately creates wisdom that fuels corporate strategy.

Data are the raw material from which information is produced; the quality, reliability, and integrity of the data must be maintained for the information to be useful. Data are the raw facts and figures that are not organized in any way. Examples are the number of hours an employee worked in a certain week or the number of new Ford vehicles sold from the first quarter (Q1) of 2015 through the second quarter (Q2) of 2017 (Figure 2.4).

Information is an organization’s most important asset, second only to people. Information provides the “who,” “what,” “where,” and “when” of data in a given context. For example,

Data describe products, customers, events, activities, and transactions that are recorded, classified, and stored.

Information is data that have been processed, organized, or put into context so that they have meaning and value to the person receiving them.

Knowledge adds understanding, experience, accumulated learning, and expertise as they apply to a current problem or activity, to information.

Creatively assess knowledge to develop innovative policies and procedures to

reverse downward trend in sales

Use information to determine reasons for consistent downward trend in sales

from June 2016 to June 2017

17, 25, 54, 12, 68, 19, 39, 42, 72 Number of new vehicles sold

DATA (Raw figures)

INFORMATION (who, what, where, when)

KNOWLEDGE (how)

WISDOM (why)

Q 1

20 15

Q 2

20 15

Q 3

20 15

Q 4

20 15

Q 1

20 16

Q 2

20 16

Q 3

20 16

Q 4

20 16

Q 1

20 17

Q 2

20 17

New Vehicle Sales by Quarter

6

5

4

3

2

1

0

FIGURE 2.4 Examples of data, information, knowledge, and wisdom.

IS Concepts and Classification 31

summarizing the quarterly sales of new Ford vehicles from Q1 2015 through Q2 2017 provides information that shows sales have steadily decreased from Q2 2016.

Knowledge is used to answer the question “how.” In our example, it would involve deter- mining how the trend can be reversed, for example, customer satisfaction can be improved, new features can be added, and pricing can be adjusted.

Wisdom is more abstract than data and information (that can be harnessed) and knowledge (that can be shared). Wisdom adds value and increases effectiveness. It answers the “why” in a given situation. In the Ford example, wisdom would be corporate strategists evalu- ating the various reasons for the sales drop, creatively analyzing the situation as a whole, and developing innovative policies and procedures to reverse the recent downward trend in new vehicle sales.

ISs collect or input and process data to create and distribute reports or other outputs based on information gleaned from the raw data to support decision-making and business processes that, in turn, produce corporate knowledge that can be stored for future use. Figure 2.5 shows the input-processing-output-storage (IPOS) cycle.

Wisdom is a collection of values, ethics, moral codes, and prior experiences that form an evaluated understanding or common-sense judgment.

Storage Temporary memory (RAM), hard disks, flash memory, cloud

People Users, clients, customers, operators, technicians, governments, companies

Sending results,

collecting data,

feedback

Communication Working with information, changing,

calculating, manipulating

Processing Data collected,

captured, scanned,

snapped from transactions

Input Showing results on screen,

hardcopy, digital copy, archive

Output

FIGURE 2.5 Input-processing-output-storage model.

Types of ISS An IS may be as simple as a single computer and a printer used by one person, or as complex as several thousand computers of various types (tablets, desktops, laptops, mainframes) with hundreds of printers, scanners, and other devices connected through an elaborate network used by thousands of geographically dispersed employees. Functional ISs that support busi- ness analysts and other departmental employees range from simple to complex, depending on the type of employees supported. The following examples show the support that IT provides to major functional areas.

1. Marketing Utilizing IBM software, Bolsa de Comercio de Santiago, a large stock exchange in Chile, is able to process its ever-increasing, high-volume trading in microseconds. The Chilean stock exchange system can do the detective work of analyzing current and past transactions and market information, learning, and adapting to market trends and con- necting its traders to business information in real time. Immediate throughput in combina- tion with analytics allows traders to make more accurate decisions.

2. Sales According to the New England Journal of Medicine, one in five patients suffers from preventable readmissions, which cost taxpayers over $17 billion a year. In the past, hospitals have been penalized for high readmission rates with cuts to the payments they receive from the government (Zuckerman et al.,  2016). Using effective management information systems (MISs), the health-care industry can leverage unstructured informa- tion in ways not possible before, according to Matt McClelland, manager of information

32 C H A P T E R 2 Information Systems, IT Architecture, Data Governance, and Cloud Computing

governance for Blue Cross Blue Shield of North Carolina. “With proper support, informa- tion governance can bridge the gaps among the need to address regulation and litiga- tion risk, the need to generate increased sales and revenue, and the need to cut costs and become more efficient. When done right, information governance positively impacts every facet of the business,” McClelland said in the Information Governance Initiative (Jarousse, 2016).

Figure  2.6 illustrates the classification of the different types of ISs used in organiza- tions, the typical level of workers who use them and the types of input/output (I/O) pro- duced by each of the ISs. At the operational level of the organization, line workers use transaction processing systems (TPSs) to capture raw data and pass it along (output) to middle managers. The raw data is then input into office automation (OA) and MISs by middle managers to produce information for use by senior managers. Next, information is input into decision support systems (DSSs) for processing into explicit knowledge that will be used by senior managers to direct current corporate strategy. Finally, corporate executives input the explicit knowledge provided by the DSSs into executive information systems (EISs) and apply their experience, expertise, and skills to create wisdom that will lead to new cor- porate strategies.

Executives

Senior Managers

Middle Managers

Line Workers

Executive Information Systems (EIS)

Decision Support Systems (DSS)

Management Information Systems (MIS)

Transaction Processing Systems (TPS)

Wisdom

Knowledge

Information

Data

FIGURE 2.6 Hierarchy of ISs, input/output, and user levels.

Transaction Processing System (TPS) A TPS is designed to process specific types of data input from ongoing transactions. TPSs can be manual, as when data are typed into a form on a screen, or automated by using scanners or sensors to capture barcodes or other data (Figure 2.7). TPSs are usually operated directly by frontline workers and provide the key data required to support the management of operations.

Organizational data are processed by a TPS, for example, sales orders, reservations, stock control, and payments by payroll, accounting, financial, marketing, purchasing, inventory con- trol, and other functional departments. The data are usually obtained through the automated or semiautomated tracking of low-level activities and basic transactions. Transactions are either:

• internal transactions that originate within the organization or that occur within the orga- nization, for example, payroll, purchases, budget transfers, and payments (in accounting terms, they are referred to as accounts payable); or

• external transactions that originate from outside the organization, for example, from cus- tomers, suppliers, regulators, distributors, and financing institutions.

TPSs are essential systems. Transactions that are not captured can result in lost sales, dis- satisfied customers, unrecorded payments, and many other types of data errors with financial

IS Concepts and Classification 33

impacts. For example, if the accounting department issued a check to pay an invoice (bill) and it was cashed by the recipient, but information about that transaction was not captured, then two things happen. First, the amount of cash listed on the company’s financial state- ments is incorrect because no deduction was made for the amount of the check. Second, the accounts payable (A/P) system will continue to show the invoice as unpaid, so the accounting department might pay it a second time. Likewise, if services are provided, but the transactions are not recorded, the company will not bill for them and thus lose service revenue.

Batch versus Online Real-Time Processing Data captured by a TPS are pro- cessed and stored in a database; they then become available for use by other systems. Processing of transactions is done in one of two modes:

1. Batch processing A TPS in batch processing mode collects all transaction for a day, shift, or other time period, and then processes the data and updates the data stores. Pay- roll processing done weekly or bi-weekly is an example of batch mode.

2. Online transaction processing (OLTP) or real-time processing The TPS processes each transaction as it occurs, which is what is meant by the term real-time processing. In order for OLTP to occur, the input device or website must be directly linked via a network to the TPS. Airlines need to process flight reservations in real time to verify that seats are available.

Batch processing costs less than real-time processing. A disadvantage is that data are inaccu- rate because they are not updated immediately, in real time.

Processing Impacts Data Quality As data are collected or captured, they are vali- dated to detect and correct obvious errors and omissions. For example, when a customer sets up an account with a financial services firm or retailer, the TPS validates that the address, city, and postal code provided are consistent with one another and also that they match the credit card holder’s address, city, and postal code. If the form is not complete or errors are detected, the customer is required to make the corrections before the data are processed any further.

Data errors detected later may be time-consuming to correct or cause other problems. You can better understand the difficulty of detecting and correcting errors by considering identity theft. Victims of identity theft face enormous challenges and frustration trying to correct data about them.

Management Information System (MIS) An MIS is built on the data provided by TPS. MISs are management-level systems that are used by middle managers to help ensure the smooth running of an organization in the short to medium term. The highly structured information provided by these systems allows managers

© J

a n _N

e vi

lle /iS

to c kp

h o to

FIGURE 2.7 Scanners automate the input of data into a transaction processing system (TPS).

34 C H A P T E R 2 Information Systems, IT Architecture, Data Governance, and Cloud Computing

to evaluate an organization’s performance by comparing current with previous outputs. Func- tional areas or departments―accounting, finance, production/operations, marketing and sales, human resources, and engineering and design―are supported by ISs designed for their particular reporting needs. General-purpose reporting systems are referred to as management information systems (MISs). Their objective is to provide reports to managers for tracking operations, monitoring, and control.

Typically, a functional system provides reports about such topics as operational efficiency, effectiveness, and productivity by extracting information from databases and processing it according to the needs of the user. Types of reports include the following:

• Periodic These reports are created or run according to a pre-set schedule. Examples are daily, weekly, and quarterly. Reports are easily distributed via e-mail, blogs, internal web- sites (called intranets), or other electronic media. Periodic reports are also easily ignored if workers do not find them worth the time to review.

• Exception Exception reports are generated only when something is outside the norm, either higher or lower than expected. Sales in hardware stores prior to a hurricane may be much higher than the norm. Or sales of fresh produce may drop during a food contamina- tion crisis. Exception reports are more likely to be read because workers know that some unusual event or deviation has occurred.

• Ad hoc, or on demand Ad hoc reports are unplanned reports. They are generated to a mobile device or computer on demand as needed. They are generated on request to learn more about a situation, problem, or opportunity.

Reports typically include interactive data visualizations, such as column and pie charts, as shown in Figure 2.8.

© D

a m

ir K

a ra

n /iS

to c kp

h o to

FIGURE 2.8 Sample report produced by an MIS.

Decision Support System (DSS) A DSS is a knowledge-based system used by senior managers to facilitate the creation of knowl- edge and allow its integration into the organization. More specifically, a DSS is an interactive application that supports decision-making by manipulating and building upon the information from an MIS and/or a TPS to generate insights and new information.

Configurations of a DSS range from relatively simple applications that support a single user to complex enterprisewide systems. A DSS can support the analysis and solution of a specific problem, evaluate a strategic opportunity, or support ongoing operations. These sys- tems support unstructured and semistructured decisions, such as make-or-buy-or-outsource decisions, or what products to develop and introduce into existing markets.

Degree of Structure of Decisions Decisions range from structured to unstruc- tured. Structured decisions are those that have a well-defined method for solving and the

IS Concepts and Classification 35

data necessary to reach a sound decision. An example of a structured decision is determining whether an applicant qualifies for an auto loan, or whether to extend credit to a new customer― and the terms of those financing options. Structured decisions are relatively straightforward and made on a regular basis, and an IS can ensure that they are done consistently.

At the other end of the continuum are unstructured decisions that depend on human intelligence, knowledge, and/or experience―as well as data and models to solve. Examples include deciding which new products to develop or which new markets to enter. Semistruc- tured decisions fall in the middle of the continuum. DSSs are best suited to support these types of decisions, but they are also used to support unstructured ones. To provide such support, DSSs have certain characteristics to support the decision-maker and the overall decision- making process.

The main characteristic that distinguishes a DSS from an MIS is the inclusion of models. Decision-makers can manipulate models to conduct experiments and sensitivity analyses, for example, what-if and goal seeking. What-if analysis refers to changing assumptions or data in the model to observe the impacts of those changes on the outcome. For example, if sales forecasts are based on a 5% increase in customer demand, a what-if analysis would replace the 5% with higher and/or lower estimates to determine what would happen to sales if demand changed. With goal seeking, the decision-maker has a specific outcome in mind and needs to determine how that outcome could be achieved and whether it is feasible to achieve that desired outcome. A DSS can also estimate the risk of alternative strategies or actions.

California Pizza Kitchen (CPK) uses a DSS to support inventory decisions. CPK has over 200 locations in 32 U.S. states and 13 other countries, including 17 California Pizza Kitchen non- traditional, franchise concepts designed for airports, universities, and stadiums. Maintaining optimal inventory levels at all its restaurants was challenging and time-consuming. The original MIS was replaced by a DSS to make it easy for the chain’s managers to maintain updated records, generate reports as and when needed, and make corporate- and restaurant-level decisions. Many CPK restaurants reported a 5% increase in sales after the DSS was implemented.

Executive Information System (EIS) EISs are strategic-level information systems that help executives and senior managers analyze the environment in which the organization exists. They typically are used to identify long-term trends and to plan appropriate courses of action. The information in such systems is often weakly structured and comes from both internal and external sources. EISs are designed to be operated directly by executives without the need for intermediaries and easily tailored to the preferences of the individual using them. An EIS organizes and presents data and information from both external data sources and internal MIS or TPS in an easy-to-use dashboard format to support and extend the inherent capabilities of senior executives.

Initially, EISs were custom-made for an individual executive. However, a number of off-the-shelf EIS packages now exist and some enterprise-level systems offer a customizable EIS module.

The ways in which the different characteristics of the various types of ISs are classified is shown in Table 2.2.

Here’s an example of how these ISs are used together to add value in an organization. Day-to-day transaction data collected by the TPS are converted into prescheduled summa- rized reports by middle managers using an MIS. The findings in these reports are then analyzed by senior managers who use a DSS to support their semistructured or unstructured decision- making. DSSs contain models that consist of a set of formulas and functions, such as statistical, financial, optimization, and/or simulation models. Corporations, government agencies, the military, health care, medical research, major league sports, and nonprofits depend on their DSSs to answer what-if questions to help reduce waste in production operations, improve inventory management, support investment decisions, and predict demand and help sustain a competitive edge.

Customer data, sales, and other critical data produced by the DSS are then selected for further analysis, such as trend analysis or forecasting demand and are input into an EIS for

36 C H A P T E R 2 Information Systems, IT Architecture, Data Governance, and Cloud Computing

TA B L E 2 . 2 Characteristics of Types of Information Systems

Type Characteristics TPS Used by operations personnel

Produce information for other ISs Use internal and external data Efficiency oriented

MIS Used by lower and middle managers Based on internal information Support structured decisions Inflexible Lack analytical capabilities Focus on past and present data

DSS Used by senior managers Support semistructured or unstructured decisions Contain models or formulas that enable sensitivity analysis, what-if analysis, goal seeking, and risk analysis Use internal and external data plus data added by the decision-maker who may have insights relevant to the decision situation Predict the future

EIS Used by C-level managers Easy-to-use, customizable interface Support unstructured decisions Use internal and external data sources Focus on effectiveness of the organization Very flexible Focus on the future

use by top level management, who add their experience and expertise to make unstructured decisions that will affect the future of the business.

Figure  2.9 shows how the major types of ISs relate to one another and how data flow among them. In this example,

1. Data from online purchases are captured and processed by the TPS and then stored in the transactional database.

2. Data needed for reporting purposes are extracted from the database and used by the MIS to create periodic, ad hoc, or other types of reports.

3. Data are output to a DSS where they are analyzed using formulas, financial ratios, or models.

ISS Exist within Corporate Culture It is important to remember that ISs do not exist in isolation. They have a purpose and a social (organizational) context. A common purpose is to provide a solution to a business problem. The social context of the system consists of the values and beliefs that determine what is admis- sible and possible within the culture of the organization and among the people involved. For example, a company may believe that superb customer service and on-time delivery are critical success factors. This belief system influences IT investments, among other factors.

The business value of IT is determined by the people who use them, the business processes they support, and the culture of the organization. That is, IS value is determined by the

IT Infrastructure, IT Architecture, and Enterprise Architecture 37

relationships among ISs, people, and business processes―all of which are influenced strongly by organizational culture.

In an organization, there may be a culture of distrust between the technology and business employees. No enterprise IT architecture methodology or data governance can bridge this divide unless there is a genuine commitment to change. That commitment must come from the highest level of the organization―senior management. Methodologies cannot solve people problems; they can only provide a framework in which those problems can be solved.

Questions

1. Name the six components of an IS. 2. Describe the differences between data, information, knowledge, and wisdom. 3. Define TPS and give an example. 4. Explain why TPSs need to process incoming data before they are stored. 5. Define MIS and DSS and give an example of each. 6. What characteristic distinguishes a DSS from an MIS? 7. What level of personnel typically uses an EIS? 8. What factors determine IS value?

2.2 IT Infrastructure, IT Architecture, and Enterprise Architecture Every enterprise has a core set of ISs and business processes that execute the transactions that keep it in business. Transactions include processing orders, order fulfillment and delivery, pur- chasing inventory and supplies, hiring and paying employees, and paying bills. To most effec- tively utilize its IT assets, an organization must create an IT infrastructure, IT architecture, and an enterprise architecture (EA) as shown in Figure 2.10.

Data

Data Data

Data are extracted, transformed, & loaded (ETL)

Data from online purchases

of transactional data

Database

Reporting MIS

Models applied to data for analysis

DSS

Processes raw data

TPS

Analytical processing of data to discover trends and learn

insights

Data Warehouse

FIGURE 2.9 Flow of data from point of sale (POS) through processing, storage, reporting, decision support, and analysis. Also shows the relationships among different types of ISs.

38 C H A P T E R 2 Information Systems, IT Architecture, Data Governance, and Cloud Computing

IT infrastructure is an inventory of the physical IT devices that an organization owns and operates. The IT infrastructure describes an organization’s entire collection of hardware, soft- ware, networks, data centers, facilities, and other related equipment used to develop, test, operate, manage, and support IT services. It does NOT include the people or process compo- nents of an information system.

IT architecture guides the process of planning, acquiring, building, modifying, inter- facing, and deploying IT resources in a single department within an organization. The IT architecture should offer a way to systematically identify technologies that work together to satisfy the needs of the departments’ users. The IT architecture is a blueprint for how future technology acquisitions and deployment will take place. It consists of standards, investment decisions, and product selections for hardware, software, and communications. The IT architecture is developed first and foremost based on department direction and business requirements.

Enterprise architecture (EA) reviews all the information systems across all departments in an organization to develop a strategy to organize and integrate the organization’s IT infrastruc- tures to help it meet the current and future goals of the enterprise and maximize the value of technology to the organization. In this way, EA provides a holistic view of an organization with graphic and text descriptions of strategies, policies, information, ISs, and business processes and the relationships between them.

The EA adds value in an organization in that it can provide the basis for organizational change just as architectural plans guide a construction project. Since a poorly crafted enterprise architecture (EA) can also hinder day-to-day operations and efforts to execute business strategy, it is more important than ever before to carefully consider the EA within your organization when deciding on an approach to business, technology, and corporate strategy. Simply put, EA helps solve two critical challenges: where an organization is going, and how it will get there.

The success of EA is measured not only in financial terms, such as profitability and return on investment (ROI), but also in nonfinancial terms, for example, improved customer satisfac- tion, faster speed to market, and lower employee turnover as diagrammed in Figure 2.11 and demonstrated in IT at Work 2.1.

EA Helps to Maintain Sustainability As you read in Chapter  1, the volume, variety, and speed of data being collected or generated have increased dramatically over the past decade. As enterprise ISs become more complex,

IT Infrastructure

IT Architecture

HR ACCTG.

PRODUCTION

SALES

FINANCE

ORGANIZATIONAL STRATEGY TO MAXIMIZE IT VALUE

Enterprise Architecture

Policy

FIGURE 2.10 Comparing IT infrastructure, IT architecture, and enterprise architecture.

IT Infrastructure, IT Architecture, and Enterprise Architecture 39

SUCCESS (PROFITABILITY, ROI,

INCREASED CUSTOMER SATISFACTION, FASTER SPEED TO MARKET,

LOWER EMPLOYEE TURNOVER)

CREATING IT

LEVERAGING IT

MAINTAINING IT

EA

FIGURE 2.11 Enterprise architecture success.

IT at Work 2.1

A New Enterprise Architecture Improves Data Quality and EIS Use Executives at a large chemical corporation were supported by an IS specifically designed for their needs—called an executive information system (EIS). The EIS was designed to provide senior managers with internal and external data and key performance indicators (KPIs) that were relevant to their specific needs. Tech Note 2.1 describes KPIs. As with any system, the value of the EIS depends on the data quality.

Too Much Irrelevant Data Unfortunately, the EIS was a failure. Executives soon discovered that only half of the data available through the EIS related to their level of analysis and decision-making at the corporate level. A worse prob- lem was that the data they needed were not available when and how they wanted them. For example, executives needed to analyze current detailed sales revenue and cost data for every strategic busi- ness unit (SBU), product line, and operating business to compare performance. But, data were not in standardized format as needed, making analysis difficult or impossible. A large part of the problem was that SBUs reported sales revenues in different time frames (e.g., daily, weekly, monthly, or quarterly), and many of those reports were not available when needed. As a result, senior management could not get a trusted view of the company’s current overall perfor- mance and did not know which products were profitable.

There were two reasons for the failure of the EIS:

1. IT architecture was not designed for customized reporting The design of the IT architecture had been based on financial accounting rules. That is, the data were organized to make it easy to collect and consolidate the data needed to prepare financial statements and reports that had to be submitted to the SEC (Securities and Exchange Commission) and other regulatory agencies. These statements and reports have well-defined or standardized formats and only need to be pre- pared at specific times during the year, typically annually or quarterly. The organization of the data (for financial reporting)

did not have the flexibility needed for the customized ad hoc (unplanned) data needs of the executives. For example, it was nearly impossible to generate customized sales performance (nonfinancial) reports or do ad hoc analyses, such as com- paring inventory turnover rates by product for each region for each sales quarter. Because of lags in reports from various SBUs, executives could not trust the underlying data.

2. Complicated user interface Executives could not easily review the KPIs. Instead, they had to sort through screens packed with too much data—some of interest and some irrelevant. To com- pensate for poor interface design, several IT analysts themselves had to do the data and KPI analyses for the executives—delaying response time and driving up the cost of reporting.

Solution: New Enterprise Architecture with Standardized Data Formats The CIO worked with a task force to design and implement an entirely new EA. Data governance policies and procedures were imple- mented to standardize data formats companywide. Data govern- ance eliminated data inconsistencies to provide reliable KPI reports on inventory turns, cycle times, and profit margins of all SBUs.

The new architecture was business-driven instead of financial reporting-driven. It was easy to modify reports—eliminating the costly and time-consuming ad hoc analyses. Fewer IT resources are needed to maintain the system. Because the underlying data are now relatively reliable, EIS use by executives increased significantly.

IT at Work Questions 1. Why was an EIS designed and implemented at the large

chemical corporation? 2. What problems did the executives have with the EIS? 3. What were the two reasons for those EIS problems? 4. How did the CIO improve the EIS? 5. What are the benefits of the new IT enterprise architecture? 6. What are the benefits of data governance?

40 C H A P T E R 2 Information Systems, IT Architecture, Data Governance, and Cloud Computing

long-range IT planning is critical. Companies cannot simply add storage, new apps, or data ana- lytics on an as-needed basis and expect those additional IT assets to work with existing systems.

The relationship between complexity and planning for the future is easier to see in physical things such as buildings and transportation systems. For example, if you are constructing a simple holiday cabin in a remote area, there is no need to create a detailed plan for future expansion. On the other hand, if you are building a large commercial development in a highly populated area, you’re not likely to succeed without a detailed project plan. Relating this to the case of enterprise ISs, if you are building a simple, single-user, nondistributed system, you would not need to develop a well-thought-out growth plan. However, this approach would not be feasible to enable you to successfully manage big data, copious content from mobiles and social networks, and data in the cloud. Instead, you would need a well-designed set of plans, or blueprints, provided by an EA to align IT with business objectives by guiding and controlling hardware acquisition, software add-ons and upgrades, system changes, network upgrades, choice of cloud services, and other digital technology investments that you will need to make your business sustainable.

There are two specific strategic issues that the EA is designed to address:

1. IT systems’ complexity IT systems have become unmanageably complex and expensive to maintain.

2. Poor business alignment Organizations find it difficult to keep their increasingly expen- sive IT systems aligned with business needs.

Business and IT Benefits of EA Having the right EA in place is important for the fol- lowing reasons:

• EA cuts IT costs and increases productivity by giving decision-makers access to information, insights, and ideas where and when they need them.

• EA determines an organization’s competitiveness, flexibility, and IT economics for the next decade and beyond. That is, it provides a long-term view of a company’s processes, sys- tems, and technologies so that IT investments do not simply fulfill immediate needs.

• EA helps align IT capabilities with business strategy―to grow, innovate, and respond to market demands, supported by an IT practice that is 100% in accord with business objectives.

• EA can reduce the risk of buying or building systems and enterprise applications that are incompatible or unnecessarily expensive to maintain and integrate.

Tech Note 2.1

Key Performance Indicators (KPIs) KPIs are a set of quantifiable measures used to evaluate factors that are crucial to the success of an organization. KPIs present data in easy- to-comprehend and comparison-ready formats to gauge or compare performance in terms of meeting an organization’s operational and strategic goals. KPIs are used in four main areas: increasing revenue; reducing costs; improving process cycle-time; and improving cus- tomer satisfaction. Examples of key comparisons include actual versus budget, actual versus forecasted, and the ROI for this year versus prior years. KPIs help reduce the complex nature of organizational perfor- mance to a small number of understandable measures, including:

• Financial KPIs: accounts payable turnover; inventory turn- over; net profit margin; sum of difference between planned and actual project budgets

• Social media KPIs: social traffic and conversions (number of visitors who are converted to customers); likes; new followers per week; social visits and leads

• Sales and marketing KPIs: cost per lead; how much revenue a marketing campaign generates; number of customer complaints; cycle time from customer request to delivery, percentage of correspondence replied to on time

• Operational and supply chain KPIs: units per transaction; carrying cost of inventory; order status; back order rate

• Environmental and carbon-footprint KPIs: energy, water, or other resource use; spend by utility; weight of landfill waste.

IT Infrastructure, IT Architecture, and Enterprise Architecture 41

TA B L E 2 . 3 Components of an Enterprise Architecture

Business architecture

How the business works. Includes broad business strategies and plans for moving the organization from where it is now to where it wants to be. Processes the business uses to meet its goals.

Application architecture

Portfolio of organization’s applications. Includes descriptions of automated services that support business processes; descriptions of interactions and interdependencies between the organization’s ISs.

Information architecture

What the organization needs to know to perform its business processes and operations. Includes standard data models; data management policies and descriptions of patterns of information production and use in an organization.

Technology architecture

Hardware and software that supports the organization. Examples include desktop and server software; OSs; network connectivity components; printers, modems.

IT at Work 2.2

EA Must Be Dynamic and Evolving In order to keep IT aligned with the business, the EA must be a dynamic plan. As shown in the model in Figure  2.12, the EA evolves toward the target architecture, which represents the company’s future IT needs. According to this model, EA defines the following:

1. The organization’s mission, business functions, and future direction

2. Information and information flows needed to perform the mission

3. The current baseline architecture 4. The desired target architecture 5. The sequencing plan or strategy to progress from the baseline

to the target architecture.

Baseline Transition Target

Im p

le m

en ta

ti o

n S

ta tu

s

Baseline architecture

Sequencing plan

Target architecture

FIGURE 2.12 The importance of viewing EA as a dynamic and evolving plan The purpose of the EA is to maintain IT–business alignment. Changes in priorities and business are reflected in the target architecture to help keep IT aligned with them (Bloomberg, 2016).

Developing an Enterprise Architecture (EA) Developing an EA starts with the organization’s goals, for example, where does it want to be in three years? and identifies the strategic direction in which it is heading and the business driv- ers to which it is responding. The goal is to make sure that everyone understands and shares a single vision. As soon as managers have defined this single shared vision of the future, they then consider the impact this vision will have on the business, technical, information, and solu- tions architectures of the enterprise. This shared vision of the future will dictate changes in all these architectures, assign priorities to those changes, and keep those changes grounded in business value.

According to Microsoft, the EA should include the four different perspectives shown in Table 2.3.

It is important to recognize that the EA must be dynamic, not static. To sustain its effective- ness, it should be an ongoing process of aligning the creation, operation, and maintenance of IT across the organization with the ever-changing business objectives. As business needs change, so must the EA, as demonstrated in IT at Work 2.2.

42 C H A P T E R 2 Information Systems, IT Architecture, Data Governance, and Cloud Computing

Questions

1. What is the purpose of the IT infrastructure? 2. How is the IT infrastructure different from the IT architecture? 3. What is the purpose of an EA? 4. What are the business benefits of EA? 5. Explain why it is necessary to ensure that an EA maintains alignment between IT and business strategy? 6. Explain KPIs and give an example.

2.3 Information Management and Data Governance As shown in Figure  2.3, data is the heart of the business and the central component of an IS. Most business initiatives succeed or fail based on the quality of their data. Effective planning and decision-making depend on systems being able to make data available in usable formats on a timely basis. Almost everyone manages information. You manage your social and cloud accounts across multiple mobile devices and computers. You update or synchronize (“synch”) your calendars, appointments, contact lists, media files, documents, and reports. Your pro- ductivity depends on the compatibility of devices and applications and their ability to share data. Not being able to transfer and synch whenever you add a device or app is bothersome and wastes your time. For example, when you switch to the latest mobile device, you might need to reorganize content to make dealing with data and devices easier. To simplify add-ons, upgrades, sharing, and access, you might leverage cloud services such as iTunes, Instagram, Diigo, and Box.

This is just a glimpse at some of the information management situations that organiza- tions face today and shows why a continuous plan is needed to guide, control, and govern IT growth. As with building construction (Figure  2.13), blueprints and models help guide and govern future IT and digital technology investments.

Information management is the use of IT tools and methods to collect, process, consolidate, store, and secure data from sources that are often fragmented and inconsistent.

Career Insight 2.1

Essential Skills of an Enterprise Architect (EA) Enterprise architects need much more than technology skills. On a daily basis, an enterprise architect’s activities can change quickly and significantly. Ideally, enterprise architects should come from a highly technical background. Even though enterprise architects deal with many other factors besides technology, it is still impor- tant to keep technical skills current. The job performance and suc- cess of such an architect―or anyone responsible for large-scale IT projects―depend on a broad range of skills.

• Interpersonal or people skills The job requires interacting with people and getting their cooperation.

• Ability to influence and motivate A large part of the job is motivating users to comply with new processes and practices.

• Negotiating skills The project needs resources―time, money, and personnel―that must be negotiated to get things accomplished.

• Critical-thinking and problem-solving skills Architects face complex and unique problems. Being able to expedite solu- tions prevents bottlenecks.

• Business and industry expertise Knowing the business and industry improves the outcomes and the architect’s credibility.

• Process orientation Thinking in terms of process is essential for an enterprise architect. Building repeatable and reusable processes as artifacts from the work they do and how they work themselves.

The most common function an enterprise architect will perform is that of overseeing a large-scale program. Programs are a group of related projects and as such, managing EA implementa- tions requires someone who is able to handle multiple aspects of a project at one time. Project management is covered in Chapter 13.

Information Management and Data Governance 43

Information Management Harnesses Scattered Data Business information is generally scattered throughout an enterprise, stored in separate sys- tems dedicated to specific purposes, such as operations, supply chain management, or cus- tomer relationship management. Major organizations have over 100 data repositories (storage areas). In many companies, the integration of these disparate systems is limited―as is users’ ability to access all the information they need. As a result, despite all the information flowing through companies, executives, managers, and workers often struggle to find the information they need to make sound decisions or do their jobs. The overall goal of information manage- ment is to eliminate that struggle through the design and implementation of a sound data gov- ernance program and a well-planned EA.

Providing easy access to large volumes of information is just one of the challenges facing organizations. The days of simply managing structured data are over. Now, organizations must manage semistructured and unstructured content from social and mobile sources even though that data may be of questionable quality.

Information management is critical to data security and compliance with continually evolving regulatory requirements, such as the Sarbanes-Oxley Act, Basel III, the Computer Fraud and Abuse Act (CFAA), the USA PATRIOT Act, and the Health Insurance Portability and Accountability Act (HIPAA).

Issues of information access, management, and security must also deal with information degradation and disorder―where people do not understand what data mean or how the data can be useful.

Reasons for Information Deficiencies Organizational information and decision support technologies have developed over many dec- ades. During that time management teams’ priorities have changed along with their under- standing of the role of IT within the organization; technology has advanced in unforeseeable ways, and IT investments have been increased or decreased based on competing demands on the budget. Other common reasons why information deficiencies are still a problem include:

1. Data silos Information can be trapped in departmental data silos (also called information silos), such as marketing or production databases. Data silos are illustrated in Figure 2.14. Since silos are unable to share or exchange data, they cannot consistently be updated. When data are inconsistent across multiple enterprise applications, data quality cannot (and should not) be trusted without extensive verification. Data silos exist when there is no overall IT architecture to guide IT investments, data coordination, and communication. Data silos support a single function and, as a result, do not support an organization’s cross- functional needs.

Data silo are stand-alone data stores. Their data are not accessible by other ISs that need it or outside that department.

© M

a rt

in B

a rr

a u d

/A la

m y

FIGURE 2.13 Blueprints and models, like those used for building construction, are needed to guide and govern an enterprise’s IT assets.

44 C H A P T E R 2 Information Systems, IT Architecture, Data Governance, and Cloud Computing

For example, most health-care organizations are drowning in data, yet they cannot get reliable, actionable insights from these data. Physician notes, registration forms, dis- charge summaries, documents, and more are doubling every five years. Unlike structured machine-ready data, these are messy data that take too much time and effort for health- care providers to include in their business analysis. So, valuable messy data are routinely left out. Millions of insightful patient notes and records sit inaccessible or unavailable in separate clinical data silos because historically there has been no easy way to analyze the information they contain.

2. Lost or bypassed data Data can get lost in transit from one system to another. Or, data might never get captured because of inadequately tuned data collection systems, such as those that rely on sensors or scanners. Or, the data may not get captured in sufficient detail, as described in Tech Note 2.2.

3. Poorly designed interfaces Despite all the talk about user-friendly interfaces, some ISs are horrible to deal with. Poorly designed interfaces or formats that require extra time and effort to figure out increase the risk of errors from misunderstanding the data or ignoring them.

4. Nonstandardized data formats When users are presented with data in inconsistent or nonstandardized formats, errors increase. Attempts to compare or analyze data are more dif- ficult and take more time. For example, if the Northeast division reports weekly gross sales revenues per product line and the Southwest division reports monthly net sales per product, you cannot compare their performance without converting the data to a common format. Consider the extra effort needed to compare temperature-related sales, such as air condition- ers, when some temperatures are expressed in degrees Fahrenheit and others in Centigrade.

5. Difficult to hit moving targets The information that decision-makers want keeps changing―and changes faster than ISs can respond to because of the first four reasons in this list. Tracking tweets, YouTube hits, and other unstructured content requires expensive investments―which managers find risky in an economic downturn.

These are the data challenges managers have to face when there is little or no information management. Companies undergoing fast growth or merger activity or those with decentral- ized systems (each division or business unit manages its own IT) will end up with a patchwork of reporting processes. As you would expect, patchwork systems are more complicated to modify, too rigid to support an agile business, and more expensive to maintain.

Information Requirements: Understandable Relevant Timely Accurate Secure

Parts Replenish

Procuring

Design

Build

Ship

Sales

Fulfillment

Billing

Support

Customer data Product data Procurement data Contract data Data order Parts inventory data Engineering data Logistics data

Data Types

Operations silos

Sourcing silos

Customer-facing silos

FIGURE 2.14 Data (or information) silos are ISs that do not have the capability to exchange data with other systems, making timely coordination and communication across functions or departments difficult.

Information Management and Data Governance 45

Factors Driving the Shift from Silos to Sharing and Collaboration Senior executives and managers are aware of the problems associated with their data silos and information management problems, but they also know about the huge cost and disruption associated with converting to newer IT architectures. The “silo effect” occurs when different departments of an organization do not share data and/or communicate effectively enough to maintain productivity. Surprisingly, 75% of employers believe team work and collaboration are essential, but only 18% of employees receive communication evaluations during performance critiques (Marchese,  2016). In the new age of efficiency of service, many companies like Formaspace, an industrial manufacturing and service corporation, must work toward complete cloud integration of old silos to increase customer service and generate more revenue. Enabling applications to interact with one another in an automated fashion to gain better access to data increases meaningful productivity and decreases time and effort spent in manual collaboration efforts. In an illustration of how silo integration is essential for a modern corporation, IT technician at Formaspace, Loddie Alspach, claims that in 2015, the company managed to increase revenues by 20% using Amazon-based cloud technology (Shore,  2015). However, companies are struggling to integrate thousands of siloed global applications, while aligning them to business operations. To remain competitive, they must be able to analyze and adapt their business processes quickly, efficiently, and without disruption.

Greater investments in collaboration technologies have been reported by the research firm Forrester (Keitt, 2014). A recent study identified four main factors that have influenced the increased use of cloud technologies, as shown in Table 2.4 (Rai et al., 2015).

Tech Note 2.2

Need to Measure in Order to Manage A residential home construction company had two divisions: stand- ard homes and luxury homes. The company was not capturing material, labor, and other costs associated with each type of con- struction. Instead, these costs were pooled, making it impossible to allocate costs to each type of construction and then to calculate the profit margins of each division. They had no way of calculating profit margins on each type of home within the divisions. Without the ability to measure costs, they did not have any cost control.

After upgrading their ISs, they began to capture detailed data at the division level. They discovered a wide profit margin on stand- ard homes, which was hiding the losses occurring in the luxury home division. Without cost control data, the profitable standard homes division had been subsidizing the luxury home division for many years. Based on the cost control data, the company decided to focus more on standard homes and adjust their pricing on luxury homes. This new cost control strategy increased the company’s long-term performance.

TA B L E 2 . 4 Key Factors Leading to Increased Migration to the Cloud

Cost Savings

Efficient Use of Resources

Unlimited Scalability of Resources

Lower Maintenance

Business Benefits of Information Management Based on the examples you have read, the obvious benefits of information management are:

1. Improves decision quality Decision quality depends on accurate and complete data. 2. Improves the accuracy and reliability of management predictions It is essential for

managers to be able to predict sales, product demand, opportunities, and competitive

46 C H A P T E R 2 Information Systems, IT Architecture, Data Governance, and Cloud Computing

threats. Management predictions focus on “what is going to happen” as opposed to finan- cial reporting on “what has happened.”

3. Reduces the risk of noncompliance Government regulations and compliance require- ments have increased significantly in the past decade. Companies that fail to comply with laws on privacy, fraud, anti-money laundering, cybersecurity, occupational safety, and so on face harsh penalties.

4. Reduces the time and cost of locating and integrating relevant information.

Data Governance: Maintaining Data Quality and Cost Control The success of every data-driven strategy or marketing effort depends on data governance. Data governance policies must address structured, semistructured, and unstructured data (discussed in Section 2.3) to ensure that insights can be trusted.

Enterprisewide Data Governance With an effective data governance program, managers can determine where their data are coming from, who owns them, and who is responsible for what―in order to know they can trust the available data when needed. Data governance is an enterprisewide project because data cross boundaries and are used by people throughout the enterprise. New regulations and pressure to reduce costs have increased the importance of effective data governance. Governance eliminates the cost of maintaining and archiving bad, unneeded, or inaccurate data. These costs grow as the volume of data grows. Governance also reduces the legal risks associated with unmanaged or inconsistently managed information.

Three industries that depend on data governance to comply with regulations or reporting requirements are the following:

• Food industry In the food industry, data governance is required to comply with food safety regulations. Food manufacturers and retailers have sophisticated control systems in place so that if a contaminated food product, such as spinach or peanut butter, is detected, they are able to trace the problem back to a particular processing plant or even the farm at the start of the food chain.

• Financial services industry In the financial services sector, strict reporting requirements of the Dodd−Frank Wall Street Reform and Consumer Protection Act of 2010 are leading to greater use of data governance. The Dodd−Frank Act regulates Wall Street practices by enforcing transparency and accountability in an effort to prevent another significant finan- cial crisis like the one that occurred in 2008.

• Health-care industry Data are health care’s most valuable asset. Hospitals have moun- tains of electronic patient information. New health-care accountability and reporting obli- gations require data governance models for transparency to defend against fraud and to protect patients’ information.

Master Data and Master Data Management (MDM) Master data is the term used to describe business-critical information on customers, products and services, vendors, locations, employees, and other things needed for operations and business trans- actions. Master data are fundamentally different from the high volume, velocity, and vari- ety of big data and traditional data. For example, when a customer applies for automobile insurance, data provided on the application become the master data for that customer. In contrast, if the customer’s vehicle has a device that sends data about his or her driving

Data governance is the control of enterprise data through formal policies and procedures to help ensure data can be trusted and are accessible.

Information Management and Data Governance 47

behavior to the insurer, those machine-generated data are transactional or operational, but not master data.

Data are used in two ways―both depend on high-quality trustworthy data:

1. For running the business Transactional or operational use 2. For improving the business Analytic use

Master data are typically quite stable and typically stored in a number of different sys- tems spread across the enterprise. Master data management (MDM) links and synchronizes all critical data from those disparate systems into one file called a master file, to provide a common point of reference. MDM solutions can be complex and expensive. Given their com- plexity and cost, most MDM solutions are out of reach for small and medium companies. Ven- dors have addressed this challenge by offering cloud-managed MDM services. For example, in 2013, Dell Software launched its next-generation Dell Boomi MDM. Dell Boomi provides MDM, data management, and data quality services (DQS)―and they are 100% cloud-based with near real-time synchronization.

Data governance and MDM manage the availability, usability, integrity, and security of data used throughout the enterprise. Strong data governance and MDM are needed ensure data are of sufficient quality to meet business needs. The characteristics and consequences of weak or nonexistent data governance are listed in Table 2.5. Data governance and MDM are a powerful combination. As data sources and volumes continue to increase, so does the need to manage data as a strategic asset in order to extract its full value. Making business data consistent, trusted, and accessible across the enterprise is a criti- cal first step in customer-centric business models. With data governance, companies are able to extract maximum value from their data, specifically by making better use of opportunities that are buried within behavioral data.

TA B L E 2 . 5 Characteristics and Consequences of Weak or Nonexistent Data Governance and MDM

• Data duplication causes isolated data silos. • Inconsistency exists in the meaning and level of detail of data elements. • Users do not trust the data and waste time verifying the data rather than analyzing them for appro-

priate decision-making. • Leads to inaccurate data analysis. • Bad decisions are made on perception rather than reality, which can negatively affect the company

and its customers. • Results in increased workloads and processing time.

Questions

1. What is information management? 2. What is the “silo effect” and how does it affect business performance? 3. What three factors are driving collaboration and information sharing? 4. What are the business benefits of information management? 5. Explain why it is important to develop an effective data governance program? 6. Explain the purposes of master data management. 7. Why has interest in data governance and MDM increased?

48 C H A P T E R 2 Information Systems, IT Architecture, Data Governance, and Cloud Computing

2.4 Data Centers and Cloud Computing Data centers and cloud computing are types of IT infrastructures or computing systems. Data center also refers to the building or facility that houses the servers and equipment. In the past, there were few IT infrastructure options. Companies owned their servers, storage, and network components to support their business applications and these computing resources were on their premises. Now, there are several choices for an IT infrastructure strategy―including cloud computing. As is common to IT investments, each infrastructure configuration has strengths, weaknesses, and cost considerations.

Data Centers Traditionally, data and database technologies were kept in data centers that were typically run by an in-house IT department (Figure 2.15) and consisted of on-premises hardware and equip- ment that store data within an organization’s local area network.

Today, companies may own and manage their own on-premises data centers or pay for the use of their vendors’ data centers, such as in cloud computing, virtualization, and software-as- a-service arrangements (Figure 2.16).

© O

le ks

iy M

ar k/

S hu

tte rs

to ck

FIGURE 2.15 A row of network servers in a data center.

© M

ic ha

el D

B ro

w n/

S hu

tte rs

to ck

FIGURE 2.16 Data centers are the infrastructure underlying cloud computing, virtualization, networking, security, delivery systems, and software-as-a-service.

Data Centers and Cloud Computing 49

In an on-premises data center connected to a local area network, it is easier to restrict access to applications and information to authorized, company-approved people and equip- ment. In the cloud, the management of updates, security, and ongoing maintenance are out- sourced to a third-party cloud provider where data is accessible to anyone with the proper credentials and Internet connection. This arrangement can make a company more vulnerable since it increases exposure of company data at many more entry and exit points. Here are some examples of data centers.

• National Climatic Data Center The National Climatic Data Center is an example of a public data center that stores and manages the world’s largest archive of weather data.

• U.S. National Security Agency The National Security Agency’s (NSA) data center, shown in Figure 2.17 is located in Bluffdale, UT. It is the largest spy data center for the NSA. People who think their correspondence and postings through sites like Google, Facebook, and Apple are safe from prying eyes should rethink that belief. You will read more about reports exposing government data collection programs in Chapter 5.

• Apple Apple has a 500,000-square-foot data center in Maiden, NC, that houses servers for various iCloud and iTunes services. The center plays a vital role in the company’s back-end IT infrastructure. In 2014 Apple expanded this center with a new, smaller 14,250-square- foot tactical data center that also includes office space, meeting areas, and breakrooms.

Since only the company owns the infrastructure, a data center is more suitable for organiza- tions that run many different types of applications and have complex workloads. A data center, like a factory, has limited capacity. Once it is built, the amount of storage and the workload the center can handle does not change without purchasing and installing more equipment.

© e

p a

eu ro

p ea

n p

re ss

p ho

to a

g en

cy b

.v ./A

la m

y

FIGURE 2.17 The NSA data center in Bluffdale, UT.

When a Data Center Goes Down, so Does Business Data center failures dis- rupt all operations regardless of who owns the data center. Here are two examples.

• Uber The startup company Uber experienced an hour-long outage in February 2014 that brought its car-hailing service to a halt across the country. The problem was caused by an outage at its vendor’s West Coast data center. Uber users flooded social media sites with complaints about problems kicking off Uber’s app to summon a driver-for-hire.

• WhatsApp WhatsApp also experienced a server outage in early 2014 that took the ser- vice offline for 2.5 hours. WhatsApp is a smartphone text-messaging service that had been bought by Facebook for $19 billion. “Sorry we currently experiencing server issues. We hope to be back up and recovered shortly,” WhatsApp said in a message on Twitter that was retweeted more than 25,000 times in just a few hours. The company has grown rapidly to 450 million active users within five years, nearly twice as many as Twitter. More than two-thirds of these global users use the app daily. WhatsApp’s server failure drove millions of users to a competitor. Line, a messaging app developed in Japan, added 2 million new registered users within 24 hours of WhatsApp’s outage―the biggest increase in Line’s user base within a 24-hour period.

50 C H A P T E R 2 Information Systems, IT Architecture, Data Governance, and Cloud Computing

These outages point to the risks of maintaining the complex and sophisticated technology needed to power digital services used by millions or hundreds of millions of people.

Integrating Data to Combat Data Chaos An enterprise’s data are stored in many different or remote locations―creating data chaos at times. And some data may be duplicated so that they are available in multiple locations that need a quick response. Therefore, the data needed for planning, decision-making, opera- tions, queries, and reporting are scattered or duplicated across numerous servers, data cent- ers, devices, and cloud services. Disparate data must be unified or integrated in order for the organization to function.

Data Virtualization As organizations have transitioned to a cloud-based infrastruc- ture, data centers have become virtualized. For example, Cisco offers data virtualization, which gives greater IT flexibility. The process of data virtualization involves abstracting, transforming, merging, and delivering data from disparate sources. The main goal of data virtualization is to provide a single point of access to the data. By aggregating data from a wide range of sources users can access applications without knowing their exact location. Using data virtualization methods, enterprises can respond to change more quickly and make better decisions in real time without physically moving their data, which significantly cuts costs. Cisco Data Virtualiza- tion makes it possible to:

• Have instant access to data at any time and in any format. • Respond faster to changing data analytics needs. • Cut complexity and costs.

Compared to traditional (nonvirtual) data integration and replication methods, data virtu- alization accelerates time to value with:

• Greater agility Speeds 5–10 times faster than traditional data integration methods • Streamlined approach 50–75% time savings over data replication and consolida-

tion methods • Better insight Instant access to data

Software-Defined Data Center Data virtualization has led to the latest development in data centers—the software-defined data center (SDDC). An SDDC facilitates the integration of the various infrastructures of the SDDC silos within organizations and optimizes the use of resources, balances workloads, and maximizes operational efficiency by dynamically dis- tributing workloads and provisioning networks. The goal of the SDDC is to decrease costs and increase agility, policy compliance, and security by deploying, operating, managing, and maintaining applications. In addition, by providing organizations with their own private cloud, SDDCs provide greater flexibility by allowing organizations to have on-demand access to their data instead of having to request permission from their cloud provider (see Figure 2.18).

The base resources for the SDDC are computation, storage, networking, and security. Typi- cally, the SDDC includes limited functionality of service portals, applications, OSs, VM hardware, hypervisors, physical hardware, software-defined networking, software-defined storage, a security layer, automation and management layers, catalogs, a gateway interface module, and third-party plug-ins (Figure 2.19).

It is estimated that the market share for SDDCs will grow from the current level of $22 billion to more than $77 billion in the next five years. As the use of SDDCs grows at this extraor- dinary rate, data center managers will be called upon to scale their data centers exponentially at a moment’s notice. Unfortunately, this is impossible to achieve using the traditional data center infrastructure. In the SDDC, software placement and optimization decisions are based on business logic, not technical provisioning directives. This requires changes in culture,

Data Centers and Cloud Computing 51

© K

itt ic

ha i/S

hu tte

rs to

ck FIGURE 2.18 Corporate IT infrastructures can consist of an on-premises data center and off-premises cloud computing.

Catalogs

3rd Party

Plug-Ins

Gateway Interface Module

A u

to m

a te

d P

o lic

ie s

M a

n a

g e

m e

n t

P ro

vi si

o n

in g

& M

o n

ito ri

n g

Security

Software-defined Networking

Software-defined Storage

Physical & Virtual Compute

Limited Functionality of Service Portals

FIGURE 2.19 SDDC infrastructure (adapted from Sturm et al., 2017).

processes, structure, and technology. The SDDC isolates the application layer from the physical infrastructure layer to facilitate faster and more effective deployment, management, and moni- toring of diverse applications. This is achieved by finding each enterprise application an optimal home in a public or private cloud environment or draw from a diverse collection of resources.

From a business perspective moving to a SDDC is motivated by the need to improve secu- rity, increase alignment of the IT infrastructure with business objectives and provision of appli- cations more quickly.

52 C H A P T E R 2 Information Systems, IT Architecture, Data Governance, and Cloud Computing

Traditional data centers had dedicated, isolated hardware that results in poor utilization of resources and very limited flexibility. Second-generation virtualization data cases improved resource use by consolidating virtualized servers. By reducing the steps needed to decrease the time it takes to deploy workloads, facilitating the definition of applications and resource needs, the SDDC creates an even more flexible environment in which enterprise applications can be quickly reconfigured and supported to provide infrastructure-as a service (IaaS). Tran- sitioning to an SDDC enables an organization to optimize its resource usage, provide capacity on demand, improve business-IT alignment, improve agility and flexibility of operations, and save money (Figure 2.20).

TRADITIONAL DATA CENTER

• Hardware silos • Limited utilization • Limited flexibility

VIRTUALIZED DATA CENTER

• Virtualized servers • Better resource use • Automate/Balance workload • Sub-optimal performance

SOFTWARE-DEFINED DATA CENTER (SDDC)

• IaaS • Optimized resource use • Increased business- IT alignment • Improved agility & flexibility • Capacity on demand • Cost savings

FIGURE 2.20 Evolution of data centers (adapted from Sturm et al., 2017).

Cloud Computing In a business world where first movers gain the advantage, IT responsiveness and agility pro- vide a competitive edge and lead to sustainable business practices. Yet, many IT infrastructures are extremely expensive to manage and too complex to easily adapt. A common solution is cloud computing. Cloud computing is the general term for infrastructures that use the Inter- net and private networks to access, share, and deliver computing resources. More specifically, IBM defines cloud computing as “the delivery of on-demand computing resources—everything from applications to data centers—over the Internet on a pay-for-use basis” (IBM, 2016).

Cloud computing is the delivery of computing and storage resources as a service to end-users over a network. Cloud systems are scalable. That is, they can be adjusted to meet changes in business needs. At the extreme, the cloud’s capacity is unlimited depending on the vendor’s offer- ings and service plans. A drawback of the cloud is control because a third party manages it. Unless the company uses a private cloud within its network, it shares computing and storage resources with other cloud users in the vendor’s public cloud. Public clouds allow multiple clients to access the same virtualized services and utilize the same pool of servers across a public network. In con- trast, private clouds are single-tenant environments with stronger security and control for reg- ulated industries and critical data. In effect, private clouds retain all the IT security and control provided by traditional IT infrastructures with the added advantages of cloud computing.

Selecting a Cloud Vendor Because cloud is still a relatively new and evolving business model, the decision to select a cloud service provider should be approached with even greater diligence than other IT deci- sions. As cloud computing becomes an increasingly important part of the IT delivery model, assessing and selecting the right cloud provider also become the most strategic decisions that business leaders undertake. Providers are not created equally, so it is important to investigate each provider’s offerings prior to subscribing. When selecting and investing in cloud services, there are several service factors a vendor needs to address. These evaluation factors are listed in Table 2.6.

Data Centers and Cloud Computing 53

Vendor Management and Cloud Service Agreements (CSAs) The move to the cloud is also a move to vendor-managed services and cloud service agreements (CSAs). Also referred to as cloud service level agreements (SLAs), the CSA or SLA is a negotiated agreement between a company and service provider that can be a legally binding contract or an informal contract. You can review a sample CSA used by IBM by visiting http://www-05. ibm.com/support/operations/files/pdf/csa_us.pdf.

Staff experienced in managing outsourcing projects may have the necessary expertise for managing work in the cloud and policing SLAs with vendors. The goal is not building the best CSA terms, but negotiating the terms that align most closely with the business needs. For example, if a server becomes nonoperational and it does not support a critical business operation, it would not make sense to pay a high premium for reestablishing the server within one hour. On the other hand, if the data on the server support a business process that would effectively close down the business for the period of time that it was not accessible, it would be prudent to nego- tiate the fastest possible service in the CSA and pay a premium for that high level of service.

In April 2015, the Cloud Standards Customer Council (CSCC) published the Practical Guide to Cloud Service Agreements, Version 2.0, to reflect changes that have occurred since 2012 when it first published the Practical Guide to Cloud Service Level Agreements. The new guide provides a practical reference to help enterprise IT and business decision-makers analyze CSAs from dif- ferent cloud service providers. The main purpose of a CSA is to set clear expectations for service between the cloud customer (buyer) and the cloud provider (seller), but CSAs should also exist between a customer and other cloud entities, such as the cloud carrier, the cloud broker, and even the cloud auditor. Although the various service delivery models, that is, IaaS, PaaS, SaaS, and so on, may have different requirements, the guide focuses on the requirements that are common across the various service models (Cloud Standards Customer Council, 2015, p. 4).

Implementing an effective management process is an important step in ensuring internal and external user satisfaction with cloud services. Table  2.7 lists the 10 steps that should be taken by cloud customers to evaluate cloud providers’ CSAs in order to compare CSAs across multiple providers or to negotiate terms with a selected provider.

TA B L E 2 . 6 Service Factors to Consider when Evaluating Cloud Vendors or Service Providers

Factors Examples of Questions to Be Addressed Delays What are the estimated server delays and network delays?

Workloads What is the volume of data and processing that can be handled during a specific amount of time?

Costs What are the costs associated with workloads across multiple cloud computing platforms?

Security How are data and networks secured against attacks? Are data encrypted and how strong is the encryption? What are network security practices?

Disaster recovery and business continuity

How is service outage defined? What level of redundancy is in place to minimize outages, including backup services in differ- ent geographical regions? If a natural disaster or outage occurs, how will cloud services be continued?

Technical expertise and understanding

Does the vendor have expertise in your industry or business processes? Does the vendor understand what you need to do and have the technical expertise to fulfill those obligations?

Insurance in case of failure Does the vendor provide cloud insurance to mitigate user losses in case of service failure or damage? This is a new and important concept.

Third-party audit or an unbiased assessment of the ability to rely on the service provided by the vendor

Can the vendor show objective proof with an audit that it can live up to the promises it is making?

54 C H A P T E R 2 Information Systems, IT Architecture, Data Governance, and Cloud Computing

Cloud Infrastructure The cloud has greatly expanded the options for enterprise IT infrastructures because any device that accesses the Internet can access, share, and deliver data. Cloud computing is a valuable infrastructure because:

1. It is dynamic, not static and provides a way to make applications and computing power available on demand. Applications and power are available on demand because they are provided as a service. For example, any software that is provided on demand is referred to as software as a service (SaaS). Typical SaaS products are Google Apps and www. Salesforce.com. Section 2.5 discusses SaaS and other cloud services.

2. Helps companies become more agile and responsive while significantly reducing IT costs and complexity through improved workload optimization and service delivery.

Move to Enterprise Clouds A majority of large organizations have hundreds or thou- sands of software licenses that support business processes, such as licenses for Microsoft Office, Oracle database management, IBM CRM (customer relationship management), and var- ious network security software. Managing software and their licenses involves deploying, provi- sioning, and updating them―all of which are time-consuming and expensive. Cloud computing overcomes these problems.

Issues in Moving Workloads from the Enterprise to the Cloud Building a cloud strategy is a challenge, and moving existing applications to the cloud is stress- ful. Despite the business and technical benefits, the risk exists of disrupting operations or cus- tomers in the process. With the cloud, the network and WAN (wide area network) become an even more critical part of the IT infrastructure. Greater network bandwidth is needed to sup- port the increase in network traffic. And, putting part of the IT architecture or workload into the cloud requires different management approaches, different IT skills, and knowing how to manage vendor relationships and contracts.

TA B L E 2 . 7 Ten Steps to Evaluate a CSA

1. Understand roles and responsibilities of the CSA customer and provider

2. Evaluate business-level policies and compliance requirements relevant to the CSA customer

3. Understand service and deployment model differences

4. Identify critical performance objectives such as availability, response time, and processing speed. Ensure they are measurable and auditable

5. Evaluate security and privacy requirements for customer information that has moved into the provider’s cloud and applications, functions, and services being operated in the cloud to provide required service to the customer

6. Identify service management requirements such as auditing, monitoring and reporting, measurement, provisioning, change management, and upgrading/patching

7. Prepare for service failure management by explicitly documenting cloud service capabilities and performance expectations with remedies and limitations for each

8. Understand the disaster recovery plan

9. Develop a strong and detailed governance plan of the cloud services on the customer side

10. Understand the process to terminate the CSA

Cloud Services and Virtualization 55

Infrastructure Issues There is a big difference because cloud computing runs on a shared infrastructure, so the arrangement is less customized to a specific company’s require- ments. A comparison to help understand the challenges is that outsourcing is like renting an apartment, while the cloud is like getting a room at a hotel.

With cloud computing, it may be more difficult to get to the root of performance problems, like the unplanned outages that occurred with Google’s Gmail and Workday’s human resources apps. The trade-off is cost versus control.

Increasing demand for faster and more powerful computers, and increases in the number and variety of applications are driving the need for more capable IT architectures.

Questions

1. What is a data center? 2. What is the difference between on-premises data centers and cloud computing? 3. What is an SDDC? 4. What are the advantages of using an SDDC? 5. How can cloud computing solve the problems of managing software licenses? 6. What factors should be considered when selecting a cloud vendor or provider? 7. When are private clouds used instead of public clouds? 8. Explain three issues that need to be addressed when moving to cloud computing or services.

2.5 Cloud Services and Virtualization Managers want streamlined, real-time, data-driven enterprises, yet they may face budget cuts. Sustaining performance requires the development of new business applications and analytics capabilities, which comprise the front end and the data stores and digital infrastructure, or back end, to support them. The back end is where the data reside. The problem is that data may have to navigate through a congested IT infrastructure that was first designed decades ago. These network or database bottlenecks can quickly wipe out the competitive advantages from big data, mobility, and so on. Traditional approaches to increasing database performance―manu- ally tuning databases, adding more disk space, and upgrading processors―are not enough when you are you are dealing with streaming data and real-time big data analytics. Cloud ser- vices help to overcome these limitations. Cloud services are outsourced to a third-party cloud provider who manages the updates, security, and ongoing maintenance.

At first glance, virtualization and cloud computing may appear to be quite similar. How- ever, cloud computing and virtualization are inherently different. Unlike cloud computing that involves multiple computers or hardware devices sending data through vendor-provided net- works, virtualization is the replacement of a tangible physical component with a virtual one. Each of these concepts are described and discussed in the following sections.

Anything as a Service (XAAS) Models The cloud computing model for on-demand delivery of and access to various types of com- puting resources also extends to the development of business apps. Figure  2.21 shows four “as a service” (XaaS) solutions based on the concept that the resource―software, platform, infrastructure, or data—can be provided on demand regardless of geolocation. As these as ser- vice solutions develop, the focus is changing from massive technology implementation costs to business-reengineering programs that enable XaaS platforms (Fresht, 2014).

Cloud services are services made available to users on demand via the Internet from a cloud computing provider’s servers instead of being accessed through an organization’s

56 C H A P T E R 2 Information Systems, IT Architecture, Data Governance, and Cloud Computing

on-premises servers. Cloud services are designed to provide easy, scalable access to applica- tions, resources, and services, and are fully managed by a cloud services provider.

Cloud computing is often referred to as a “stack” or broad range of services built on top of each other under the name cloud. These cloud services can be defined as follows:

• Software as a service (SaaS) is a widely used model in which software is available to users from a service provider as needed. A provider licenses a SaaS application to customers as an on-demand service, through a subscription, a pay-as-you-go model, or free of charge (where revenue can be generated by other means, such as through sale of advertisements).

• Platform as a service (PaaS) is a computing platform that enables the quick and easy creation, testing, and deployment of web applications without the necessity of buying and maintaining the software and infrastructure underneath it. It is a set of tools and services that make coding and deploying these applications faster and more efficient.

• Infrastructure as a service (IaaS) is a way of delivering servers, storage, networks, work- load balancers, and OSs as an on-demand service.

• Data as a service (DaaS) is an information provision and distribution model in which data files (including text, images, sounds, and videos) are made available to customers over a network by a service provider.

Software as a Service (SaaS) SaaS is a rapidly growing method of delivering soft- ware and is particularly useful in applications in which there are considerable interactions bet- ween the organization and external entities that do not confer a competitive advantage, for example, e-mail and newsletters. It is also useful when an organization is going to be needing a particular type of software for a short period of time or for a specific project, and for software that is used periodically, for example, tax, payroll, or billing software. SaaS is not appropriate for accessing applications that require fast processing of real-time data or applications where regulation does not permit data being hosted externally.

Other terms for SaaS are on-demand computing and hosted services. The idea is basically the same: Instead of buying and installing expensive packaged enterprise applications, users can access software applications over a network, using an Internet browser. To use SaaS, a service provider hosts the application at its data center and customers access it via a standard Web browser.

The SaaS model was developed to overcome the common challenge to an enterprise of being able to meet fluctuating demands on IT resources efficiently. It is used in many business functions, primarily customer relationship management (CRM), accounting, human resources (HR), service desk management, communication, and collaboration.

© V

al le

p u/

S hu

tte rs

to ck

FIGURE 2.21 Four as a service solutions: software, platform, infrastructure, and data as a service.

Cloud Services and Virtualization 57

There are thousands of SaaS vendors. www.Salesforce.com is one of the most widely known SaaS providers. Other examples are Google Docs and collaborative presentation soft- ware Prezi. For instance, instead of installing Microsoft Word on your own computer, and then loading Word to create a document, you use a browser to log into Google Docs. Only the browser uses your computer’s resources.

Platform as a Service (PaaS) PaaS provides a standard unified platform for devel- oping, testing, and deploying software over the Web. This computing platform allows the creation of Web applications quickly and easily without the complexity of buying and main- taining the underlying infrastructure. Without PaaS, the cost of developing some applications would be prohibitive. Examples of PaaS include databases, Web servers, development tools, and execution runtime. PaaS is particularly useful when multiple software developers are working on a software development project of when other external parties need to interact with the development process and for when developers want to automate testing and deploy- ment services. It is less useful in those instances where application performance needs to be customized to the underlying hardware and software or an application needs to be highly por- table in terms of where it is hosted. Some examples of PaaS include Microsoft Azure Service, www.Force.com, and Google App Engine.

Infrastructure as a Service (IaaS) Rather than purchasing all the components of its IT infrastructure, organizations buy their computing resources as a fully outsourced Infra- structure as a Service (IaaS) on demand. Generally, IaaS can be acquired as a Public or Private infrastructure or a combination of the two (Hybrid). A public IaaS is one that consists of shared resources deployed on a self-service basis over the Internet. On the other hand, a private IaaS is provided on a private network. And, a hybrid IaaS is a combination of both public and private. IaaS is useful where organizations experience significant highs and lows in terms of demand on the infrastructure, for new or existing organizations who have budgetary constraints on hardware investment and in situations where an organization has temporary infrastructure needs. Some IaaS providers you may be familiar with include Amazon Web Services (AWS) and Rackspace.

Data as a Service (DaaS)—The New Kid on the Block DaaS is the newest entrant into the XaaS arena. DaaS enables data to be shared among clouds, systems, apps, and so on regardless of the data source or where they are stored. Data files, including text, images, sound, and video, are made available to customers over a network, typically the Internet. DaaS makes it easier for data architects to select data from different pools, filter out sensitive data, and make the remaining data available on demand.

A key benefit of DaaS is that it transfers the risks and responsibilities associated with data management to a third-party cloud provider. Traditionally, organizations stored and managed their data within a self-contained storage system, however, as data become more complex, it is increasingly difficult and expensive to maintain using the traditional data model. Using DaaS, organizational data are readily accessible through a cloud-based platform and can be delivered to users despite organizational or geographical constraints. This model is growing in popularity as data become more complex, difficult, and expensive to maintain. Some of the most common business applications currently using DaaS are CRM and enterprise resource planning (ERP). For an example of Daas, see IT at Work 2.3.

IT at Work 2.3

Slack Slack, the successful social chat app for companies and their executives and/or employees, has announced a “deep product partnership” with Salesforce (Lunden,  2016). The partnership includes a new data sharing platform for businesses to easily share

information about conversations they are having within the app. More specifically, businesses will be able to share details about client accounts in real time with automatic updates for new leads about the accounts. The new partnership will allow Slack and its users to be even more effective in collaboration and data sharing across many platforms and departments (Lunden, 2016).

58 C H A P T E R 2 Information Systems, IT Architecture, Data Governance, and Cloud Computing

As a Service Models Are Enterprisewide and Can Trigger Lawsuits The various As a Service models are used in various aspects of business. You will read how these specific services, such as CRM and HR management, are being used for operational and strategic purposes in later chapters. Companies are frequently adopting software, platform, infrastructure, data management, and starting to embrace mobility as a service and big data as a service because they typically no longer have to worry about the costs of buying, maintaining, or updating their own data servers. Both hardware and human resources expenses can be cut significantly. Ser- vice arrangements all require that managers understand the benefits and trade-offs―and how to negotiate effective SLAs and CSAs. Regulations mandate that confidential data be protected regardless of whether the data are on-premises or in the cloud. Therefore, a company’s legal department needs to get involved in these IT decisions. Put simply, moving to cloud services is not simply an IT decision because the stakes around legal and compliance issues are very high.

Going Cloud Cloud services can advance the core business of delivering superior services to optimize busi- ness performance. Cloud can cut costs and add flexibility to the performance of critical busi- ness apps. And, it can improve responsiveness to end-consumers, application developers, and business organizations. But to achieve these benefits, there must be IT, legal, and senior man- agement oversight because a company still must meet its legal obligations and responsibilities to employees, customers, investors, business partners, and society.

Virtualization and Virtual Machines There are many types of virtualization, such as virtual storage devices, virtual desktops, virtual OSs, and virtual servers for network virtualization. You can think of virtualization as a model for a physical component that is built into computer code, to create a software program that acts in the same way as the physical component it is modeling. For example, a virtual machine is a software representation of a computer, rather than an actual computer and a virtual server sends and receives signals just like a physical one, even though it doesn’t have its own circuitry and other physical components.

You might ask why organizations want to virtualize their physical computing and net- working devices. The answer is a gross underutilization of inefficient use of resources. Computer hardware had been designed to run a single OS and a single app, which leaves most computers vastly underutilized. Virtualization is a technique that creates a virtual (i.e., nonphysical) layer and multiple virtual machines (VMs) to run on a single physical machine. The virtual (or virtual- ization) layer makes it possible for each VM to share the resources of the hardware. Figure 2.22 shows the relationship among the VMs and physical hardware.

Application

Virtualization Layer

Hardware Layer

Operating System

Application Operating

System

Application Operating

System

Virtual Machines

FIGURE 2.22 Virtual machines running on a simple computer hardware layer.

Cloud Services and Virtualization 59

What Is a Virtual Machine? Just as virtual reality is not real, but a software-created world, a virtual machine is a software-created computer. Technically, a virtual machine (VM) is created by a software layer, called the virtualization layer, as shown in Figure 2.22. That layer has its own Windows or other OS and apps, such as Microsoft Office, as if it were an actual physical computer. A VM behaves exactly like a physical computer and con- tains its own virtual―that is, software-based―CPU, RAM (random access memory), hard drive, and network interface card (NIC). An OS cannot tell the difference between a VM and a physical machine, nor can applications or other computers on a network tell the difference. Even the VM thinks it is a “real” computer. Users can set up multiple real com- puters to function as a single PC through virtualization to pool resources to create a more powerful VM.

Virtualization is a concept that has several meanings in IT and therefore several defini- tions. The major type of virtualization is hardware virtualization, which remains popular and widely used. Virtualization is often a key part of an enterprise’s disaster recovery plan. In gen- eral, virtualization separates business applications and data from hardware resources. This separation allows companies to pool hardware resources―rather than dedicate servers to applications―and assign those resources to applications as needed.

Different types of virtualization include:

• Storage virtualization is the pooling of physical storage from multiple network storage devices into what appears to be a single storage device managed from a central console.

• Server virtualization consolidates multiple physical servers into virtual servers that run on a single physical server.

• Desktop virtualization is software technology that separates the desktop environ- ment and associated application software from the physical machine that is used to access it.

• Application virtualization is the practice of running software from a remote server rather than on the user’s computer.

• Network virtualization combines the available resources in a network by splitting the network load into manageable parts, each of which can be assigned (or reassigned) to a particular server on the network.

• Hardware virtualization is the use of software to emulate hardware or a total computer environment other than the one the software is actually running in. It allows a piece of hardware to run multiple OS images at once. This kind of software is sometimes known as a virtual machine.

Virtualization Characteristics and Benefits Virtualization increases the flexi- bility of IT assets, allowing companies to consolidate IT infrastructure, reduce maintenance and administration costs, and prepare for strategic IT initiatives. Virtualization is not primarily about cost-cutting, which is a tactical reason. More importantly, for strategic reasons, virtual- ization is used because it enables flexible sourcing and cloud computing.

The characteristics and benefits of virtualization are as follows:

1. Memory-intensive VMs need a huge amount of RAM (random access memory, or pri- mary memory) because of their massive processing requirements.

2. Energy-efficient VMs minimize energy consumed running and cooling servers in the data center―representing up to a 95% reduction in energy use per server.

3. Scalability and load balancing When a big event happens, such as the Super Bowl, millions of people go to a website at the same time. Virtualization provides load balanc- ing to handle the demand for requests to the site. The VMware infrastructure automati- cally distributes the load across a cluster of physical servers to ensure the maximum performance of all running VMs. Load balancing is key to solving many of today’s IT challenges.

60 C H A P T E R 2 Information Systems, IT Architecture, Data Governance, and Cloud Computing

Virtualization consolidates servers, which reduces the cost of servers, makes more effi- cient use of data center space, and reduces energy consumption. All of these factors reduce the total cost of ownership (TCO). Over a three-year life cycle, a VM costs approximately 75% less to operate than a physical server. IT at Work 2.4 describes one example of how virtualization can help organizations provide higher levels of customer service and improve productivity.

Key Terms ad hoc report 34 batch processing 33 cloud computing 52 cloud service agreements (CSAs) 53 customer-centric 47 data 30 data as a service (DaaS) 56 data center 38 data governance 46

data silo 43 database 33 decision support systems (DSS) 32 dirty data 26 enterprise architecture (EA) 26 exception report 34 executive information systems (EISs) 32 goal seeking 35 information 30

information management 42 information systems (ISs) 28 infrastructure as a service (IaaS) 56 IT infrastructure 38 IPOS 28 knowledge 30 management information systems (MIS) 34 master data 46 master data management (MDM) 47

IT at Work 2.4

Business Continuity with Virtualization Liberty Wines supplies to restaurants, supermarkets, and indepen- dent retailers from its headquarters in central London. Recipient of multiple international wine awards—including the Interna- tional Wine Challenge on Trade Supplier of the Year for two years running—Liberty Wines is one of the United Kingdom’s foremost wine importers and distributors.

IT Problems and Business Needs As the business expanded, the existing servers did not have the capacity to handle increased data volumes, and maintenance of the system put a strain on the IT team of two employees. Existing systems were slow and could not provide the responsiveness that employees expected.

Liberty Wines had to speed up business processes to meet the needs of customers in the fast-paced world of fine dining. To provide the service their customers expect, employees at Liberty Wines needed quick and easy access to customer, order, and stock

information. In the past, the company relied on 10 physical servers for applications and services, such as order processing, reporting, and e-mail.

Virtualized Solution Liberty Wines deployed a virtualized server solution incorporating Windows Server 2008 R2. The 10 servers were replaced with 3 physical servers, running 10 virtual servers. An additional server was used as part of a backup system, further improving resilience and stability.

By reducing the number of physical servers from 10 to 4, power use and air conditioning costs were cut by 60%. Not only was the bottom line improved, but the carbon footprint was also reduced, which was good for the environment.

The new IT infrastructure cut hardware replacement costs by £45,000 (U.S. $69,500) while enhancing stability with the backup system. Applications now run faster, too, so employees can pro- vide better customer service with improved productivity. When needed, virtual servers can be added quickly and easily to support business growth.

Questions

1. What is SaaS? 2. What is PaaS? 3. What is IaaS? 4. How might companies risk violating regulation or compliance requirements with cloud services? 5. In what ways is a virtualized information system different from a traditional information system? 6. Describe the different types of virtualization. 7. What is load balancing and why is it important?

Assuring Your Learning 61

master file 47 model 26 online transaction processing (OLTP) 33 platform as a service (PaaS) 56 private cloud 52 public cloud 52

real-time processing 33 service level agreement (SLA) 61 software as a service (SaaS) 54 software-defined data center (SDDC) 50 stack 56 structured decisions 35

transaction processing systems (TPS) 32 unstructured decisions 35 virtualization 59 virtual machine (VM) 59 what-if analysis 35 wisdom 31

Assuring Your Learning

Discuss: Critical Thinking Questions

1. Why is a strong market position or good profit performance only temporary?

2. Assume you had: a. A tall ladder with a sticker that lists a weight allowance only five pounds more than you weigh. You know the manufacturer and model number.

b. Perishable food with an expiration date two days into the future.

c. A checking account balance that indicates you have sufficient funds to cover the balance due on an account.

In all three cases, trusting the data to be exactly correct could have negative consequences. Explain the consequences of trusting the data in each instance. How might you determine the correct data for each instance? Which data might not be possible to verify? How does dirty data impact your decision-making? 3. If business data are scattered throughout the enterprise and not synched until the end of the month, how does that impact day-to-day decision-making and planning?

4. Assume a bank’s data are stored in silos based on financial product―checking accounts, saving accounts, mortgages, auto loans, and so on. What problems do these data silos create for the bank’s managers?

5. Why do managers and workers still struggle to find information that they need to make decisions or take action despite advances in digital technology? That is, what causes data deficiencies?

6. According to a Tech CEO Council Report, Fortune 500 companies waste $480 billion every year on inefficient business processes. What factors cause such huge waste? How can this waste be reduced?

7. Explain why organizations need to implement EA and data governance.

8. What two problems can EA solve? 9. Name two industries that depend on data governance to comply with regulations or reporting requirements. Given an example of each.

10. Why is it important for data to be standardized? Give an example of unstandardized data.

11. Why are TPSs critical systems? 12. Discuss why the cloud acts as the great IT delivery frontier. 13. What are the functions of data centers? 14. What factors need to be considered when selecting a cloud vendor? 15. What protection does an effective SLA or CSA provide? 16. Why is an SLA or a CSA a legal document? 17. How can virtualization reduce IT costs while improving performance?

Explore: Online and Interactive Exercises

1. When selecting a cloud vendor to host your enterprise data and apps, you need to evaluate the service level agreement (SLA).

a. Research the SLAs of two cloud vendors, such as Rackspace, Amazon, or Google.

b. For the vendors you selected, what are the SLAs’ uptime per- centages? Expect them to be 99.9% or less.

c. Does each vendor count both scheduled downtime and planned downtime toward the SLA uptime percentage?

d. Compare the SLAs in terms of two other criteria. e. Decide which SLA is better based on your comparisons. f. Report your results and explain your decision.

2. Many organizations initiate data governance programs because of pressing compliance issues that impact data usage. Organizations may

need data governance to be in compliance with one or more regula- tions, such as the Gramm−Leach Bliley Act (GLB), HIPAA, Foreign Cor- rupt Practices Act (FCPA), Sarbanes−Oxley Act, and several state and federal privacy laws.

a. Research and select two U.S. regulations or privacy laws. b. Describe how data governance would help an enterprise com- ply with these regulations or laws.

3. Visit www.eWeek.com Cloud Computing Solutions Center for news and reviews at www.eweek.com/c/s/Cloud-Computing. Select one of the articles listed under Latest Cloud Computing News. Prepare an executive summary of the article.

4. Visit Rackspace.com and review the company’s three types of cloud products. Describe each of those cloud solutions.

62 C H A P T E R 2 Information Systems, IT Architecture, Data Governance, and Cloud Computing

5. Visit Oracle.com. Describe the types of virtualization services of- fered by Oracle.

6. Visit YouTube.com and search for two videos on virtualization. For

each video, report what you learned. Specify the complete URL, video title, who uploaded the video and the date, video length, and num- ber of views.

Case 2.2 Business Case: Data Chaos Creates Risk Data chaos often runs rampant in service organizations, such as health care and the government. For example, in many hospitals, each line of business, division, and department has implemented its own IT applications, often without a thorough analysis of its relationship with other departmental or divisional systems. This arrangement leads to the hospital having IT groups that specifically manage a particu- lar type of application suite or data silo for a particular department or division.

Data Management When applications are not well managed, they can generate terabytes of irrelevant data, causing hospitals to drown in such data. This data chaos could lead to medical errors. In the effort to manage excessive and massive amounts of data, there is increased risk of relevant infor- mation being lost (missing) or inaccurate—that is, faulty or dirty data. Another risk is data breaches.

• Faulty data By 2015, 96% of health-care organizations had adopted electronic health records, or EHRs (Office of the National Coordinator for HIT, 2016). It is well known that an unintended consequence of EHR is faulty data. According to a study pub- lished in the Journal of the American Medical Association, data in EHR systems may not be as accurate and complete as expected (Conn,  2016). Incorrect lab values, imaging results, or physi- cian documentation lead to medical errors, harm patients, and damage the organization’s accreditation and reputation.

• Data breaches More than 25 million people have been affected by health-care system data breaches since the Office for Civil Rights, a division of the U.S. Department of Health and Human Services, began reporting breaches in 2009. Most breaches involved lost or stolen data on laptops, removable drives, or other portable media. Breaches are extremely expensive and destroy trust.

Accountability in health-care demands compliance with strong data governance efforts. Data governance programs verify that data input into EHR, clinical, financial, and operational systems are accu- rate and complete—and that only authorized edits can be made and logged.

Vanderbilt University Medical Center Adopts EHR and Data Governance Vanderbilt University Medical Center (VUMC) in Nashville, TN, was an early adopter of EHR and implemented data governance in 2009. VUMC’s experience provides valuable lessons.

VUMC consists of three hospitals and the Vanderbilt Clinic, which have 918 beds, discharge 53,000 patients each year, and count 1.6 mil- lion clinic visits each year. On average, VUMC has an 83% occupancy rate and has achieved HIMSS Stage 6 hospital EHR adoption. HIMSS (Healthcare Information and Management Systems Society, himss.org) is a global, nonprofit organization dedicated to better health-care out- comes through IT. There are seven stages of EHR adoption, with Stage 7 being a fully paperless environment. That means all clinical data are part of an electronic medical record and, as a result, can be shared

Analyze & Decide: Apply IT Concepts to Business Decisions

1. Financial services firms experience large fluctuations in business volumes because of the cyclical nature of financial markets. These fluctuations are often caused by crises―such as the subprime mort- gage problems, the discovery of major fraud, or a slowdown in the economy. These fluctuations require that executives and IT leaders have the ability to cut spending levels in market downturns and quickly scale up when business volumes rise again. Research SaaS solutions and vendors for the financial services sector. Would invest- ment in SaaS help such firms align their IT capacity with their business needs and also cut IT costs? Explain your answer.

2. Despite multimillion-dollar investments, many IT organizations cannot respond quickly to evolving business needs. Also, they cannot

adapt to large-scale shifts like mergers, sudden drops in sales, or new product introductions. Can cloud computing help organizations improve their responsiveness and get better control of their IT costs? Explain your answer.

3. Describe the relationship between enterprise architecture and organizational performance.

4. Identify four KPIs for a major airline (e.g., American, United, Delta) or an automobile manufacturer (e.g., GM, Ford, BMW). Which KPI would be the easiest to present to managers on an online dashboard? Explain why.

Case 2.3 63

across and outside the enterprise. At Stage 7, the health-care organi- zation is getting full advantage of the health information exchange (HIE). HIE provides interoperability so that information can flow back and forth among physicians, patients, and health networks (NextGen Healthcare, 2016).

VUMC began collecting data as part of its EHR efforts in 1997. By 2009, the center needed stronger, more disciplined data management. At that time, hospital leaders initiated a project to build a data governance infrastructure.

Data Governance Implementation VUMC’s leadership team had several concerns.

1. IT investments and tools were evolving rapidly, but they were not governed by HIM (Healthcare Information and Manage- ment) policies.

2. As medical records became electronic so they might be trans- mitted and shared easily, they became more vulnerable to hacking.

3. As new uses of electronic information were emerging, the medi- cal center struggled to keep up.

Health Record Executive Committee Initially, VUMC’s leaders assigned data governance to their traditional medical records committee, but that approach failed. Next, they hired consultants to help develop a data governance structure and organ- ized a health record executive committee to oversee the project. The committee reports to the medical board and an executive commit- tee to ensure executive involvement and sponsorship. The commit- tee is responsible for developing the strategy for standardizing health record practices, minimizing risk, and maintaining compliance. Mem- bers include the chief medical information officer (CMIO), CIO, legal counsel, medical staff, nursing informatics, HIM, administration, risk management, compliance, and accreditation. In addition, a legal medical records team was formed to support additions, corrections, and deletions to the EHR. This team defines procedures for removal of

duplicate medical record numbers and policies for data management and compliance.

Costs of Data Failure Data failures incur the following costs:

• Rework • Loss of business • Patient safety errors • Malpractice lawsuits • Delays in receiving payments because billing or medical codes

data are not available.

Benefits Achieved from Data Governance As in other industries, in health care, data are the most valuable asset. The handling of data is the real risk. EHRs are effective only if the data are accurate and useful to support patient care. Effective ongoing data governance has achieved that goal at VUMC.

Questions 1. What might happen when each line of business, division, and

department develops its own IT apps?

2. What are the consequences of poorly managed apps? 3. What two risks are posed by data chaos? Explain why. 4. What are the functions of data governance in the health-care

sector?

5. Why is it important to have executives involved in data gover- nance projects?

6. List and explain the costs of data failure. 7. Why are data the most valuable asset in health care?

Case 2.3 Video Case: Cloud Computing at Coca-Cola Is Changing Everything When organizations say they are “using the cloud,” they can mean a number of very different things. Using an IaaS service such as Amazon EC2 or Terremark is different from using Google Apps to outsource e-mail, which is different again from exposing an API in Facebook.

In this video Alan Boehme, CIO of the Coca-Cola Company discusses how Coca-Cola uses cloud computing to more effectively interact with its customers and describes the challenges Coca-Cola is facing in establish- ing SaaS partnerships with new start-ups.

Complete these three steps:

1. Visit https://www.youtube.com/watch?v=hCxmsSED2DY 2. View the 13-minute video. 3. Answer each of the three parts of the following question.

Question 1. Explain the value of Coca-Cola’s cloud partnerships with start-up

companies to: a. Coca-Cola b. The start-up companies c. Coca-Cola’s customers

Sources: Compiled from NextGen Healthcare (2016), Office of the National Coor- dinator for HIT (2016), and Conn (2016).

64 C H A P T E R 2 Information Systems, IT Architecture, Data Governance, and Cloud Computing

IT Toolbox

Accurately Measuring the Value of Data Governance When developing a data governance program, it’s important to pre- sent a strong business case to get buy-in from top executives and stakeholders. A crucial part of the business case is an estimate of the data governance program’s return on investment (ROI) to show how it will add value to the company. You will need to justify the ROI based on both business and IT strategy to ensure that available funds are used to best meet the business objectives.

To do this you will need to carefully analyze the IT infrastruc- ture with regard to how different components of the IT infrastructure work together to support business processes, how data needed by one system can be received and used by another, how easily data can be communicated and/or repurposed. You will also need to factor in risks and adverse events such as costs associated with rework in data collection, costs associated with unreliable or unfit data, and delays associated with untimely or unavailable data. Now, all of these costs must be quantified and your level of confidence in the corporate data has to be calculated to ensure your business case accurately reflects the value of a data governance program.

One metric used to make this calculation is the confidence in data- dependent assumptions metric, or CIDDA (Reeves & Bowen, 2013). The CIDDA identifies specific areas of deficiency.

So, to sum up, when building a data governance model, it is necessary to:

1. Establish a leadership team 2. Define the program’s scope 3. Calculate the ROI using the CIDDA.

CIDDA is computed by multiplying three confidence estimates using the following formula:

CIDDA G M TS where

G = Confidence that data are good enough for their intended purpose

M = Confidence that data mean what you think they do

TS = Confidence that you know where the data come from and trust the source.

CIDDA is a subjective metric for which there are no industry benchmarks, yet it can be evaluated over time to gauge improve- ments in data quality confidence.

To ensure your understanding of this IT Toolbox item, calculate the CIDDA of Company A over time, using the stated levels of confi- dence in the different aspects of its corporate data over Q1–Q4 2017:

Q1_ 2017 : 40%, 50%, 20% Q2 _ 2017 : 50%, 55%, 30% Q3_ 2017 : 60%, 60%, 40% Q4 _ 2017 : 60%, 70%, 45%

G M TS G M TS G M TS G M TS

References Bloomberg, J. “Change as Core Competency: Transforming the Role

of the Enterprise Architect.” Forbes, June 16, 2016. Cailean, I. “What Role Do Algorithms Play in Programmatic Advertis-

ing?” Trade Mod, January 6, 2016. http://www.trademob.com/what- role-do-algorithms-play-in-programmatic-advertising

Cloud Standards Customer Council. Practical Guide to Cloud Service Agreements, Version 2.0. April 2015. http://www.cloud-council.org/ deliverables/CSCC-Practical-Guide-to-Cloud-Service-Agreements .pdf

Conn, J. “EHRs vs. Paper: A Split-decision on Accuracy.” Modern Healthcare, July 8, 2016.

Fresht, P. “The Ten Tenets Driving the As-a-service Economy.” Horses for Sources, October 6, 2014. http://www.horsesforsources.com/as a service-economy_100614

IBM. “What is Cloud Computing?” IBM, June 6, 2016. https://www. ibm.com/cloud-computing/learn-more/what-is-cloud-computing

Jarousse, L. A. “Information Governance for Hospitals.” Hospitals & Health Networks, February 18, 2016.

Keitt, T. J. “Collaboration Technology Should Be Part of Your Cus- tomer Experience Tool Kit.” Forrester.com, June 30, 2014.

Lunden, I. “Enterprise Chat App Slack Ties up with Salesforce in a Deep Product Partnership.” Tech Crunch, September 27, 2016. https://techcrunch.com/2016/09/27/enterprise-chat-app-slack-ties- up-with-salesforce-in-a-deep-platform-partnership

Marchese, L. “How the ‘Silo Effect’ Is Hurting Cross Team Collabora- tion.” Trello, May 10, 2016.

NextGen Healthcare. “Health Information Exchange (HIE).” NextGen Healthcare, March 31, 2016.

Office of the National Coordinator for Health Information Technology. “Percent of Hospitals, By Type, that Possess Certified Health IT.” Office of the National Coordinator for Health Information Technology, May 31, 2016.

Porter, M. Competitive Advantage: Creating and Sustaining Superior Performance. Free Press, 1998.

Rai, R., G. Sahoo, and S. Mehfuz. “Exploring the Factors Influencing the Cloud Computing Adoption: A Systematic Study on Cloud Migra- tion.” Springerplus, April 25, 2015, 4, 197.

Reeves, M. G. and R. Bowen. “Developing a Data Governance Model in Health Care.”Healthcare Financial Management, February 2013, 67(2): 82–86.

Schneider, M. “Case Study: How MEDIATA Increased Campaign Perfor- mance with Hyperlocal Targeting.” Skyhook Wireless, July 22, 2014.

Schneider, M. “Solving the Dirty Data Problem in Location-Based Advertising.” Street Fight, January 7, 2015.

Shore, J. “Cloud-Based Integration Seeks to Tear Down Data Silos.” Tech Target, August 19, 2015.

Sturm, R., C. Pollard, and J. Craig. Application Performance Manage- ment in the Digital Enterprise. Elsevier, March 2017.

Zuckerman, M.P.H., Sheingold, Ph.D., Orav, Ph.D., Ruhter, M.P.P., M.H.S.A., and Epstein, M.D. “Readmissions, Observation, and the Hospital Readmissions Reduction Program.” The New England Jour- nal of Medicine, April 21, 2016.