\documentclass[11pt]{article}
\usepackage{graphics} \usepackage{html} \usepackage{palatino}
\title{K12LTSP and Public Access Terminals using Linux}
\begin{document}
\section{Overview} This document should be seen as both a guide for strategists, in how to deploy public access terminals using an open source solution, but also a useful document for technical people wishing to build and develop systems using things like the Linux Terminal Server Project application or the K12LTSP distribution.
This document should be read alonside other documentation, including:-
Linux Terminal Server Project (LTSP): Documentation on installation and configuration and troubleshooting. http://www.ltsp.org K12LTSP:http://linux.riverdale.k12.or.us/k12ltsp.html Public-Web-Browser mini-HOWTO:
There are also a number of useful mailing lists: LTSP: http://www.ltsp.org K12LTSP: http://linux.riverdale.k12.or.us/k12ltsp.html Suse Schools Mailing List:
\section{Introduction} This document is aimed at educational or public-facing institutions who wish to maximise the number of computers available to their students or customers, but at the same time wish to minimise both the overall running costs and the management overhead.
\subsection{Why thin client?} Thin client machines have been in and out of fashion for a number of years, they do however have a number of significant advantages over the 'traditional' fat clients (ie standard PCs) that have been dominant for the last 10-15 years. These include: \begin{enumerate} \item Management: By having centralised configuration, the last of client management is significantly easier. \item Resilience: There is simply less to go wrong with a thin client (no local hard disk for example). Improvements in network design and performance also means that this is now a significantly lower risk that previously. \item Flexibility: Thin client networks can be configured in a number of ways, providing a range of desktops and services. This can theorectically enable mutlifunctional labs and access points to be set up. \item Purchase Cost: Thin client hardware is again becoming cheaper to obtain, and high prices in the past compared with standard PC hardware have prevented widespread uptake. \item Total Cost of Ownership (TCO): It is the lower ongoing costs that make open source thin clients an attractive solution. By removing licensing costs, increasing reliability and reducing the management overhead, using thin client equipment can significantly reduce the cost of IT provision. \end{enumerate}
\subsection{Why open source?} There is a growing amount of evidence, including case studies, of the potential benefits of using open source software as the basis for developing viable and robust IT solutions. Open source software is 'freely available', that is free to use, distribute and development, usually under the conditions of a suitable license \footnote{Information on licenses can be found on the \htmladdnormallink{OSI web site}{http://www.opensource.org/licenses/index.html}}
By using open source software, the developer and the end user benefit from not only having to pay ongoing licensing costs (not insignificant in propretiary thin client solutions), but also from not being tied to individual manufacturers or applications. The open source community has developed some significant applications (the Linux kernel and Samba, the SMB networking server for example) and as shown below has also developed some excellent applications for thin client solutions, such as the \htmladdnormallink{Linux Terminal Server Project}{http://www.ltsp.org}.
\subsection{Introduction}
\subsection{Performance} In terms of reliability and resilience, thin client can greatly increase performance.
\subsection{Management}
\subsection{Flexibility}
\subsection{Security} Public access terminals need to be very secure and by using a thin client strategy this can be done effectively.
\subsection{Scaling}
While Citrix has raised the awareness of thin client, its big limiter is the excessive license costs, which have a high entry level even for academic pricing. Although it is not a linear increase, there are cost considerations for extending the network. This is not a consideration for a LTSP solution.
\section{Design}
\subsection{Lab/Desktop Terminal}
\subsection{Public Access Terminal}
\subsection{K12LTSP Design} The K12LTSP distribution has specifically
\section{Building Your Server}
\section{Constructing Clients}
\subsection{Introduction}
\subsection{Legacy Equipment}
\subsubsection{Introduction}
\subsubsection{Selecting Your Hardware}
\begin{itemize} \item CPU. Anything from an Intel 66MHz 8x486 will work well enough to give you a viable client, \item Memory. 8Mb will work, but most machines should have at least 16Mb. \item Graphics. You really need something that can support SVGA graphics and that has a decent amount of graphics memory. It should allow for a resolution of at least 800x600 pixels. Check if the chip is supported by XFree86, and that ideally it runs using XF86 SVGA server. Other accelerated servers are supported, but it requires a little additional configuration. \item Network Card \item Other issues \end{itemize}
\subsection{New Equipment}
\section{Desktop Configuration}
\subsection{Getting the right desktop}
\subsection{Enabling Desktops}
\subsection{Gnome (Ximian)}
\subsection{KDE}
\subsection{Nautilus (Eazel)}
\subsection{Netscape}
\subsection{Star Office}
\subsection{Alternatives}
\section{Remote booting other OS's}
\subsection{Bpbatch}
\section{Remote Booting Servers}
\subsection{Why?}
\subsection{Web Server Example}
\section{Links and Further Reading}
Bpbatch - Remote Booting: http://www.bpbatch.org
K12Linux Project: http://k12linux.riverdale.k12.or.us
Linux Terminal Server Project: http://www.ltsp.org
Linux Web Station (formerly Public Web Browser) mini-HOWTO: http://www.linuxdoc.org/HOWTO/mini/Public-Web-Browser.html
Remote Boot mini-HOWTO: http://cui.unige.ch/info/pc/remote-boot/howto.html (using Bpbatch) \end{document}
This document was generated on 17 April 2001 using texi2html 1.56k.