NetGuido

Develop a new implementation kernel of the whole protocol tcp / ip, quality, and walk as well as existing products, was not an easy task. The decision not from an existing implementation was taken at a time when there was a doubt about any restrictions on copyrights, because of U.S. court decisions, and at a time when there were many enthusiasm to do differently and perhaps even better than what had been done before.


The first volunteered to lead the development was Ross Biro <biro@yggdrasil.com>. Ross produced an implementation of simple routines, incomplete, but perfectly usable, to which was added a driver for the Ethernet network interface card WD-8003. That was enough for many people try the software and even some s’arrangèrent to connect with this configuration, on the Internet live. The pressure of the Linux community who was attending to the development of network support grew, and finally, convergence of this unjust pressure and its own obligations prevailed on the merits Ross pulled it therefore stopped its task as coordinator development. Ross’s efforts to start the project, its acceptance of responsibility to do something really useful in such turbulent circumstances, were the starting point for any further work and therefore an essential element for the success of current product.
Orest Zborowski <obz@Kodak.COM> produced the first BSD socket interface for the Linux kernel. It was a great step forward and enabled many existing network applications to be brought under Linux without major changes.
At about this time Laurence Culhane <loz@holmes.demon.co.uk> developed the first pilot Linux to support the protocol SLIP. This enabled many people who lacked access to an Ethernet network to try the software network. Then some used this driver to connect to the Internet. This gave more people a glimpse of what is possible if Linux had full support for the network and increase the number of users using and experimenting with this software network.

One who has also actively worked on the construction of network support was Fred van Kempen <waltje@uwalt.nl.mugnet.org>. After the period of uncertainty that followed the withdrawal of Ross, Fred offered his time and accepted the role of driver development without encountering opposition. Fred had some ambitious plans about the direction that he wanted to bring the network software Linux, and he began to make progress in these directions. Fred produced a series of networking code called the kernel code `NET-2 ‘(the code` NET’ being that of Ross), which allowed many people to use it with interest. Then Fred mit number of innovations in the further development, such as the interface device dynamic, protocol support amateur radio AX-25 and an implementation network designed more modular. The code NET-2 Fred was used by a large number of enthusiasts, this number increasing as the use of software in the world. The network software at that time still consisted of a large number of patches that should be applied to the kernel code and was not included in the normal distribution. The document NET-FAQ and his successor NET-2-HOWTO described the rather complex procedure so it works. Fred focused on the development of innovations and it took time. The user community s’impatientait because she wanted to have something functioning properly and that would meet 80% of users then, as with Ross, the pressure on the responsible development increased.

Alan Cox <iialan@www.uk.linux.org> proposed a solution to improve the situation. He offered to take the code NET-2 Fred, debug it, make it reliable and stable so it would satisfy the user base forward, releasing the pressure on Fred could continue his work. Alan began to work with some success and its first version of Linux networking code was called `Net-2D (ebugged;). The code worked reliably with several typical configurations and the user base was happy. Alan really had ideas and skills to contribute to the project and many discussions about the direction that should take the code NET-2 were followed up. He then developed two separate schools in the Linux community, one with the principle that `it works and then improve on then ‘and other` improve first. Linus arbitra and finally offered his assistance to the development efforts of Alan and includes its code in the standard distribution of the kernel. This placed Fred in a delicate situation. Any development of long-term suffering for lack of use and testing by the user base and it would mean that progress would be long and difficult. Fred continued to work for some time, then finally retired, and Alan became the new driver for development of network code in the Linux kernel.

Donald Becker <becker@cesdis.gsfc.nasa.gov> quickly proved his talents in all aspects of low-level network and produced a huge quantity of pilots Ethernet, almost all those included in the current pits is to him. There are other people who have made a significant contribution, but Donald’s work is prolific and therefore deserves special mention.
Alan continued to refine the code NET-2-D (ebugged) for some time, while progressing on some issues that remained unresolved in the list of `TODO ‘(note:` To Do’). While the sources of Linux kernel 1.3 .* made their first steps, network code migration to the distribution NET-3, on which current versions are based. Alan worked on many aspects of networking code and, with the assistance of many talented people from the Linux community, developed the code in all sorts of directions. Alan produced a pilot network devices, the first standard AX.25 and implementations IPX. Alan continued to rafistoler the code, restructuring and gradually bringing its level today.
The PPP support was added by Michael Callahan <callahan@maths.ox.ac.uk> and Al Longyear <longyear@netcom.com>, what was important to increase the number of people using Linux willing to go on the network.
Jonathon Naylor <jsn@cs.nott.ac.uk> brought its contribution to improving the code AX.25 Alan and adding protocols NetRom and Rose. The support AX.25/NetRom itself is quite significant, because no other operating system that Linux can boast of having a native support for this protocol.
There were obviously hundreds of others who have made a significant contribution to the network layer of Linux. You will find some later in the paragraphs dealing with specific technologies, others have collaborated on the modules, drivers, bugfixes, suggestions, test reports and moral support. In any case everyone is entitled to have played a role and offered what he could. The network code Linux is an excellent example of what can be achieved with a style anarchic development of Linux, if you did not even surprised, and we see yet, development has not stopped.