This article throws light upon the three main services widely used by sample networks. The services are: 1. Novell NetWare 2. Arpanet 3. NSFNET.
Service # 1. Novell NetWare:
Novell NetWare is the most popular standard in the PC world, where it is widely used for Local Area Networks (LANs). Its origin is based on the sudden requirements which came in the market because many companies decided to change from main frame systems to distributed data processing in the form of a PC-based network.
Where these new networks were geographically limited to a small physical area and could, therefore, have a LAN network, Novell NetWare became very popular as a networking tool. A network based on Novell NetWare is a client-server network, with a few servers, mostly acting as file servers although other services (such as data-base services) are also sometimes loaded on these servers.
Novell Netware is an adaptation of an old Xerox networks the XMS. It does not follow the OSI model or the TCP/IP model, although it does have some similarities with the TCP/IP reference model. There are numerous protocols used for the Application Layer, depending upon the applications used. It does not have the Presentation Layer or the Session Layer and just below the Application Layer is the Transport Layer.
ADVERTISEMENTS:
The Transport Layer really provides the most important services in Novell NetWare. It uses a protocol called Network Core Protocol (NCP) that provides many services beside the standard Transport Layer functions. Sometimes a protocol, called SPX, is also used to provide the standard Transport Layer services.
But SPX, unlike NCP, does not provide other functions. To that extent NCP, which is a connection-oriented protocol, may be called the representative protocol for Novell NetWare. As for the Physical Layer and the Data Link Layer, Novell NetWare can use Ethernet (IEEE 802.3) or Token ring (IEEE 802.6) or even ARCnet. For the Application Layer, Novell NetWare uses a Service Access Protocol (SAP).
The Network Layer uses a protocol called IPX. This at once points to a limitation in the use of Novell NetWare for Wide Area Networks. Wide Area Networks typically need the TCP/IP protocol, but IPX does not provide this protocol. This limitation of Novell NetWare has probably been the chief reason for it to remain an accepted yet popular network, but only for Local Area Networks.
Service # 2. Arpanet:
In the march towards various Wide Area Networks and the Internet, the ARPANET was a necessary station. In the mid-60s, the US Department of Defence needed a network for the purposes of command and control.
ADVERTISEMENTS:
Obviously, for this purpose, a Wide Area Network was required. It created this network through its Advanced Research Projects Agency (ARPA). This was really a misnomer because ARPA did not have any researchers to research into the requirements of networks.
Instead, ARPA gave grants to several universities and contracts to several companies to develop a suitable Wide Area Network and thus ARPANET was born. ARPANET is a packet-switched network with a communications subnet and hosts.
The task of creating such a packet-switched network was given to a consulting company called BBN from Boston. BBN chose to use DDP-316 computers from Honeywell as the Interface Message Processors (IMPs) and 56-kbps leased telephone lines for its communications subnet.
At the head of the communications subnet were the IMPs and the protocols used consisted of IMP-IMP protocol, source IMP to destination IMP protocol, Host to IMP protocol and a Host to host protocol. This is illustrated in Fig. 7.8.
The ARPANET was the precursor to today’s Wide Area Network in general and to the Internet in particular. When ARPANET started, the first host software was written by graduate students at Snowbird, Utah.
These graduate students designed the network without the aid of a network specialist. Finally, when the software was ready, the network went into operation in December, 1969 with four nodes—UCLA, UCSB, SRI and the University of Utah.
These four universities were probably selected because they all had large contracts from the Advanced Research Projects Agency and also probably because they all had widely different machines to act as host computers. This meant that the incompatibility would be able to test the effectiveness of the network.
Another possible reason probably was the fact that these incompatible host computers provided an appropriate challenge for the graduate students. This network started to grow very rapidly as more and more IMPs were installed in the network, until it spanned the whole of the United States. At about this time, the concept of a Terminal Interface Processor (TIP) was developed.
ADVERTISEMENTS:
This TIP permitted nodes to be attached to the network without going through a host. Subsequent developments included having several hosts connected to each IMP and also each host being connected to several IMPs.
This would be of great help in case of the failure of a particular IMP. IMPs were also separated by large distances. ARPANET also utilised satellites for communication within the network. These developments also demonstrated that the protocols used by ARPANET, were not suitable for multiple networks.
Therefore, a new set of protocols, called TCP/IP, was developed. In 1983, control of the ARPANET was handed over to the Defence Communications Agency (DCA). They removed the military network from ARPANET and formed a separate network called MILNET. The rest of the network was left as a research network.
As new networks kept on being added to ARPANET, finding hosts for each new network started to become an expensive process. To get over this problem, the concept of a Domain Name System (DNS) was developed. The Domain Name System mapped hosts to IP addresses and organised machines into domains.
ADVERTISEMENTS:
Since then, there has been a great deal of development in the DNS, culminating in the Internet. While we shall not go into any further details of ARPANET, it may be interesting to trace the history of the development of the Internet. This we shall do by looking at another network, the NSFNET, which provided another step in the development of the Internet.
Service # 3. NSFNET:
By the late 70s, the success of ARPANET was obvious. University research was one of the main beneficiaries of the ARPANET; scientists and researchers all over the US were able to share data, the results of their work and their conclusions over the ARPANET.
Unfortunately, the use of the ARPANET was strictly controlled because of the involvement of the US Department of Defence, since only those universities that had contracts with the Advanced Research Projects Agency of the Department of Defence could use ARPANET.
The National Science Foundation of the US observed this fact and decided to set up their own Wide Area Network to use largely for the purpose of communication between research scientists.
In association with BBN, the consulting agency from Boston, they set up a virtual network called CSNET. BBN provided the main server and while the network was based on dial-up lines, it could be connected to other networks including the ARPANET.
CSNET worked, but within its severe limitations. However, the US National Science Foundation (NSF) wanted to create a high-speed network that could be used by research scientists without the limitations imposed by ARPANET. So, they decided to design a high-speed network and in the mid-80s they succeeded.
They started with a backbone of six super-computer centres located at the universities at Boulder, Champaign, Ithaca, Pittsburgh, Princeton and San Diego. These were connected to each other over 56-kbps leased lines and more importantly using TCP/IP. The NSFNET was thus the first Wide Area Network using TCP/IP. NSF also provided funds to connect some two dozen Local Area Networks to this super-computer network.
These consisted of libraries and universities. NSFNET was connected to ARPANET through Carnegie-Mellon at Pittsburgh. NSFNET was a runaway success, right from the start and it became obvious to NSF that an advanced successor would soon be required. For this, they used an organisation called MERIT from Michigan.
They used 448-kbps leased lines from MCI and RS6000 computers as routers. This leased line was soon upgraded to 1.5 Mbps. However, at this point NSF realised that while the controls over the NSFNET were much less than that over ARPANET, it still was not available absolutely freely.
This was because NSFNET could not be used for commercial purposes or by commercial organisations, as per the Charter of NSF. Therefore, a non-profit organisation called Advanced Networks and Services (ANS) was formed.
NSF, in fact, did the prompting for the formation of this organisation and it was necessary to form this organisation because the Charter for NSF did not permit any network set up by it to be used for commercial purposes. Therefore, in order to form a freely usable public network, it was necessary for NSF that it did not finance such a public network.
When NSFNET was taken over by Advanced Networks and Services, it was upgraded by ANS and called ANSNET. This upgrade consisted largely of upgrading the 1.5-Mbps link to 45 Mbps, in 1991. In 1995, ANSNET was sold to America Online since the US Congress wanted a Gigabit network. Meanwhile, several commercial companies had established commercial networks, all of which were using the Internet Protocol or IP.
But the regional networks of NFSNET had to be disconnected from ANSNET as a result of this sale. For this, NSF asked several network companies to establish a Network Access Point.
This in one sense was the start of the Internet as it is today. Other such networks have also come up in other countries, such as EBONE and Europa NET in Europe and the NICNET in India. This entire chain of events have ultimately led to the development of the Internet.