Thin Client Vs. Fat Client & Network Design Considerations

Thin Client vs. Fat Client & Network Design Considerations

Introduction

Network computing was created in an effort to allow users of a computer application to share data more easily than using stand alone computers. Clients on a client/server network store their application data on a central server. There are two categories of clients on a network. They were originally categorized by their hardware design, but today clients are categorized by the software application design and where the bulk of the processing is done and where the bulk of the application software is stored.

A thin client, sometimes called a lean client, is a low-cost, centrally-managed computer. The term derives from the fact that small computers in networks tend to be clients and not servers. Since the idea is to limit the capabilities of these computers to only essential applications, they tend to be purchased and remain "thin" in terms of the client applications they include. The majority of the processing required for applications is done on a server. An example might be accessing a banking software application via telnet or visiting H&R Blocks website to do your taxes online.

The opposite of a thin client network application design is a fat client design.

In a client/server architecture, a fat client performs the bulk of the data processing operations. The data itself is still stored on the server. Although the term usually refers to software, it can also apply to a network computer that has relatively strong processing abilities. An example might be QuickBooks for Windows or Microsoft Word.

Part A

Centralized computing has been through an obvious cycle in the last 30 years. During the 70s computer hardware was very expensive. It made sense at that time to invest in a server and have dummy terminals to do the computing. The dummy terminals consisted of inexpensive hardware, while the server tended to be expensive but low in numbers. During the late 80s and 90s computer desktops began to drop in price. Networks were growing in numbers and it began to be difficult to keep all the processing and application software on the server. Furthermore operation managers liked the ability of a fat client to stay working, while the network was down. Popular software languages and tools such as visual basic and Visual studio were created, which spurred fat client design.

During the 90s networks and servers became more reliable. An extremely reliable server with redundant parts could be purchased for under $10,000. Advances in network design from token ring to Ethernet allowed for more reliable network communication. The internet and long range network communications become prevalent as TCP/IP proved itself as a reliable network protocol.

Other developments on the software side which have made centralized computing possible again are numerous. In the 1970s computing consisted of basically alpha numeric digits. During the late 1980s and 1990s applications with graphics came into being. Graphical applications did not lend themselves to thin clients, so a bulk of the application resided on the workstation. During the 1990s a thin client language called HTML was popularized by a client, which could display graphical HTML. The client was Mosiac and it was invented by Marc Andreesen from the University of Illinois. Especially important was the inclusion of the "image" tag which allowed it to include images on web pages. Earlier HTML browsers allowed the viewing of pictures, but only as separate files. Mosaic made it possible for images and text to appear on the same page.

New add-ons to Microsoft Windows are also making thin clients revive again. A product made by Citrix Systems actually simulates a Microsoft Windows desktop from a thin 400KB client. The Citrix product does have server capacity problems when sending multimedia over a network, but basic word processing

...