Technology Reference Guide

	Technology Reference Guide 2. Hardware C. Application Servers

What is it?
In the traditional world of mainframes and minicomputers, the user had no local processing power. Using a terminal, the user could only request that the host system execute his required processes, and transmit the output when it was done. As the PC network evolved, it was based on the premise that the user would do their own processing, using their local CPU. They would access the network to gain access to passive resources such as file storage, printing services, etc. It soon became apparent that some applications, requiring high levels of computing power, or manipulation of large amounts of data, could not efficiently be processed by a local workstation, and would cause excessive traffic across the network. The first solution to this problem was to retain the host system architecture, and use software on the PC that would emulate a traditional terminal. This was inefficient in that it meant that the high costs associated with proprietary host architectures had to be maintained, and it wasted the processing potential inherent in the users workstation. These factors drove the development of the application server, and the modern client/server model, which implies sharing the computing tasks in an optimal way between the local workstation and a centralized processor dedicated to running the application.

An application server is a system, which may be a PC or a higher-end system, that is chosen for its suitability to run a specific application. A common example is a database server. This system typically runs a multi-user operating system such as UNIX, or a multitasking OS such as OS/2 which allows multiple service requests to be handled concurrently, much the same way as a host system does. In contrast to a host system, many application services are designed to split the processing load with a user workstation. For example, a database application may have a client portion residing on the local workstation that provides the user interface, structures the requests going to the server, and the data coming back. The server portion is designed to efficiently process these requests, and transmit the minimal amount of data required to fulfill the request. When the client workstation receives the data, it will structure and present it in the way desired by the user.

In terms of reliability and redundancy, an application server for a mission-critical application may employ redundancy techniques similar to a file server, such a multiple processors, redundant disks, dual power supplies, etc.

What does it do?
An application server offers the use of an application to authorized clients. It typically provides for the use of an application that is too compute-intensive or data-intensive to be handled by the users workstation. An application server often exists as a node on a LAN, that can be accessed by users with the appropriate client software and authorization. Novells solution involves the use of UNIXware application servers in the Netware environment, while the non-dedicated file services provided by OS/2 and Windows NTAS may allow application services to be provided, along with file services, by a single system. Typically, however, a dedicated system will be required to provide the capacity required by the application.

How is it used?
Application servers are an enabling technology for downsizing initiatives. Although office applications such as word processing and spreadsheets moved quickly into the realm of PC LANs, other applications requiring high levels of compute power or data manipulation remained chained to the mainframe. Thus, many companies remained committed to both their legacy mainframe environment, as well as investing in PC LAN technology. Through the use of application servers, the customer can often break the reliance on mainframes, and fully utilize the potential of an open microprocessor-based LAN environment.

Where do I get more information?
Contact your local CBV Office.