The ARPANET, the Internet and TCP/IP
In 1958, in response to the launch the previous year of the Russian Sputnik, the
United States Government set up the Advanced Research Projects Agency (ARPA) to
try and establish a US lead in science and technology within the armed forces.
In 1965 ARPA funded a study on a 'cooperative network of time-sharing
computers'. The TX-2 at the Massachusetts Institute of Technology (MIT)
Lincoln Lab and the Q-32 at System Development Corporation (Santa Monica CA)
were directly linked as a result.
The design and specification for the experimental ARPA network was written by Larry Roberts who oversaw all the work on it between 1966 and 1973. By 1969 ARPANET was commissioned by the Department of Defense (DoD) for research into networking. It was initially a four node network comprising UCLA, Stanford Research Institute, UCSB and the University of Utah. The network consisted of the host computers at the research centres which were connected to Interface Message Processors (IMP). The IMPs were based on the Honeywell DDP516 computer and developed by Bolt Beranek and Newman (BBN) who had been given the contract to set up the system. basically this was a network of IMPs which communicated with the host computers. (See diagram at right). Communication between the IMPs was by packet switching.
ARPANET was originally conceived to allow users at research centres (funded by ARPA) to use computers at other research centres. By it's design, messages could be routed in more than one direction, thus the system could continue to operate even if parts of the network were destroyed or out of action.
In 1970, ARPANET hosts started using Network Control Protocol (NCP) which was a host to host protocol, but would only function between hosts on the ARPA network, it would not allow communication between networks. To do this, Transmission-Control Protocol (TCP) would be required. By 1978 TCP had been split into two protocols, TCP and Internet Protocol (IP), officially known as TCP/IP.
ARPANET and Internet:- The difference:-
Basically the difference between ARPANET and the Internet is, ARPANET was a network which one was invited to join, and the Internet is a 'Global network of networks' open to anyone.
In 1962 the US Air Force commissioned a study for a design which would allow the US command and control systems to withstand a nuclear attack. Paul Baran, a researcher at Rand Corporation, produced a report entitled, 'On Distributed Communications' in which he envisaged a distributed network with no central node which could continue to function even if part of the network was destroyed, rather than a centralized or decentralized network.
His other innovation was to split the messages into small packages, he called them ‘message blocks’. These digital messages would be passed, using a network of computerized switches, from one node to another without any need for a central control, each node choosing the best route for the next stage. Around the same time, Donald Watts Davies of the National Physical Laboratory (NPL) in the UK was formulating his idea of ‘packet switching.', which was very similar to Baran’s ‘message blocks’.
The initial plan for the ARPANET was released in October 1967 and was for a network of four sites. The sites or hosts chosen were University research centres carrying out ARPA funded research. The network consisted of the four hosts, which were each linked to their own Interface Message Processors which would ‘be built as a messenger, a sophisticated store-and-forward device, nothing more. Its job would be to carry bits, packets, and messages: To disassemble messages, store packets, check for errors, route the packets, and send acknowledgments for packets arriving error-free; and then to reassemble incoming packets into messages and send them up to the host machines – all in a common language.’ 1.
By December 1969 all four host computers were linked and the fledgling ARPANET network had been born. Nodes were then added at the rate of about one a month and by August 1972 there were 29 nodes in the ARPANET. A Terminal IMP (TIP) interface had been developed which enabled numbers of terminals to be linked to an IMP. Around this time also, BBN finally acquired their own IMP and this allowed them to take control of the network for the remote installation of software upgrades and diagnostic checks.
The ARPANET kept expanding during the 1970s and other networks came into being. In 1970 Aloha net packet-radio network in Hawaii was one of the first, Cyclades in France, NPL system in the Britain and Satnet the satellite network followed, these networks used packet-switching to communicate within their own networks, but the problem was that they could not communicate with each other.
In 1973 ARPA initiated the Internetting Project. This came about mainly because Bob Kahn, who had organised the first public demonstration of the ARPA network in 1972, recognised that single networks run by one organisation were not really able to expand appreciably. He realised that for heterogeneous networks to be connected, they would need standard interfaces and protocols between the networks to send the packets to their appropriate destinations. He originally called them ‘gateways’; today they are called ‘routers’.
Bob Kahn teamed up with Vint Cerf and based on Cerf’s ideas, they worked out the details of a new protocol that would allow different networks to connect together into one big network- an Internet. ‘Our thought was that, clearly, each gateway had to know how to talk to each network that it was connected to,’ Cerf said. ‘Say you’re connecting the packet-radio net with the ARPANET. The gateway machine has software in it that makes it look like a host to the ARPANET IMPs. But it also looks like a host on the packet-radio network.’ 2
Fig. 2. Vint Cerf's first diagram of his Internet design ideas.
By the end of 1973 they had completed their paper ‘A Protocol for Packet Network Intercommunication’. The protocol they invented they called Transmission Control Protocol (TCP). By 1978 it had been refined and developed and split into two protocols and was now known as Transmission Control Protocol and Internet Protocol (TCP/IP). Meanwhile the ARPANET had grown into quite a large network of very varied host computers, but the subnet of IMPs was still an integrated operation, built and operated by BBN, who could monitor, control and reboot the IMPs from their control room in Cambridge Massachusetts and therefore by definition controlled the network.
In January 1983 ARPA officially adopted TCP/IP as the protocol suite for ARPANET and a gateway was opened between ARPANET and Computer Science Research Network (CSNET). This is regarded by many as the true birth of the Internet. The gateways opened the floodgates, so to speak, as networks opened gateways between each other. With TCP/IP protocols, the hosts were now responsible for the end-to-end transmission of packets along with error correction and retransmission as required.
As the network grew, it became obvious that there were going to be large numbers of host computers with the same name and this would cause problems addressing them. In 1984 the Domain Name System (DNS) was introduced, it consisted of seven ‘top level’ domains: edu, com, gov, mil, net, org and int, now there could be seven computers with the same name at different sites. At a summit meeting of all the major networks in 1986, agreement was reached to fully implement the DNS.
TCP/IP and DNS are two of the major pieces of technology that distinguished the Internet from ARPANET. The Internet allows a host computer using TCP/IP protocols, to communicate end-to-end from one network, via gateways to a host computer on another network, with no central controlling authority. The network has no control over what it transmits, that belongs to the users, the hosts. ARPANET could allow host-to-host communication via IMPs, which were responsible for sending/receiving data, checking for errors, routing data and verifying that messages had arrived but only within the ARPANET and was controlled by BBN.
Anyone can become part of the Internet if they have a computer, the software and can pay for the connection, there is no regulatory body to decide who can join. The ARPANET was restricted to research centres funded by ARPA and the research centres had to develop their own software to communicate with the IMPs.
TCP/IP:-
Technically, what distinguishes the Internet is its use of a set of protocols called TCP/IP. The development of TCP/IP was important in that it is the basic communication language or protocol of the Internet. Every computer with direct access to the Internet has a copy of TCP/IP, whether the operating system is Windows, Macintosh or UNIX.
TCP/IP is a two-layer system. The TCP layer assembles the messages or files into packets, which are transmitted over the Internet and received at their destination by a TCP layer, which reassembles the packets into the original message. The IP layer handles the address of each packet and sends it to the right destination. As the packet travels over the networks each gateway computer checks this address and forwards it. Some of the packets from the same message may be routed differently from the others, but they are reassembled at the destination. Note, the gateway (router) computer retains no information about the packets passing through. This was designed to speed up the traffic but it also removes a means of control and censorship.
At the same time as the TCP/IP protocol was being developed an addressing scheme encompassing all the networks was introduced. The original ARPANET addresses were 8 bits that limited the network to 253 hosts. The IP protocol introduced 32 bit addresses that allowed millions of addresses.
By 1989 the number of hosts exceeded 100,000, by 1990, 300,000 and by 1992 it was over 1,000,000. this rapid growth was due to the ideas underpinning the TCP/IP protocol, namely that each network would develop its own applications without restraint and require no modification to participate in the Internet. The gateway software would retain no information about the data passing through, therefore no censorship. Packets would be routed by the fastest available route. The operating principles of TCP/IP would be freely available to all networks.
In 1990 the ARPANET was decommissioned. It had served it's purpose. Its legacy is the Internet, the Global network of networks. The future; I can visualize free access, the bandwidth increasing to give greater interactive activities on the WWW and dramatically reducing the access time. The Internet being used for voice, high quality TV and radio services and TCP/IP being further developed to accommodate these advances.
References and acknowledgments:-