Calculating Operating Costs | Microsoft Windows Server 2003 Terminal Services

Why should a company trust in a technology such as terminal server technology? What are the problems with conventional PC environments? Before PCs were networked, they were the most cost-efficient computers. Through networking and identification of hidden maintenance and training costs, this point of view has changed drastically. The term total cost of ownership (TCO) is used to describe the total cost incurred by a system during its entire life.

In the past, the sheer number of Microsoft Windows concepts and applications meant that networking greatly improved computer systems. On the other hand, however, it was and still is challenging (and may be expensive) to introduce and maintain networked systems. This is true despite the fact that corporate administration tools such as Microsoft Systems Management Server were developed at great expense and effort. Unfortunately, these tools are not always the optimum choice, particularly for heterogeneous network environments.

In these cases, terminal servers can be a great alternative, especially because the existing infrastructure (network, server, clients ) does not necessarily have to be replaced . Terminal servers are a central instance that can be managed comparatively easily, even if different clients need to be integrated. Nevertheless, before terminal servers are accepted for some business units or even for the entire company, the profitability of this solution still needs to be justified in detail.

Cost Categories

The following section deals with individual cost items and how to categorize them as they relate to terminal servers. The exact costs depend on the target environment and will not be stated in absolute figures, but rather in qualitative terms.

Investment Costs

Investments are those costs that need to be incurred for the initial purchase of hardware and software. They are mainly made up of central components, infrastructure, decentralized components , hiring staff, and costs of location:

Central components Hardware, peripheral devices, operating systems, user authentication software, file servers, central printing, e-mail servers, database servers, Web servers, CD servers, and backup tasks .
Infrastructure Network, including cables, routers, switches, and firewalls.
Decentralized elements Hardware, peripherals, operating systems, and software for clients that the end users will work on.
Staff Costs incurred in finding and hiring suitable personnel.
Location Air conditioning, uninterruptible power supply, security, and fire protection.

When terminal servers come into play, investment in decentralized elements and infrastructure usually decreases. It is not necessary to modernize an existing network including clients to achieve improved application performance. Only the connection between the servers should be as fast as possible, which might require some investment in suitable network components.

Expenses for central elements will increase because terminal servers require powerful hardware. As this hardware requires a lot of space, fitting out the location will be quite expensive. Even though less staff will be necessary to manage the clients, more personnel will be needed for administration. However, the larger an installation, the lower the administration cost per workstation.

When a conventional environment reaches a certain size , it requires central administration and installation tools. The cost of these tools must be considered when comparing a conventional environment with a terminal server environment. However, central installation assistants or cloning processes will be needed for very large terminal server environments as well.

In general, the investment costs incurred by a conventional PC environment and those incurred by terminal servers are basically the same, or are perhaps, if anything, a little more favorable for terminal servers. Terminal server environments do not allow for major savings yet. Nevertheless, it is important to see that the costs of client/server hardware often represent less than 15 percent of the total cost incurred in providing applications.

Fixed Current Costs

Fixed current costs regularly arise for the environment s general operation. These costs include administrative tasks, personnel costs, training activities, end user support, siting costs, expendables, maintenance agreements, and frequent hardware disposal:

Administrative tasks User administration, installation and administration of applications, performance management, security, licensing and data administration (availability, security, and virus protection).
Personnel costs Salaries, rent, administration, and expenses.
Training activities Administrator and end user training, training material, and management information.
End user support Troubleshooting, hotline, helpdesk, and task logs.
Siting costs Rent for server rooms, power and communication charges for clients, server, and network.
Expendable Procurement, storage, and management of data carriers , literature, and spare parts .
Maintenance agreements Frequent hardware component maintenance, and software support agreements.
Disposal Removing and disposing of client and server hardware at the end of its useful life.

Administrative tasks on terminal servers are highly centralized and therefore require less effort, which, of course, results in significantly lower personnel costs. This fact can be realized especially when working with low-maintenance clients.

In regard to training activities, there are almost no differences between terminal server and conventional environments. However, the difference is extreme when it comes to end user support. A homogeneous and central environment simply does not require as much support for configuration and applications.

Siting costs are higher with servers; thin clients require considerably less space and power. Compared to a conventional PC client environment, expenses remain about the same. The same goes for expendables. Even though fewer variations of material are needed, the amount required makes up for any difference.

Basically, terminal server maintenance agreements are quite simple because they encompass only a few components. However, depending on the number of client types, the bandwidth of their individual maintenance agreements can still be fairly comprehensive. On the other hand, the hardware replacement is less often because both clients and servers work a lot longer on server farms than before.

All in all, current fixed costs are considerably decreased when terminal servers are used, thus presenting obvious savings potential.

Note

Some fixed current costs stay the same while administration costs, support, and disposal costs are reduced. These are fixed current costs because you have to pay the system administration staff and the support staff, even if there is less work in system administration and support. Typically, the income of a supporter is not a variable cost for the company because they have to pay his salary even if he has to answer fewer support calls. You cannot fire half a support person. You can reduce your support personnel, and thus the related fixed current costs, if you find out that your average number of support calls remain on a reduced level after the introduction of terminal servers.

Variable Current Costs

Because variable current costs are exceptions, they are not easy to calculate. They are incurred by unplanned downtime, necessary changes to the runtime environment, and user support.

Unplanned downtime System failures, work interruption, and failed cases of user self-help .
Changes to the runtime environment Evaluation of new software, piloting, project management, software distribution, and changes to application configurations.
User support Help for new users and after introduction of new software versions.

Central, well-managed servers reduce downtime in terminal server environments. Even mistakes caused by self-help in conventional environments are no longer important because users have no opportunity to interfere with terminal servers.

In a production terminal server environment, software components are usually not changed without prior tests. However, it is relatively easy to integrate a test system with new operating system or application software components into the existing environment and let selected users test it. This allows risk-free installation of new software versions. Downtimes are reduced to a minimum through new operating system versions and application configurations. One definite precondition for this is that a test environment be available where the new runtime parameters can be determined prior to general installation.

User-support expenses are also significantly reduced in terminal server environments. All terminal server users work with the applications selected by the administrators, and users no longer have the opportunity to manipulate applications. Another advantage is remote control. It enables the administrator to take over control of a user session and guide the user through solving the problem.

All in all, terminal server environments incur considerably fewer variable current costs than conventional environments. Nevertheless, this calculation strongly depends on the quality of both the server installation and the organization of the administrators in charge.

Application Costs

The cost qualification model described above focuses on hardware. If networks and computer systems already exist in the target environment, other criteria play a more important role, especially the costs incurred by owning and providing applications for use on heterogeneous networks and clients. For this reason, the preceding cost analysis is adjusted accordingly and thus becomes total cost of application ownership (TCA).

Cost Structure

These four factors influence costs relating to applications:

Where is the location of the application executable? If an application executable is located on a client rather than on a central server, costs for administration, installation, and configuration are higher.
Where is the application executed? If an application runs on the client, the client needs the corresponding attributes. This might even imply a physical exchange or extension of client hardware when a new application version is installed. This is even true for applications that run partly on the client and partly on the server, or whose executable files are first loaded by the server. Executing the application on the server allows central adjustment of hardware for many users.
Where is the data saved? User data saved on the client is a risk when it comes to security and availability. For this reason, user data should be located on a central server.
What is the user location and type of connection? Distributed user locations and the bandwidth of the network connection impact costs depending on the different models. Support costs and the necessary infrastructure play an important role, too.

The terminal server model is optimal regarding all four factors: applications are saved and executed on the server. Data is saved on a server. Clients are no longer relevant and can be located anywhere . The network connection requires a relatively low and easy-to-provide bandwidth.

However, the disadvantages of a terminal server environment also need to be considered. They include reduced flexibility for the user, extreme dependency on a network that functions continuously, and problems with processing local data (for example, printing or scanning documents).

Conclusion

Using terminal servers reduces the overall costs for providing applications for the entire company. It is possible to use the existing infrastructure without subjecting it to substantial changes. Framework conditions include powerful server components and a minimum network bandwidth of 10 to 20 kilobits per second for each client. Furthermore, well-trained administrators are the basis for stable and thus cost-effective operation of this type of environment.