Previous | Table of Contents | Next |
Another way the browser can be fattened up is by bulking up on components written in ActiveX, Microsofts answer to Suns Java. Like Java, ActiveX is an object-oriented development tool that can be used to build such components as Excel spreadsheet interpreters and data entry programs. Functionally, the two development tools are headed for increasing levels of convergence.
For example, the Microsoft Java VM is an ActiveX control that allows Microsoft Internet Explorer 3.0 users to run Java applets. The control is installed as a component of Internet Explorer 3.0. The Java VM supports integration between other ActiveX controls and a Java applet. In addition, the Java VM understands the COM and can load COM classes and expose COM interfaces. This means that developers can write ActiveX controls using Java.
Browsers can also fatten up by running functions written in scripting languages like Netscapes JavaScript and Microsofts VBScript. VBScript is a Web-adapted subset of VBA, Microsofts standard Basic syntax. Both JavaScript and VBScript are used to manipulate HTML from objects like check boxes and radio buttons, as well as add pop-up windows, scroll bars, prompts, digital clocks, and simple animations to Web pages.
The important thing to remember about these tools is that the features they create rely on scripts that are embedded within the HTML document itself, initiating extensive local processing. Browsers are becoming universal clients, so much so that Microsofts next release of Windows 95 will even have the look and feel of a browser.
Most PCs today come bundled with a browser. Several vendors, including Microsoft, have endorsed the idea of offering a new breed of computer that relies on a browser as the graphical user interface, Java or ActiveX as the operating system, and servers for the applications. With Java and ActiveX, a network-centric computing solution is emerging that can potentially offer major improvements in simplicity, expense, security, and reliability vs. many of the enterprise computing environments in place today.
How fat the client is may be less important than how the code is delivered and executed on the client machine. Because Java applications originate at the server, clients only get the code when they need to run the application. If there are changes to the applications, they are made at the server. Programmers and network administrators do not have to worry about distributing all the changes to every client. The next time the client logs onto the server and accesses the application, it automatically gets the most current code. This method of delivering applications also reduces support costs.
Fat may be interpreted as how much the client application has to be fed in order to use it. For example, a locally installed emulator may have the same capabilities as a network-delivered, Java-based emulator, but there is more work to be done in installing and configuring the local emulator than the Java-based emulator that is delivered each time it is needed. The traditional emulator takes up local disk space whether it is being used or not. The Java-based emulator, in contrast, takes no local disk space.
ActiveX components are a cross between locally installed applications and network-delivered applications. They are not only sent to the client when initially needed, but are also installed on the local disk for future use. Local disk space is used even if the component was only used once and never used again. Updates are easy to get because they can be sent over the network when required. With Java, the component is sent each time it is needed unless it is already in the browsers cache. This makes Java components instantly updateable.
Because Java is platform-independent, a Java-based T27 emulator for Unisys hosts or a 3270 emulator for IBM hosts, for example, can run on any hardware or software architecture that supports the Java virtual machine. This includes Windows, Macintosh, and UNIX platforms as well as new network computers. Thus, any Java-enabled browser has access to legacy data and applications.
As with most issues, the answer is it depends. There is no right answer for all applications and all environments. Each has advantages and disadvantages, so it is necessary to do a cost/benefits analysis first. Even if a significant number of desktops must stay with the fat-client approach, there still may be enough incentive to move the others to the thin-client approach.
According to The Gartner Group (Stamford, CT), the annual cost of supporting fat clients Windows 95/NT, UNIX, OS/2, and Macintosh is about$11,900 per seat. Substantial savings could be realized for as many as 90%of an enterprises clients, with only 10% of users needing to continue with a fat client for processing-intensive applications. Thus, the support costs for moving from a fat-client to a thin-client architecture could be as much as $84.6 million annually for a company with 10,000 clients.
Intranets are becoming pervasive because they allow network users to easily access information through standard Web browsers and other World Wide Web technologies and tools to provide a simple, reliable, universal, and low-cost way to exchange information among enterprise network users. However, the resulting changes in network traffic patterns require upgrading the network infrastructure to improve performance and prevent slow network response times.
The corporate network may need to be upgraded to accommodate:
LAN switches traditionally operate at layer 2 of the OSI model, or the data link layer, providing high-performance segmentation for workgroup-based client/server networks. Routing operates at layer 3, or the network layer, providing broadcast controls, WAN access, and bandwidth management vital to intranets. Most networks do not contain sufficient routing resources to handle the new inter-subnetwork traffic demands of enterprise intranets.
The optimal solution intranet switching is to add layer 3 switching, the portion of routing functionality required to forward intranet information between subnetworks, to existing layer 2 switches. This solution allows network managers to cost-effectively upgrade the layer 3 performance in their networks. This is the approach being taken by new intranet switches and software upgrades to existing switches.
Previous | Table of Contents | Next |