Over the last decade, total spending on desktop personal computing has escalated into the hundreds of billions of dollars, yet the cost of owning a client-server computing system goes far beyond the initial expenditure for hardware and software.
It is estimated that over two trillion dollars has been spent for support, administration, and end-user operations such as training, help desk, network management, upgrades, and disposal. In this age of Business Process Redesign (BPR) and Activity Based Costing (ABC), many organizations have focused their efforts on determining the real costs of distributed computing. It has become imperative for organizations to know the total cost of owning, managing, and supporting a client/server PC.
This White Paper examines a number of components of the Total Cost of Ownership (TCO) equation, and how TCO can be effectively, and successfully, applied to reduce operating expenditures. It explores the four major cost components of client-server computing, examines the cost differences between desktop computing and network computing, and discusses cost containment/reduction strategies.
Since the mid 1980's, Total Cost of Ownership (TCO) models have become popular, in an effort to help understand the complexity of hard and soft dollar costs associated with owning and operating client/server-computing equipment. TCO has become a major focus in the PC industry since less complex and cheaper computing platforms like Intranet environments and associated network computers (NC's) may be the next evolution of distributed client server technology.
This section examines the four major cost components of client-server computing, the TCO concept, and three potential TCO models.
To fully analyze distributed client server technology and its overall effect on cost, the cost of ownership must first be defined. Client-server cost of ownership is defined as the total cost of computing over a specified period of time and includes all the elements in delivering the information users want when it is needed.
Client-server computing costs consist of the following four major categories:
The concept of "total cost of ownership" has been a highly charged debate for over a decade among information technology research firms and vendors. A basic model provides a comprehensive life-cycle approach to measure the costs associated with owning and operating a desktop PC and to evaluate end-user computing investment decisions. Such a model is platform defined (e.g., Microsoft Windows 98, Microsoft NT/ 2000), analytically driven to assess costs such as several soft, hidden cost components that are difficult to quantify, and show how costs will change over time, typically a three-year period (the approximate life-cycle of a desktop PC).
Additional models are geared to environments that may include mainframe computers, geographically dispersed computing architectures, communications-intensive network environments, point-of-sale systems, campus-wide debit-systems, and so forth. Depending on the business environment, the TCO models may be combined into a single overall view of the technical environment. However, for those technologically diverse environments where different life-cycle phases are present, the TCO model may be partitioned into sections to account to variable lapsed periods, such as where PCs may have a period of 3 years, but a communications relay may have 5 years.
This section discusses two-tiered versus three-tiered architectures, explores the Network Computer evolution and compares the cost differences between the desktop computer and the network computer.
Though client/server architecture can be very complex, there are typically two opposing architectures: the "devolved" two-tier (i.e., desktop PC) and the "centralist" three-tier (i.e., network computer). The "centralist" view is represented by the terminal-oriented systems of the mainframe environment, while the distributed client-server model represents the "devolved" view. Approximately 95% of existing UNIX and PC client server architecture are "devolved" two-tier systems utilizing graphical user interfaces.
The TCO revolution has led many PC vendors to deploy "thin-client" computing architectures, such as network computers (NC) and Intranet applications, as the cost solution for similar, cheaper computing platforms. The network computer is simply a non-programmable terminal, which utilities on-board CPU chips and random access memory acting with a terminal emulation interface. New computing architectures such as "thin-clients" are being deployed to minimize the cost of supporting client PC's and realize the benefits of open systems.
Depending on the type of client server architecture, the total cost of ownership can vary significantly. Generally, the costs associated with maintaining a PC are higher than those of a network computer (NC), since the NC provides fewer internal components that need to be managed at the desktop level. In fact, the NC, as an example, enables more effective use of centralized resources to maintain and update software, and reduces the amount of end-user introduced variables that can increased required support levels.
Network computers have fewer hardware components and the software is stored on a central server that virtually eliminates the need to troubleshoot individual desktop computers. In addition, the initial acquisition costs for a networked computer versus a PC can be substantially different, particularly when purchased in quantity. Significant cost savings can be obtained through the deployment of network computers, particularly in the administrative support and executive levels of organizations. However, an organization must determine if a networked computer with its restricted end-user control suits the business needs.
This section discusses strategies for improving a wide variety of functions related to owning, operating and managing PC/LAN technology. Information concerning the use of standards, improved coordination among support staff, electronic software distribution tools, automated inventory and software metering tools, remote access maintenance, virus detection and repair, data backup and restoration, and JITT (Just-in-time-training) tools are discussed below.
The use of standards is critical for managing the total cost of ownership. Standards play an important role by establishing organizational clarity on technology, business processes, and procedures. Technology standards include not only the technology itself, but also how that technology is configured, managed and supported. Standards must also be applied to the business processes and procedures utilized in managing an organization's networking environment, particularly if that organization utilizes remote sites, standards and guidelines are crucial to establish a productive IT environment at a reasonable cost.
Standards must be applied where they make the most sense and an organization does not have I00% conformance to a standard. The ultimate goal is organizational effectiveness and not universal conformance to a standard. Standards enable an organization to take better advantage of volume pricing, decrease acquisition processing costs, decrease support and training costs, and improve the organization's ability to share data and applications. Standardization needs to be considered on five fronts - hardware, user interface, applications, infrastructure and processes.
Many distributed computing environments are developed with multi-layered support structures and lack an overall coordination of effort. For example, those that support telecommunication equipment installation typically do not consult with those installing the desktop computer for a user. A lack of coordination results in severe levels of duplicated services and limits an organization's ability to implement standards and apply system management tools. An organization's Corporate Information Systems Department needs to define the boundaries and roles of PC, telecommunications and user support groups. The important goal is the clear delineation of responsibilities for supporting the desktop-computing environment. Organizations need to realize that the PC asset base needs to be administered and managed, and not ignored. Organizations that have implemented centralized management and operation of servers are experiencing measurably lower operations costs.
It is estimated that software product updating typically accounts for up to 55% of a desktop system's total cost, whereas the initial purchase and support account for less than 45%. The task of updating software products is manual and cumbersome and many organizations have shifted the updating task to the end user community. Electronic software distribution (ESD) packages offer a cost-effective solution for automating the distribution and installation process. In addition, ESD also provides capacity checking, auditing and management reports.
A current inventory of hardware and software assets can assist in eliminating duplicative maintenance, improve asset disposal, eliminate duplicate software licensing fees, and improve disaster recovery planning to reduce support costs. Manual tracking of the inventory is a difficult job requiring physical inspection of each workstation to determine the installed devices and scan the hard disk drive. Automated inventory systems significantly accelerate the effort and are less costly than traditional manual inspections.
Software-metering tools allow an organization to easily identify the software used by the networked PC's and assist in determining the correct number of software licenses. Metering tools can also assist support staff in determining concurrent use of software applications when the software is distributed to all users and licenses are determined according to the number of executions. Metering systems play an important role in security and software distribution by enforcing limited-use and site licenses for particular software packages. An effective inventory control system should effectively track both hardware and software assets.
When systems are distributed it is extremely difficult to perform routine maintenance and inspection of the system, conduct enterprise-wide virus detection and repair, and perform data backup and restoration. In a networked environment, the vast majority of maintenance functions can be performed through automated remote-access tools that allow information technology staff to open a session on other systems in the network, run programs and diagnostics, and inspect and alter files.
Network virus detection programs allow information technology support staff to execute virus scans either on demand or on a scheduled basis and initiate automatic repair or system isolation.
The backup process wastes both time and equipment when it is performed by the end user and may also lead to lost, misplaced and stolen backup media. An organization should conduct both scheduled full and incremental backups for data. In addition, an organization should design their LAN systems to include appropriate bandwidth and storage capacity for more efficient on-line storage utilization. It is critical to reduce the support costs associated with routine maintenance and virus detection and repair.
Traditional classroom training for software applications continues to be progressively less practical relative to the overall cost. Just-In-Time-Training (JITT) tools provide on-demand training to users through computer-based training programs. JITT enables self-paced learning and when it is integrated into the application, it enables the user to interactively learn a function and apply the function to a specific task. Finally, JITT reduces the time and cost of classroom training and learning labs. JITT is a powerful learning tool that reduces the cost of formal training, informal training and troubleshooting.
This White Paper has examined the major cost components of client-server computing, discussed to TCO modeling, discussed the cost differences between desktop computing and network computing, and how to leverage cost reduction strategies. TCO is a valuable too that can offer an organization benefits in multiple areas: a tool to measure costs accurately, an approach to gather technology resource data in a consistent, coherent, useable manner, and an approach to create a cost-sensitive baseline on which future analyses can be based.
Applying TCO in any organization must be based on the unique facets of that business, it's geographic dispersion, and a balance between non-quantifiable metrics, such as service and flexibility. However, those organizations that understand the overall cost complexities of technological evolution can objectively assess the cost efficiency and establish a baseline for measuring future cost saving initiatives.