Today Many Business Firms Continue to Operations Legacy Systems
CONTEMPORARY HARDWARE PLATFORM TRENDS
Although the cost of computing per se has fallen exponentially, the cost of the IT infrastructure has not followed suit, but rather has expanded as a percentage of corporate revenues and budgets. Why? The cost of computing services (consulting, systems integration) has risen dramatically, the cost of software remains high, and the intensity of computing and communicating has increased as other costs have declined. For instance, employees now use much more sophisticated applications, requiring more powerful and expensive hardware of many different types (laptop, desktop, handheld, and tablet computers).
Firms also face a number of other challenges. They need to integrate information stored in different applications, on different platforms (telephone, legacy systems, intranet, Internet sites, desktop, and mobile devices). Firms also need to build resilient infrastructure that can withstand huge increases in peak loads and routine assaults from hackers and viruses. Because customer and employee expectations for service are increasing, firms need to increase their service levels to meet customer demands.
Each of the trends in hardware and software platforms we now describe seeks to address some or all of these challenges.
The Integration of Computing and Telecommunications Platforms
Arguably the most dominant theme in hardware platforms today is the convergence of telecommunications and computing platforms to the point where, increasingly, computing takes place over the network. You can see this convergence at several levels.
At the client level, communication devices such as cell phones are taking on functions of handheld computers, whereas handheld personal digital assistants (PDAs) are taking on cell phone functions. For instance, the Palm Treo 600 digital handheld integrates phone, camera, and handheld computer in one device. Soon full-length films and short commercials will be distributed over cell-phone-like devices. Television, radio, and video are moving toward all-digital production and distribution. There is little doubt that personal computers of some sort will be the core of the home entertainment center and the mobile personal entertainment center of the next five years as a storage device and operating system.
The Palm Treo 600 combines a mobile phone, e-mail, personal organizer, Web access, and camera all in one device. The convergence of computing and communications technologies has turned cell phones into mobile computing platforms.
At the server and network level, the growing success of Internet telephone systems (now the fastest-growing type of telephone service) demonstrates how historically separate telecommunications and computing platforms are converging toward a single network �the Internet. Chapter 8 describes the convergence of computing and telecommunications in greater depth.Other major trends in hardware platforms described here are based in large part on computing over high-capacity networks. The network in many respects is becoming the source of computing power, enabling business firms to expand their computing power greatly at very little cost.
Grid Computing
Grid computing involves connecting geographically remote computers into a single network to create a virtual supercomputer by combining the computational power of all computers on the grid. Grid computing takes advantage of the fact that most computers in the United States use their central processing units on average only 25 percent of the time for the work they have been assigned, leaving these idle resources available for other processing tasks. Grid computing was impossible until high-speed Internet connections enabled firms to connect remote machines economically and move enormous quantities of data.
Grid computing requires software programs to control and allocate resources on the grid, such as open-source software provided by Globus Alliance (www.globus.org) or other private providers. Client software communicates with a server software application. The server software breaks data and application code into chunks that are then parceled out to the grid�s machines. The client machines can perform their traditional tasks while running grid applications in the background.
The business case for using grid computing involves cost savings, speed of computation, and agility. For example, Royal Dutch/Shell Group is using a scalable grid computing platform that improves the accuracy and speed of its scientific modeling applications to find the best oil reservoirs. This platform, which links 1,024 IBM servers running Linux, in effect creates one of the largest commercial Linux supercomputers in the world. The grid adjusts to accommodate the fluctuating data volumes that are typical in this seasonal business. Royal Dutch/Shell Group claims the grid has enabled the company to cut processing time for seismic data, while improving output quality and helping its scientists pinpoint problems in finding new oil supplies.
In another example, the University of Pennsylvania is using grid computing to promote early detection of breast cancer, a disease that affects one out of every eight women today. The grid captures, manages, stores, and analyzes patient files from multiple locations throughout North America for fast retrieval and diagnostic evaluation. The system is being commercialized as the National Digital Mammography Archive (NDMA). The NDMA can now deliver images and physician notes in less than 90 seconds to remote grid users.This improved access to patient records enables physicians to make rapid and well-informed diagnoses in breast cancer screening, potentially saving numerous lives through early detection (IBM, 2004).
On-Demand Computing (Utility Computing)
On-demand computing refers to firms off-loading peak demand for computing power to remote, large-scale data processing centers. In this manner, firms can reduce their investment in IT infrastructure by investing just enough to handle average processing loads and paying for only as much additional computing power as the market demands. Another term for on-demand computing is utility computing, which suggests that firms purchase computing power from central computing utilities and pay only for the amount of computing power they use, much as they would pay for electricity.
IBM is investing $10 billion to bring this vision to reality and has created four on-demand computing centers around the United States where businesses can experiment with the concepts. HP�s Adaptive Enterprise offers similar capabilities.
In addition to lowering the cost of owning hardware resources, on-demand computing gives firms greater agility to use technology. On-demand computing shifts firms from having a fixed infrastructure capacity toward a highly flexible infrastructure, some of it owned by the firm, and some of it rented from giant computer centers owned by IBM and HP. This arrangement frees firms to launch entirely new business processes that they would never attempt with a fixed infrastructure.
For example, the Canadian Imperial Bank of Commerce (CIBC) outsources a significant portion of its infrastructure to Hewlett-Packard�s Adaptive Computing Center. HP provides support for 28,000 CIBC e-mail users, 41,000 desktop PCs, 4,500 automatic teller machines (ATMs), and 10,000 point-of-sale terminals. CIBC management believed that it was more cost-effective to outsource this portion of its infrastructure and focus on banking rather than operating its own IT infrastructure for these functions (HP, 2004).
In another example, specialty food and gift specialist Harry and David uses IBM�s On Demand services to deal with an annual traffic surge before the gift-giving season. About 65 percent of annual sales takes place between mid-November and late December. Rather than purchase enough infrastructure to handle this peak load, Harry and David rents this capacity when they need it (Shankland, 2003; IBM, 2004b).
Additional examples of value from on-demand computing can be found in the Window on Management. Qantas Airways contracted with IBM for a flexible on-demand computing arrangement and as a way of lowering operational costs, helping it remain competitive against other carriers. Ford Motor Company Europe switched to an IBM on-demand service for its help desk. Both companies pay IBM a basic fee for an anticipated level of service and then pay IBM additional fees for any spikes if they use more computing resources than planned.
Autonomic Computing
Computer systems have become so complex today that some experts believe they may not be manageable in the near future. With operating systems, enterprise, and database software weighing in at millions of lines of code, and large systems encompassing many thousands of networked devices, the problem of managing these systems looms very large (Kephardt and Chess, 2003).
It is estimated that one-third to one-half of a company�s total IT budget is spent preventing or recovering from system crashes. About 40 percent of these crashes are caused by operator error. The reason is not because operators are not well trained or do not have the right capabilities. Rather, it is because the complexities of today�s computer systems are too difficult to understand, and IT operators and managers are under pressure to make decisions about problems in seconds.
Today, the consequences of an outage are far more serious than in the past. For example, energy firms stand to lose close to $3 million in revenue for every hour systems are not operational. Therefore, firms must spend extraordinary resources to ensure systems do not fail. You can find out more about the impact of system outages and preventive measures in Chapter 10.
One approach to dealing with this problem from a computer hardware perspective is to employ autonomic computing. Autonomic computing is an industry-wide effort to develop systems that can configure themselves, optimize and tune themselves, heal themselves when broken, and protect themselves from outside intruders and self-destruction. Imagine, for instance, a desktop PC that could know it was invaded by a computer virus. Instead of blindly allowing the virus to invade, the PC would identify and eradicate the virus or, alternatively, turn its workload over to another processor and shut itself down before the virus destroyed any files.
By choosing the word autonomic, computer firms such as HP and IBM are making an analogy with the autonomic nervous system. The autonomic nervous system controls lower-level but vital functions of the human body without raising them to a conscious level. The vision of autonomic computing is that it will free system administrators from many of today�s routine system management and operational tasks so administrators can devote more of their IT skills to fulfilling the needs of their core businesses (Ganek and Corbi 2003).
Table 6-4 describes the four dimensions of autonomic computing. Currently, the bare outlines of some of these capabilities are already present in desktop operating systems. For instance, virus and firewall protection software can detect viruses on PCs, automatically defeat the viruses, and alert operators. These programs can be updated automatically as the need arises by connecting to an online virus protection service such as McAfee. Other key elements of autonomic computing are still missing but are sure to be on the agenda of research centers in the next decade. Without progress toward self-managing systems, the management resources required to operate a large-scale computing environment may be cost prohibitive.
TABLE 6-4 Four Aspects of Self-Management as They Are Now and Would Be with Autonomic Computing
ON-DEMAND COMPUTING AT QANTAS AIRWAYS AND FORD MOTOR COMPANY EUROPE
Qantas Airways is the leading airline in Australia, flying to about 140 destinations in Australia and 33 other countries. Although recent profits have been strong, the company faces cost pressures from high fuel prices and lower levels of global airline traffic since 2000. Moreover, two-thirds of the international airlines flying to Australia are owned or subsidized by their governments and Qantas must try to keep prices in line with its competition. Keeping costs low and service to customers high are very big priorities.
When Qantas decided to replace its 30-year-old data center, it had a choice: It could build a new one to be maintained in-house or it could outsource its IT infrastructure. IBM Global Services offered an arrangement for on-demand computing with flexible pricing that was too good to pass up. In May 2004, Qantas signed a 10-year contract with IBM worth more than $450 million for these services.
Qantas CIO Fiona Balfour believes that airlines no longer obtain a competitive advantage because they possess information technology. The advantage lies in �how they use it. Running IT at a low unit cost becomes a competitive advantage,� says Balfour. On-demand computing with IBM enables Qantas to better match its IT infrastructure to actual business needs and to �get out of the capital investment cycle.� Qantas is able to predict its computing requirements by looking at how it is filling aircraft seats.
Qantas�s arrangement with IBM covers primarily the infrastructure and systems that support its core applications. Most of Qantas�s 450 Unix servers, which had been running at less than 50 percent of capacity, will be replaced by Linux servers at IBM�s hosting center in Sydney, Australia. The contract established a baseline usage of computing resources that is less than half of what Qantas was paying for fixed IT costs. Qantas can adjust its usage upward or downward, depending on business conditions, and should be ensured fast response times for its processing.
Qantas will continue to develop its own software applications, while IBM will handle change management for application development and maintenance and the migration of applications from testing to production status. IBM will also provide Qantas with services for information system procurement and security and a service desk for resolving issues with all the providers of its IT services.
According to Brendon Riley, General Manager for IBM Global Services Australia and New Zealand, �It is not only IBM�s data center infrastructure and variable cost model that will provide increased flexibility and efficiency to Qantas. IBM gives Qantas access to thousands of professionals with an extensive range of skills as well."
Qantas awarded a $500 million outsourcing contract to Telstra Corporation to convert its telecommunications infrastructure into a network based on the Internet Protocol (IP) and to provide data, voice, and desktop services.
Outsourcing much of Qantas�s IT infrastructure to IBM has not only reduced fixed costs for information technology but also has freed its new employees to pursue technology-modernization projects that may have higher value. By outsourcing to IBM and Telstra, Qantas was able to cut down its information systems staff from 900 to 700 people.
IBM believes the travel industry is especially well suited for on-demand computing because travel companies are looking for ways to shift their focus from IT operations to core business issues.
Another company that has benefited from IBM�s on-demand computing is Ford Motor Company Europe. Ford Motor Company is the world�s second largest car and truck manufacturer, with 350,000 employees across the globe. Ford Motor Company Europe has a Company to Dealer Systems Division that is responsible for infrastructure and application support to its European Dealer network.
Ford�s lines of business, which are the �customers� of Ford�s Company to Dealer Systems (CDS), wanted more variable and immediate support. CDS maintained an automated help desk service that was fixed in price and capabilities, with little flexibility to address constantly changing dealer application needs.
Ford worked with IBM Business Consulting Services/AMS to design a �pay-as-you-go� help desk solution hosted and staffed by IBM. Computing resources for the help desk are allocated based on a quarterly computer capacity plan reflecting estimated volume of calls, user registrations, and application launches. IBM bills Ford for usage over the capacity plan at a premium to encourage efficient planning.
This on-demand arrangement has made Ford�s CDS a more responsive provider of IT services, adding to retail customer satisfaction and to Ford Europe�s brand reputation. By using efficient demand forecasting and capacity planning, combined with on-demand computing services, Ford Europe is saving 10 percent per help desk call.
Source: Tony Kontzer, �Airline Taps IBM for Flexible Pricing Deal,� Information Week, May 24, 2004 and �Qantas Airways Hands over IT Management to IBM, Information Week, May 17, 2004; �Case Studies: Ford Motor Company Europe�Company to Dealer Systems,� www.ibm.com, accessed September 10, 2004; IBM Media Relations, �Qantas Awards IBM Global Services a $650 million IT Services Contract,� May 17, 2004; and Qantas Corporate Communication, �Qantas Faces Significant Challenges,� August 19, 2004. To Think About: What are the management benefits of on-demand computing for these companies? How do IBM�s on-demand services provide value for them?
Edge Computing
Edge computing is a multitier, load-balancing scheme for Web-based applications in which significant parts of Web site content, logic, and processing are performed by smaller, less expensive servers located nearby the user. In this sense, edge computing is another technique like grid computing and on-demand computing for using the Internet to share the workload experienced by a firm across many computers located remotely on the network.
Figure 6-12 illustrates the components of edge computing. There are three tiers in edge computing: the local client; the nearby edge computing platform, which consists of servers positioned at any of the 5,000-plus Internet service providers in the United States; and enterprise computers located at the firm�s main data center. The edge computing platform is owned by a service firm such as Akamai, which employs about 15,000 edge servers around the United States.
FIGURE 6-12 Edge computing platform
Edge computing involves the use of the Internet to balance the processing load of enterprise platforms across the client and edge computing platform.
There are four main business benefits to edge computing:
-
Technology costs are lowered because the firm does not need to purchase infrastructure at its own data center to handle all of the requests from customers and can instead focus on building infrastructure to contain corporate databases and business logic.
-
Service levels are enhanced because response time of Web-based applications is reduced. Local client computers can be served with most static and dynamic content immediately, and customer response time is dramatically reduced.
-
Flexibility of the firm is enhanced because it can respond to business opportunities quickly without going through a lengthy infrastructure acquisition cycle by making use of service provider platforms that already exist.
-
Resilience is increased because seasonal or even daily spikes in the workload can be shared across the Web, and failure in edge nodes will not disable the entire system.
Source: https://paginas.fe.up.pt/~acbrito/laudon/ch6/chpt6-3fulltext.htm
0 Response to "Today Many Business Firms Continue to Operations Legacy Systems"
Post a Comment