This week course introduces the history "NSFnet" the first Internet. In the 1960s and 1970s, she established the first wide area packet-switched network for the US Department of Defense, a number of studies. In the 1980s, the first "Internet" form to allow academic researchers access to supercomputer resources.
Supercomputers Justify a National Network
This section describes the actual experience of 60 years and is connected, communication and computing 70s.
1970s, people began to use the teletype, just like in the video. At the time, the communication between the city through a copper cable, every time someone calls between the two cities, we must find a wire and connect them, so call audio will be via telephone lines, transmission to the other side. Their prices are very high and the cost is also associated with distance. Therefore, when the cost of maintaining, a leased line is common, and the node establish and maintain a distance between each node, such as 1-2 km, and then connected to the computer through a dial-up close to this computer behind is a computer remote. So that people can communicate with the users to connect with the remote computer.
But this would be a problem, the problem is network congestion, the information sent by the user in a computer node will be queued in front of the information forwarded again after the completion of transmission, but a computer to the next node, even if forwarded , will enter the waiting, so that the sender can not determine the state information, and a large delay. And if you want to establish a direct connection, of course, solve the problem of delay, but the cost is very expensive, so when the eventual establishment of a network infrastructure to encourage more jump, which means it encourages more delay. Because by encouraging universities or other institutions in various places to build their own networks and their own networks connected to each other, you can spend a very low cost, so that transmit the information to the area.
1960s to 1980s, the US Department of Defense is investing in a research network called the ARPANET. ARPANET was funded by the military, ARPA, the Defense Advanced Research Projects Agency. The main motivation of this project is to improve their computer equipment for military purposes, making it easier to access. The second motivation is actually with the reliability, redundancy and resistance to the relevant part of the fault, this is more for the attack on the battlefield. If a node in the network is destroyed during the war, how to make the network keeps on running? ARPANET network solution is switched by a packet, a packet oriented transfer and exchange, which is a store and forward switching mode, i.e. prior to reaching the packet switch memory for temporarily storing and processing the corresponding output until the idle circuit when re-sent.
The First "Internet"
The title of this section Internet was marked by quotation marks, introduced during that period, although the true sense of the Internet has not yet appeared, but with the rapid development of various Internet projects, in the era of the Internet has been great strides son, soon He arrived.
At the time, Larry Smarr and people to create a supercomputer to convince Congress to authorize the establishment of the National Science Foundation for funding the network. It will use to build the ARPANET TCP / IP protocol. But at that time the University of Michigan did not join ARPANET. Because the University of Michigan established its own network. The use of leased lines has established a network, which is a three-node network between Michigan State University, University of Michigan and Wayne State University. ARPANET and was not seen as many people are so important. University of Michigan as a large university, really want to build a supercomputer center. As some strategic mistakes, Michigan did not get a supercomputer center.
Doug Van Houweling know the current in the larger environment, universities are less likely to have their own supercomputer (expensive). So he wants to build a network of existing computers, expanding the use of surface supercomputers. Houweling became the overall leader of Merit project, but the lack of pre-budget project, but Houweling put forward the idea of industrialization, with IBM coordination, overcome suffering, and ultimately solve the funding problem, and with impressive results.