Each distinct physical network would have to stand on its own, connecting it to the Internet would require no further internal change.

Communication between sources would be on the basis of best-effort, that is to say that if a packet didn't make it to it's destination it would be re-transmitted from the source, algorithms would be developed to enable this facility.

Networks would be connected by a 'black-boxes', now known as gateways and routers, the gateways would be unintelligent, retaining no information about packets transmission.

There would be no global control at the operations level

Techniques for host to host flow control

A need for global addressing

Gateway functionality to allow for correct forwarding of packets, this would include recognition of IP (addressing) headers for routing, handling interfaces and breaking packets into smaller pieces if required.

The requirement for end-to-end checksums, re-assembly of packets after fragmentation and detection of corrupted or duplicated packets.

Interfacing with different operating systems

Some basic approaches between Kahn and Cerf were then developed.

Communication between two processes would consist of a long stream of bytes (they called them octets), the position of any octet in the stream would be used to identify it.

Flow control would be implemented using sliding windows and acknowledgements (acks), the destination would select when to acknowledge and each ack returned would be cumulative for all packets received at that point.

Kahn and Cerf designed a system using a 32-bit IP address, the first 8 bits signified the network and the other 24 bits the host on the network, the assumption that 256 networks were sufficient needed much re-consideration as LAN's grew throughout the 70's and 80's. Although Ethernet was being developed at the time at Xerox PARC, the current proliferation of LAN and desktop computers was not envisaged, this initial design flaw has led to an inherited problem of a dearth of IP addresses within LAN's15.

Jumping through history for the sake of brevity, a number of occurrences developed leading to the Internet of the present time. LAN's, PC's and workstations flourished throughout the 1980's and the fledgling Internet began to gain momentum, the original ARPANET model with its notion of a handful of hosts was revised and a definition of three network classes followed.

Class ALarge national scale networks (small number of networks having a large number of hosts).

Class BRegional scale networks.

Class CLocalised networks, many networks, few hosts.

Because of the proliferation of networks, people found it difficult to remember the unique numerical IP addresses assigned to hosts, this problem was alleviated by the introduction of the Domain Name System (DNS)16, by Paul Mockapetris of USC/ISI. The DNS gave a mechanism by which hierarchical host names such as 'http://www.byte.com' may be resolved into the corresponding IP address, in this case 198.83.40.72 which a computer can then interpret.

The major milestone came with the adaptation of TCP/IP as the host protocol by the ARPANET on January 1st 1983, thus the older and limited NCP was finally laid to rest and TCP/IP recognised as the way ahead for a global communication system. Such an action complemented the earlier adaptation of TCP/IP by the defense bodies in 1980.

As a final point of interest, the ARPANET itself was split into the ARPANET and MILNET in 1983, the MILNET became integrated with the Defense Data Network created in 1982 whilst the ARPANET itself was finally decommissioned in 1989.

Communication using protocols is a complex business and this dissertation can only skim the surface of such topics, a number of texts give a much greater explanation of the topic54.

It is more important to realise that the following features now define the function of the TCP/IP protocol suite.

The following schematic gives a crude representation of the processes involved in moving binary data using the TCP/IP protocol, for the purposes of the diagram, we have assumed that the destination IP address has already been resolved by DNS.

The original data request is broken down into smaller and more manageable chunks, each of these packets then becomes encoded with in a 'packet' of TCP encoding. TCP encoding gives the packet its own identity and the information is used to verify the integrity of the packet if and when it reaches its destination.

At this stage the packet knows no destination, the TCP packets are then encoded with an IP address to give them a location to be routed to. On reaching the destination, the IP and TCP information is stripped away to reveal the original data request, a process very analogous to opening an addressed letter that has been posted to you so you can read the contents.

All of the considerable development work highlighted previously has mutated into the Internet of today. The Internet is now a giant global network comprising of military, industrial and academic networks.

The physical backbone of the Internet is comprised of large fixed links between nodes, these nodes tend to belong to the Internet Service Provider's (ISP's). Individual home-users are free to become members of the Internet too, this is normally achieved by taking out an account with an ISP for a small monthly charge, the academic community is served by JANET, the joint academic network.

The central point of information exchange is the LINX (London Internet Exchange)17, the LINX is based in the Docklands and serves as the highest level point of exchange in Great Britain for the transfer of Internet traffic between both ISP's, JANET and the outside world.

The following table gives an indication of the organisations which combine together to form the LINX. Although the names of a lot of ISP's are mentioned, it is important to note that although such organisations serve the home user, they also provide larger connections for industry, thus D2D is actually connected to the 'outside' world via a gateway supplied by UNET PIPEX as are many similar companies.

The LINX consists of a number of Catalyst Ethernet Switches connected by an

FDDI (fibre) ring, the network has the IP address 194.68.130.0/24.

Having discussed the physical nature and the data transport aspects of the Internet, it becomes more interesting to see the range of services which are available to the Internet community.

Although TCP/IP provides the major transport mechanism for the Internet, a number of other protocols e.g. FTP, SMTP and NNTP are also used to implement the various services which the Internet provides, these protocols are highlighted where required in the following section.

The use of the file transfer protocol (FTP)18 allows for the transfer of files from one computer to another. FTP was one of the original Internet services, driven by the fact that the computers of the time didnt have facilities such as disk drives for the transfer of information. Many Universities and other large institutions maintain considerable archives of material available for download via FTP, security is usually handled by requiring the user to log on with a user name and password, though the users e-mail identity and the password 'anonymous' are usually all that is required for the downloading of such public domain material. FTP client programs e.g. 'Fetch' (Macintosh) and 'Rapid Filer' (PC) tend to just show a list of downloadable files without any fancy GUI's.

Electronic mail was invented in about 1972 as previously mentioned, and was considered a killer application at the time. For those who havent encountered e-mail previously, it is best to imagine e-mail as simply an electronic version of posting a letter. The message is composed using an editor e.g. 'Eudora' (Mac), 'First Class' (PC) or 'Pine' (UNIX) and posted to an address e.g. jbloggs@home.co.uk. The name is resolved by the DNS and the message is transmitted, reaching its destination after being routed via one or a number of mail servers.

E-mail addresses come in several flavours depending upon the nature of the hosting organisation1(a). Electronic mail programs and protocols have improved steadily over the years, you are now able to not only send text messages but to attach formatted documents to them e.g. you could post a large manual in electronic form to New Zealand and it would be received within a matter of hours, not days as per the more conventional surface mail.

SMTP19 and MIME20 are the most commonly encountered mail protocols, the former deals with the transportation process whilst the latter is concerned with the encoding of attachments.

Of all of the uses of the Internet, the feature that has really captured the imagination of the public is the World Wide Web or WWW as it is often referred to. The WWW is the visual side of the Internet and was initially developed by Tim Berners Lee at CERN, the European Laboratory for Particle Physics in 1989. Common usage of the WWW exploded with the introduction of the first browser (WWW page viewer), which was developed in 1992. The WWW is a global hypertext system, hypertext being a computer based system of linking documents to related documents. The links are usually visible as highlighted text and are mouse-clickable.

Each unit of information is known as a resource, be it an imbedded graphic image, sound clip or document, these resources are identified by their universal resource locator (URL), e.g. http://www.ford.co.uk/. The standard WWW retrieval mechanism is the hypertext transfer protocol (HTTP)21, another Internet protocol.

The logical structure of the text is described using a mark-up language, hypertext mark-up language (HTML), currently version 2.022, this language tells the browser (the client program that displays the hypertext) how to display the various elements which make up the page.

In the early days, the WWW only offered a display of static pages and a lot of early browsers couldn't even handle graphics. Progress has been rapid however and the stage has been reached where it is possible to read from and write to databases as well as to embed application programs written in languages like Java within web pages.

This level of interactivity has added an extra dimension to the WWW and business concerns have not lost any time in utilising the web for their own purposes, both to the outside world as means of advertising their wares and gathering customer profiles and in-house, through the use of Intranets, a subject that will be dealt with in its own right shortly.

The use of so-called 'Newsgroups' makes for an interesting concept, one which it was frequently necessary to explain at D2D. The best way to view a Newsgroup is as an electronic open-discussion forum, imagine sitting in a room with a group of friends when one of them advances an idea, the likelihood is that one or more people will give a response to that opinion, not necessarily agreeing with it and perhaps advancing a new set of opinions instead. Discussing politics or favourite football teams is a likely scenario one might imagine, what tends to develop is a whole hierarchy of related discussions, all traceable back to that initial comment.

Most of the Newsgroups on the Internet come under the umbrella of USENET23, USENET pre-dates the Internet and is not a dedicated network or piece of software, it is a set of voluntary rules for those who maintain and use Newsgroups. USENET is made up of seven basic categories.

Besides these groups which have been recognised and adopted by USENET, any server administrator is free to create their own Newsgroup locally, some of these local Newsgroups have been distributed as widely as the 'official' USENET ones, to which end groups such as 'alt' and 'gnu' tend to crop up frequently.

In the current world of GUI's, one tends to find that news items are read not through the use of the browser but through so called helper applications, these are usually developed as so-called 'plug-in's' which extend the functionality of the browser. An exception to this is the WWW type interface available when using 'Dejanews' [24] this allows Newsgroups to be examined from a straight WWW browser such as Netscape or Internet Explorer.

As a final note, it should be noted that Newsgroups adhere to the use of the Network News Transfer Protocol , (NNTP)25, this protocol was originally designed to enhance the performance of USENET when running on a TCP/IP protocol suite.

This type of protocol allows someone to log onto a remote computer over a network and to work as if they were operating that computer, the net effect of which is to make the operators computer function as a dumb terminal, every character that typed is sent straight over the network to be interpreted by the host machine at the other end. Many specialist services such as browsing library catalogues and similar large databases are accessed by Telnet programs.

A number of other Internet services are available, these include Gopher, Archie, and Veronica. These services are no less relevant than the likes of the WWW, e-mail and newsgroups, indeed, they predate some of these services, it is more likely however that the latter services are more easily comprehended and are more likely to be implemented upon an Intranet.

This is also an appropriate juncture to introduce the term client-server architecture as the concept is fundamental to the implementation of Internet and Intranet technology.

The terms client and server cause a lot of confusion for many users as people often think of metal boxes with computer components inside when the word server is mentioned. In fact, the term client-server actually refers to a conceptual software architecture rather than any physical collection of objects.

A client is a piece of software which handles user input and output and which requests services from a third party piece of software, the server. Clients usually have a graphical user interface (GUI), e.g. Netscape, Internet Explorer (WWW) and Eudora, First Class (E-mail).

The server handles requests from the client, a server may well make further requests to another server becoming a client in its own right. A client and server may well be located on the same physical machine or they may communicate across a LAN, WAN or if the Internet is involved, quite possibly the other side of the world.

By means of example therefore, someone wishing to view a WWW page uses a piece of client software, a WWW browser, the client issues a request for service to the server, a piece of software dedicated to serving HTTP requests in this case. The server processes the request as appropriate and returns the information requested to the client.

Whilst the example chosen has been one of browsing a WWW page, the same analogy extends to the use of FTP, mail, news and all of the other Internet services. The major point to realise is that if one wishes to provide a particular Internet service then the appropriate type of server software must be run on a machine which is capable of handling requests using the TCP/IP protocol and furthermore, that the machine must have a physical connection to the Internet so that user may access the service.

Having discussed at length the attributes that constitute the Internet and its services, it remains to introduce the concept of an Intranet. A simple definition would say that an Intranet functions as an internal company Internet, that is to say that a network based on the TCP/IP protocol is implemented offering the same kind of facilities, News, FTP, WWW etc. as the can be found on the Internet. The main difference is that Intranets are aimed primarily as a tool for the employees of the company and as a consequence, confidential or even secret information may well be placed upon the system.

If the company Intranet is self contained then any security problems are likely to come from the inside. It is more likely however that the Intranet will run on the same infrastructure as the rest of the company network and will connect to the outside world, most companies allow employees to access the outside world by the use of WWW browsers for example.

For this more likely scenario, it is desirable to keep out external snoopers, be they of the malicious or merely curious intent, in regard of this aspect, a level of security called a firewall is often implemented26.

There are two basic ways of implementing a firewall, at the network and application levels. The firewall can act as a traffic router sitting between the Internet and the Intranet, rejecting undesirable IP packets, on the other hand, application level firewalls that run proxy servers can be used instead. These applications offer greater security because they are better at controlling access, though such applications have the disadvantage of being slower and less transparent to the user than network level firewalls27.

The use of Intranet technology offers the prospect of considerable improvements in efficiency and concomitant cost benefits in an environment where time is money28.

Bill Gates among others has noted that it's easier to cruise the Internet (WAN), find information, and have fun than it is to find something on most corporate LAN'sI1(a)

It is a fact of life that the typical company network hasn't received a major overhaul in some time and is comprised of a heterogeneous anarchy of hardware and applications nestling alongside more modern installations of desktop PC's running legacy software alongside more modern products such as Microsoft Office.

In such an environment, it is typical to find that only a handful of personnel are able to operate the older and more complex software systems and that fewer still are conversant with all of the applications in use throughout the company. Where Intranet technology earns its credit is in its role of simplifying the desktop environment for users.

Personnel need no longer require the special training necessary to use applications such as Microsoft Powerpoint and Microsoft Access. Suddenly, presentations may be viewed as HTML documents whilst simple forms allow personnel to read from, modify and contribute to databases, all with a few clicks of the mouse.

Of course, the thinking behind these concepts is complex, but the beauty of using such technology is that the user is shielded from the inner workings of the Intranet system.

Suddenly, it becomes possible to store and retrieve huge amounts of electronic information, software manufacturers have aided this process by incorporating 'save as HTML' options in their application programs, a multi-megabyte powerpoint presentation can be transformed into HTML within a few minutes on a moderately powered desktop PC for example.

Whilst working at D2D and talking with personnel, the comments that 'I spend ages looking for or designing presentations that have already been made' and 'I didn't realise that it (the Intranet) was so easy to use' were frequently encountered.

It becomes possible to lodge the company style sheets for official literature where everyone can access them, there is suddenly no need to print thousands of copies of standard company documents, the list becomes endless. With the correct hierarchy of contributors organised, the sharing of information within the company becomes streamlined and rapidly implementable.

The following lines briefly summarise just some of the aspects which need to be considered when implementing an Intranet. The majority of these topics were discussed at the meeting of the Intranet Development Council attended by the author on Friday 4th July, 1997.

Unification of software required leads to greater simplicity and reduction in man hours lost due to training.

How can it best be implemented ? Password protection of applications inside organisation, implement firewalls to secure from outside attacks.

Training Huge differences in capabilities and awareness of staff regarding Intranet issues.

This first section has been necessarily lengthy and detailed in order to show the large historical and technical development is encompassed by the use of Intranet technology. Those involved in pushing this technology within D2D have been mindful of such benefits. Chapter 2 deals with the application end of Intranet technology and shows how web page design can be extended to the use of database connectivity and the incorporation of news facilities, such features being found in everyday use on the Internet.