Terminal server
From Wikipedia, the free encyclopedia
A terminal server is a device to connect multiple, possibly remote, input/output devices to a central processing unit.
Contents |
[edit] History
Historically, a terminal server was a device that attaches to serial RS-232 devices, such as green screen 'VT' terminals or serial printers, and transports this traffic via TCP/IP TELNET, SSH or other vendor-specific protocol (i.e. LAT)
Originally, the first terminal servers were devices providing a connection between a so-called "green screen" dumb terminal and a host computer via an Ethernet connection. These terminals were also referred to as 80x24 since they included 24 lines of text displayed up to 80 columns across. Digital Equipment Corporation's DECserver 100 (1985), 200 (1986) and 300 (1991) are early examples of this technology. (An earlier version of this product, known as the DECSA Terminal Server was actually a test-bed or proof-of-concept for using the proprietary Local Area Transport (LAT) protocol in commercial production networks.) With the introduction of inexpensive flash memory components, Digital's later DECserver 700 (1991) and 900 (1995) no longer shared with their earlier units the need to download their software from a 'load host' (usually a Digital VAX or Alpha) using Digital's proprietary MOP protocol. In fact, these later terminal server products now also included much larger flash memory and full support for the TELNET part of the TCP/IP protocol suite.
Many other companies entered the terminal server market with terminal servers pre-loaded with software fully compatible with LAT and TELNET. Some manufacturers also stated specifically that they had emulated Digital's command set for terminal server management. Besides retaining the ability of the older terminal servers to obtain their run-time code from a load host, most were able to bootstrap from on-board flash memory or from a floppy disc held in a drive in the terminal server. Some Xyplex terminal servers could act as load host for each other; one would hold the code on a PCMCIA flash card and serve it to another.
Starting in the mid-1990s, several manufacturers such as U.S. Robotics produced "modem terminal servers". Instead of having RS-232 ports, these would directly incorporate an analog modem. These devices were commonly used by Internet service providers to allow consumer dial-up. Modern versions interface to an ISDN PRI instead of having analog modem ports.
As of 2006 serial terminal servers are often used for connection to the console ports of Unix servers. This then allows system administrators to connect to the servers over the network. This is important for rebooting the system and for hardware debugging, where the operating system will not boot correctly.
[edit] Modern usage
Lately the term 'terminal server' can mean either a network access server or a server operating system that provides a graphical user interface (GUI) of a Windows or a Linux desktop to user terminals that don't have this capability themselves. Alternatively, the desktop is provided to a remote computer in order to enable teleworking. Other examples of remote desktop delivery for the purpose of remote access include Microsoft's RDP and the UNIX community's X11.
The terminal server client is referred to as a thin client. Protocols that the client and server use to communicate with each other are Remote Desktop Protocol, Citrix ICA and NX technology.
Connection to Terminal Server remains fast, because in a Terminal Server environment, only what is actually displayed on the screen needs to be sent across the Internet/WAN, and mouse and keyboard commands are the only things that need to be sent back to the server.
Since all of the processing and storage occurs at the server, the requirements for client devices are minimal. Client devices can be anything from a thin client (network computer) to a fully configured personal computer (thick client). The speed and power of the client computer matters very little since it is doing very little in the process.
[edit] Modern terminal servers
Modern terminal servers are used in many different ways. They are usually implemented with one terminal server which can emulate up to 40 or 50 machines simultaneously. The end-user uses a workstation (typically a rather inexpensive computer) to connect to the terminal server. The workstation typically acts as if it were running a full version of Linux or Windows (by using Terminal Services). This is advantageous for several reasons:
- One only needs to purchase a single very expensive terminal server rather than hundreds of expensive machines.
- Any single instance on the terminal server has access to whatever resources are not being used at the moment. This setup is ideal in a situation where the end-user needs to perform resource intensive tasks, but only intermittently. Since the terminal server may have very impressive specifications, this can help everyone have access to a powerful computer should they at some point need to perform such a task.
- The user's computer state is saved on the terminal server. Many systems are set up so that the end-user can login to their workstation from any location that has internet access (for example, by using Remote Desktop Connection).
- Lastly, it provides a centralized location where administrators can backup a single terminal server, rather than hundreds of individual machines.
[edit] See also
- Computer terminal
- Console server
- Linux Terminal Server Project (LTSP)
- Terminal Services (a.k.a. TSE)
- Virtual Network Computing (VNC)
- X Window System
- Out-of-band Management
- Reverse Telnet
[edit] External links
- Interix in a Multi-User Windows TSE Environment paper about the Unix multi-user model and MS-Windows NT TSE