Just as for other complex technologies, the introduction of terminal servers fails in many companies. Frequently, it is not technical issues but rather the organizational issues that result in unsatisfactory solutions. Still, users do not care about the reason for failure—they care about the impact on their everyday work. If production work becomes impossible, system administrators and management will be affected, too. Acceptance can drop to zero, even for a technology with a great concept. This failure can even cast doubt on an entire corporate strategy.
To help avoid organizational or conceptual mistakes, we will briefly examine project planning and, the selection of a system integrator plus we will discuss establishing guidelines to avoid mistakes.
So what is the difference between a successful and a failed terminal server project? The technology itself works, as has been proved, just not under all circumstances. No one expects a sports car to be perfectly suited to all tasks in construction or manufacturing. A four-wheel-drive truck would be a lot better for these tasks, although their basic structure (a vehicle with four tires and a steering wheel) doesn’t differ much from a sports car. The same is true for terminal servers. They are better suited for some tasks than for others. Finding out which areas of a company terminal servers can be effectively used and implementing successfully, from a technical point of view is the objective of a terminal server project.
Terminal server projects are typically divided into five phases. Either an internal or an external service provider normally handles the implementation of such a project.
Pre-study, including test In this phase, the service provider analyzes the target environment, the IT terminal server strategy is defined, required core applications are defined, the total cost of ownership (calculating the total costs) is determined for different operation models, and the test environment is installed. The test environment outlines the basic feasibility of any future project. After all pre-study activities have been completed, you will have an analysis document describing the actual situation, a rough concept describing the target situation, an estimate of all costs, and presentation material for all business units involved. Do not forget that the results of the pre-study might lead to the conclusion that terminal servers are not recommended for a particular implementation. If desired, a corporate concept can be drawn up during the pre-studies for a future production environment based on terminal servers.
Planning phase After the successful pre-study, a complete technical concept and set of specifications are drawn up. The technical concept describes the technical procedures, such as selecting hardware and operating system components, how to integrate the terminal servers in the existing infrastructure, and installing all required applications. The specifications explain the formal processes during the project, for example, decision making, time schedules, measurable results and milestones, approval procedures and criteria, exit scenarios, and escalation mechanisms. Both documents are absolutely essential for the successful conclusion of the entire project. Therefore, changes to the technical concept or specifications during the project should always be checked and approved by all project partners involved. During the planning phase, the customer will receive some initial training, if this has not been done before, to establish a common basis for know-how and terminology.
Pilot phase In this phase, there will be a so-called “clean room installation” with no direct connection to the existing production target network environment. This procedure helps prove the basic functionality of the project concept. It could even require separate network components and individual domain controllers, file servers, print servers, database servers, and so on. Only if very extensive third systems (for example, mainframe computers or comprehensive databases) need to be connected should compromises to the clean room installation be made. As a result of this phase, there will be documented tests in line with the specifications, which will help determine reference parameters.
Field-test phase This project section deals with transitioning and integrating the clean room installation into the existing production environment. The focus lies on establishing connections to existing user accounts, backend servers, and legacy systems. Furthermore, it is often necessary to adapt logon scripts, profiles, and policies. Sometimes, complete strategic corporate processes for network, user, and server administration are questioned during this phase. Test users and approval procedures ensure that the installation is qualified. The last part of this phase sees the composition of approval protocols, project documentation, and, if applicable, complete operational concepts.
Production phase In this phase, the end users start to work with the system productively. In the beginning, the service provider will offer on-site assistance. If this results in essential changes in the users’ work routines, there will be acceptance difficulties for some time. Continuously informing end users and setting up a special help desk will help make the people involved more willing to cooperate. Furthermore, operations and maintenance concepts will start to be effective and assure uninterrupted operation of the terminal servers.
Each of the above phases is concluded with a milestone. This milestone represents the end of the phase and determines whether the project will go on or be aborted. All criteria regarding the following step need to be set in the specifications from at least the planning phase on.
If weaknesses in the project concept become evident during the pilot or field- test phase, it will be necessary to backtrack to the planning phase. Even though this might lead to deadline attainment problems, this procedure will prevent a system that does not work properly from being transitioned to the production phase. Once in the production phase, conceptual mistakes can no longer be corrected and usually lead to catastrophic results.
Mistakes happen in every project! The earlier they are identified and solved, the less problematic they are. For this reason, efficient escalation management is essential for the success of a project. Determining corresponding procedures and defining project abortion criteria in the specifications is therefore essential.
What is the difference between a TV and a toaster on one hand, and a corporate server and a mainframe computer on the other? The former category can be used by untrained persons without any technical introduction. The latter category is based on technology that needs to be explained and should therefore be set up and operated only by specialists. These specialists must have undergone extensive training and acquired a great deal of experience.
Naturally, the world is not black and white. There are so many nuances, and this also goes for the technological categories described earlier. Skillful operation of a sophisticated VCR or motor vehicle requires comprehensive studying of manuals or attending special practical training. After completing their studies and/or training, normal people can use these devices. Due to improved user interfaces, setting up and using a PC is becoming common knowledge. However, in large companies, this can lead to a fatal mixture: untrained staff operating seemingly easy-to-use but at the same time company-critical technology that actually needs to be explained.
The “self-made administrator problem” is especially obvious in corporate networks with terminal servers. Because of its graphical administration tools, an application server configured as a terminal server at first glance seems to be easy to operate. However, this computer as a host that has many terminals connected to it requires relevant in-depth knowledge. If several terminal servers are combined for load balancing to provide an environment for multiple users, the corresponding administration task is anything but trivial. For this reason, system integrators assist with installing, training administrators, and operating a terminal server environment.
So what services should be considered for the introduction and production use of this type of environment? What does a good system integrator offer? Usually, five selection criteria are essential: consulting, installation, support during operation, training sessions, and support tools combined with effective project management. Let us take a closer look at these criteria.
Consulting Before project start, the target environment needs to be analyzed in detail, involving not only the existing and planned technology, but also the potential acceptance of the system by management, users, and administrators. Additionally, a plausible concept plays a major role. The overall objective is to draw up technical specifications to ensure proper project process of the terminal server environment installation. Tests can also be part of this pre-phase of the project. However, a suitable reference environment (for example, a test lab) needs to be available to conduct these tests.
Installation Installing a complex terminal server environment is the most critical part of a project and requires extensive experience and conscientiousness. Specially trained system engineers will help avoid mistakes by keeping strictly to the specifications and the time schedule. An experienced project manager should support them. The project manager must always be informed of the latest developments and check installation progress on site frequently. At the end, there will be an approval protocol and follow-up support in case of problems that might occur. The customer is presented with a well-functioning and well-documented production environment.
Support during operation After a new production terminal server environment is installed, technical problems may occur. They are usually caused by certain system software weaknesses or insecurity on the part of the administrators in charge. For this reason, each system integrator should operate a proper customer support hotline for a certain time after the actual project is finalized. Several invoicing methods might be implemented for this service. In addition to calling phone numbers subject to a charge (for example, 900 numbers), so- called call-IDs may be purchased and used as coupons for calling the hotline. Another type of hotline support is the call-by-call method, meaning that one technical question can be asked for a certain charge. This method has the advantage of involving no further contractual obligations and requiring that the answer must be presented within a certain time frame. Large companies in particular, however, prefer paying an annual fee that allows them as many technical inquiries as they need (premier support). The relevant modalities are all agreed upon in a contract. Of course, support inquiries should be possible not by phone only, but also by e-mail or fax. All inquires should be processed by the system integrator at a central helpdesk. All helpdesk staff have access to a database-supported system that archives all problems and documents all solution approaches and answers (tracking system). This method also simplifies time and content control by a superior escalation instance. Frequently asked questions should be made available to the customer on the Internet or on a document or data carrier that the users can subscribe to.
Training sessions A system integrator who offers the installation of a technology as complex as the terminal server should also be able to provide the customer with the required training. Proper training makes handing over the installed production environment to the customer and possible cooperation with a help desk much easier for everyone involved. Staff training with a fixed date, location, and content should come first. Another option is exclusive training for selected participants. Special training with individualized content would round off this service. Workshops including extensive practical exercises generally result in the participants’ ability to understand the technology and safely work with it.
Support tools A system integrator can usually present a number of reference projects relating to the required technology. This implies that the system integrator has built up an internal knowledge base that will help deal with all aspects of the issues described above, including all activities before project start (pre-sales, consulting), during the project (installation), and after the project (post-sales, support). Furthermore, the system integrator develops special products, tools, mechanisms, or procedures before or during the project, thus ensuring quick and smooth work in follow-up projects. This so-called “toolbox” is similar to the toolbox of a skilled craftsman. It usually consists of automated analysis and measurement tools, support tools for system installation, configuration aids, and defined research options for troubleshooting.
Implementation of the criteria listed earlier often requires highly specialized personnel to reach the targets set for the installation of terminal servers. Before you decide on your terminal server project provider, ask all potential system integrators about their products and general conditions relating to the preceding issues and about the availability of the required personnel. Only when suitable staff can cover all relevant areas is the probability of project success high enough to warrant implementation. Also, it is always legitimate to ask a system integrator for on-topic publications and strategic partnerships with the producers of the system software.
When drawing up a server-based computing concept for your company, always focus on the human aspect. This includes the management, administrators, and, of course, the users. If these three groups do not accept the strategy, even the best technical concept can only lead to unsatisfactory results. If these people do not sufficiently accept the solution offered, achieving the discipline initially required for the use of Windows terminal servers will be difficult. System stability will never be reached if the exclusive administration of the Windows terminal server by trained administrators and the terminal server’s explicit treatment as a vital corporate component are not ensured.