540 Part IV • The Information Management System
1 .Equipment Location.Clearly, most organizations today operate in a distributed computing environment. However,
the physical location of the hardware on a network can be a critical issue from cost, control, and security standpoints.
Physically distributing equipment that supports enterprise systems may create additional costs for managing the hard-
ware and safeguarding data (e.g., many computers and telecommunications switches benefit significantly from being
housed in a secure, environmentally controlled location), as well as extra costs for specialized personnel. Nevertheless,
there may also be good reasons for locating servers near plant or division managers, rather than in a corporate or
regional data center many miles away. Likewise, some countries may be better hosts than others for the location of
complex data centers.
2 .Client/Server Allocations.Policies on the design and role of IT servers and workstations are also driven by not only cost
concerns, but also convenience and security trade-offs. Questions that need to be addressed include: What is the most
appropriate location of each type of computing work—at the workstation or at a central server or at a remote hardware
resource? Which workstations should have independent intelligence and which should be a network device slaved to
some central server? Should telecommunications, such as with voice over the Internet protocol (VoIP), and computer
components of the workstation be physically integrated? Should video conferencing capability be integrated into the
manager workstation, on a local area network, or at some departmental server? What level of access should a given client
have to outside resources on the Internet?
3 .Operating System Standards.Some vendors of technology hardware still offer a proprietary operating system, although
more commonality exists now than in the past. How many and which operating systems will the organization support? In
particular, will the organization support an “open source” operating system such as Linux, for servers and/or clients? Each
different operating system creates more difficulty in sustaining a seamless network, and support costs increase rapidly as
new operating systems are added. Confining the company to one operating system, however, reduces bargaining power,
limits access to the best software, and makes the organization more dependent on the fortunes (and security practices) of a
particular vendor.
4 .Network Redundancy.Because organizations are so dependent today on their networks, many business managers
want full redundancy of the key nodes and paths in the IT network. Yet full redundancy can be very expensive. How
much redundancy should there be in the design of the network? Should there be full redundancy only for major nodes
and high-volume pathways? The cost for full path redundancy can be very expensive because there must be at least two
different paths to every node in the network from every other node. Likewise, “hot” backup sites that allow failed crit-
ical nodes to return to operation quickly are also expensive. The lack of redundancy, however, can be very expensive
in terms of lost user time if the network or a critical node is not available for some period. The needs of business man-
agers for specific types of applications are key inputs to trade-off decisions between the cost of downtime and the cost
to provide continuous access.
5 .Bandwidth Capacity.What bandwidth, or transmission capacity, should be provided between hardware nodes in the net-
work? The decision is, of course, dependent on the applications to be used. Image and graphical applications require much
greater transmission rates for effective use than do text-only applications. Content-rich applications are growing rapidly on
the Internet. Should every client on the network have broadband connectivity? Specifications about the desired technical
infrastructure to meet the vision for information use are critical to help drive bandwidth investments.
6 .Network Response Time.In many organizations, hundreds of users are simultaneously interacting with the network, and
each of them is directly affected by the response time of the system—the delay between when the return key is pressed
and when the response from the system appears on the screen. If this delay is reasonably short, and consistent, users are
generally satisfied. If the delay is perceived by the user as excessively long—three or four seconds when one is used to
subsecond responses—it can be very frustrating and significantly hamper productivity. Yet the costs needed to reduce
response delays tend to increase exponentially at some level, so input from business managers is again critical in making
such decisions.
7 .Security, Privacy, and Network Access.If steps are taken to make the network and its nodes more secure, quite often the
result is to reduce ease of access for users of the network. Organizations should make an explicit decision to operate
somewhere along the spectrum between maximum ease of access and maximum security, based on an up-to-date under-
standing of the legal and regulatory environments for their business. In addition, how ubiquitous should access be to spe-
cific networks and data? Should all managers in the organization have essentially the same access rights, or is there a
need to provide different levels of access support—including remote access—to specific employees? To what extent
should employees have unrestricted Internet access? At many firms, viewing the results of athletic events (or even the
events themselves) is permitted from the desktop, but others prohibit such access.FIGURE 13.3 Technology Trade-off Decisions