Centralization and Decetraliztion
When hardware was expensive, all data, software, and employees were centralized. As the use of midrange and personal computers expanded, hardware became decentralized and software and data began to follow it. For example, decision makers stored spreadsheets, analyses, and reports on their personal computers. But this decentralized data is more difficult to share, even with networks. At this stage the Internet technologies become important—the search engines and browsers were designed to make it easier to find and view data in many forms.
Applying these technologies within the company (called an intranet) provided easier access to corporate data. Network technologies expanded to provide faster access to data and services on distant computers. Central computer services became standardized and companies emphasized purchasing software (particularly ERP systems around 2000). Servers became cheaper and simpler to operate. Some companies outsourced many tasks, including server and network management. Decentralization created some security issues, so conflict often arose over control of data.
By 2010, the pieces were in place for companies to re-centralize their operations by using Web servers and software as a service. Software and data could be stored on cloud-based servers with users accessing everything through browser-based platforms including tablets and cell phones. It is still not clear whether this transfer to cloud-based computing will be a trend or simply a tool to support certain tasks. But this return to server-based computing highlights the tradeoffs and conflicts between centralization and decentralization.
The basic argument for centralization revolves around the need to coordinate activities and efficiencies that can be gained from large-scale operations. Proponents of decentralization argue that moving control to smaller units produces a more flexible system that can respond faster to market changes, encourage individual differences, and innovate. As with many arguments, there are different answers for different circumstances, and it is rare that the extreme choices are best. Wise managers will attempt to gain the advantages of both approaches.
With information systems, four basic areas are subject to centralization or decentralization: hardware, software, data, and staffing. Determining the best way to organize information resources requires that managers understand the advantages and disadvantages for each of these areas.
Hardware
Today, hardware is relatively inexpensive. Even centralized servers have come down in cost, often using systems based on server farms consisting of hundreds or thousands of inexpensive computers. Similarly, on the user side, prices of personal computers have dropped substantially. Even portable devices are relatively inexpensive. But the biggest change is the use of virtualization to create servers
as a service in the cloud.
Centralization
The biggest advantage of centralized IS hardware is that it is easier to share hardware, software, and data with multiple users. If a company installs an expensive printer in one user’s office, it will be difficult for other users to get access to the printer. On the other hand, with only one central computer, all of the hardware, software, and data will be located in one place. All users can be given equal access to these facilities. By keeping all hardware, software, and personnel in one location, it is easier to avoid duplication and keep costs down. Along the same lines, centralized hardware also makes it easier to control user access to the information system. By storing all data on one machine, it is easy to monitor usage of the data. In a sense, all user access to data must first be approved by the MIS department. Any data alteration or transfer is much easier to control if it takes place on one machine.
Centralized purchasing can also be used to save money. It is easier to standardize equipment if it is all located in one place. It is generally possible to obtain discounts from vendors by buying larger quantities. Centralized purchases also make it easier to keep track of the amount of money spent on information technology. When user departments are responsible for all IT purchases, the lack of centralized control can lead to duplication of hardware.
Decentralization
Decentralization of hardware carries its own advantages. First, there is less chance of a total breakdown. If your computer breaks, everyone else can continue working. You might even be able to borrow someone else’s machine. The 1980s saw the adoption of lower-priced mid-range computers. As hardware prices continued to drop, managers adopted personal computers to handle new tasks. The lure of total control over powerful, low-cost computers led users to adopt them and use them for many computing tasks—particularly decision support analyses. During that time, spending on all computers, including servers, increased. But the growth of PC-spending vastly exceeded all others.
The PC technology eventually spread into servers, driving down their costs as well. Today, the differences between servers and personal computers are subtle; but some vendors such as IBM, Hewlett-Packard, and Oracle/Sun sell specialty servers. Many of these servers are used for Web based applications, so they do represent a return to centralization of hardware in some areas.
Earlier, in 2006, the value of money spent on laptops exceeded that spent on desktops (because laptops were more expensive). By 2009, the number of laptops sold exceededbthat of desktops. More interestingly, almost as many tablets were sold as desktops. And almost 10 percent of the market was for Google’s Chromebook, which looks like a notebook but only runs a browser. All applications and data are stored on cloud-based servers—notably Google Docs. This trend to mobile computing is important because it shows that people want to have their computers and data, or at least Web access, with them at all times. It also reflects the increasing performance and shrinking size of hardware components.
Another aspect of decentralization is that users can obtain personalized equipment. Perhaps a financial analyst needs an extremely fast machine to process complex equations. Or maybe a marketing representative needs a portable computer to collect data from clients. An advertising specialist could use high-resolution graphics to help design promotions. In each case, the company saves money by buying each user exactly what he or she needs and not forcing everyone to use one
standardized product.
Software
Wherever there is hardware, it is also necessary to provide software. Nonetheless, it is possible to centralize some aspects of software even with decentralized computers. The goal is to capture the advantages of both methods at the same time. Data files are similar to software files, but some additional features need to be considered when choosing where to store the data.
Software Centralization
If software applications are standardized and purchased centrally, it is possible to negotiate lower prices from software vendors. Besides, if everyone uses the same basic software, fewer compatibility problems arise and it is easy for users to exchange data with coworkers. Similarly, upgrades, training, and assistance are much simpler if there are a limited number of packages to support. Imagine the time and effort involved if the company needs to upgrade different spreadsheets on 5,000 separate machines. Some companies have reported that by the time they managed to upgrade the entire company, an even newer version was released.
Cloud-based software highlights the benefits of centralized software. Ultimately, everyone runs a copy that is stored in one location. Updating the software simply requires replacing that one copy. Smart-phone apps are slightly different but similar in that one copy is stored in the central store. The phone device simply checks for updates and downloads any new releases. Automating the process helps ensure that everyone is running the newest version.
Software Decentralization
Forcing users to choose identical packages can lead to arguments between users and the MIS department. Many times users have different requirements or perhaps they simply have different preferences. If one group of users prefers the software that is different from the corporate standard, why should everyone in the company be forced to use the same tools? Cost becomes less of an issue as software prices drop. Support might be a problem, but major software packages today are similar.
Data incompatibilities often can be resolved with conversion software. To some extent, users should have the ability to customize their software to match their preferences. Today, most software packages enable users to choose colors, mouse or keyboard responsiveness, and locations to store personal files. If this software is moved to a centralized environment, you have to be careful to preserve this ability.
One complication with enabling users to choose different software is that it can be difficult to determine the configurations of each machine. If a user has a problem, the MIS support person needs to know what software is installed on the machine. When installing new hardware and software, the support team needs to know what software exists on each target machine. Managers also need to track software usage when they purchase upgrades and to verify compliance with software licenses. Several software tools exist to help the MIS department track software usage and report on the configuration of each computer. A small file is installed on each computer that reports on the software, hardware, and configuration of each machine.