Upgrading and Repairing Networks (5th Edition)

Does every user need a desktop computer in the first place? Years ago, when typewriters were the main "word processor," they weren't deployed on every desktop ”that's what administrative assistants were used for. Today, other applications, such as email, are essential for many users. However, that isn't always true for every employee. For managers who have administrative assistants, a single desktop computer can suffice in many cases. And the administrative assistant ”who is often already responsible for receiving emails and passing on important ones to the managers ”can perform this function. Is it really necessary to put a computer on every desktop?

I've found it to be typical in many corporate environments to deploy a wide array of applications to all desktops, or to segmented groups of user desktops, that are never used and aren't required for the user to perform his job function. Does every user in your company need every application for which you have purchased a site license? But the downside to this is that it might be more expensive to tailor desktops on a user-by-user basis than by groups or the entire enterprise as a whole. It's a balancing act. Which is less expensive? Configuring (and paying for) applications that not every user needs, or spending your budget for administrative and help desk personnel to determine just what applications a user does need? I like to refer to this as application overload .

Both the Windows and Unix/Linux operating systems have their own strengths and weaknesses, and each can be used to solve certain problems more effectively than the other. By integrating these two systems into the same network, you can take advantage of each one's best capabilities and enhance performance of the services provided to users.

For example, Microsoft makes the majority of its profits based on sales of operating systems and the Microsoft Office suite of products. There are many alternatives. One of these is offered by two different organizations. The OpenOffice initiative is an open source suite of applications that, although not as sophisticated as Microsoft Office, is free. And the spreadsheet and word processor applications can read/write files based on Microsoft formats, for the most part. In a recent test, I found that the OpenOffice spreadsheet program does a better job of printing these files to fit-to-page than Microsoft Office does in some cases.

Visit the Web site http://www.openoffice.org. You'll even find a Mac OS X version at this site. If you want to purchase similar software, with enhancements and support from the vendor, you can contact Sun Microsystems using the URL www.sun.com. There you'll find StarOffice, which can be installed on Solaris, Linux, as well as Windows systems. This product isn't free, but does have a respected vendor backing it, so you can expect to get a good amount of support if you decide to adopt it in your network.

Whichever you choose, OpenOffice and StarOffice both provide a subset of the functionality of Microsoft's Office product. Unless you need all the bells and whistles provided by Microsoft Office, consider this alternative. And because Microsoft Office products continue to incorporate new features that many users don't need, you might find that OpenOffice and StarOffice are all that you need for your business. This book was written using all three.

There are two approaches you can take to integrating these two kinds of systems in one network. First, you can use the features Microsoft provides for the Windows environment, which were derived from standards that were developed in the Unix world ”such as TCP/IP and the standard suite of utilities that have been written around it (FTP and Telnet, for example). Second, you can use third-party applications, such as Samba, that have been created to allow Windows Server Message Block (SMB) and Common Internet File System (CIFS) functionality to be installed on Unix/Linux platforms. This last approach seems to be the path that Microsoft will be taking in the near future, heralded by the release of the Microsoft Services for Unix (SFU) version 3.0. SFU provides components from Microsoft and other vendors that can make life easier for a Unix administrator who inherits a network of Windows clients .

Refer to Chapter 35, "File Server Protocols," for more about Samba, SMB, and CIFS.