History of TCP / IP
Internet Protocol was first developed by Defense Advanced Research Projects Agency (DARPA) in the year 1970 as the beginning of efforts to develop a protocol that can perform a variety of interconnected computer networks that separately, each network uses different technology. Protocol, the main project is produced by the Internet Protocol (IP). The same research also developed some of the high-level protocol designed to work with IP. The most important of the project is Transmission Control Protocol (TCP), and all group is replaced with the protocol TCP / IP suite. First TCP / IP implemented in the ARPANET, and began to develop after the University of California at Berkeley began to use the TCP / IP with the UNIX operating system. In addition to Defense Advanced Research Projects Agency (DARPA) is developing the Internet Protocol, also developed a TCP / IP is a Department of defense (DOD).
The terms of the Internet Protocol
There are some terms that are often found in discussions about the TCP / IP, which are:
Host or end-system, a customer of the service communication network. The host is usually the form of individual workstations or personal computers (PC) from the host where a task usually is to run this application and server software program that functions as a user, and implementing network communication services.
Internet, which is a set of network (network of networks) is a protocol and use TCP / IP to connect, such as virtual networks.
Node, which is the term applied to the router and host.protocol, which is a standard procedure or rules for pendefinisian and the transmission of data between computers.
Router, is an devais used as a connector between two or more networks. Router with a different host as the router bisanya not a goal or data traffic. Routing of IP datagram generally has been done with the software. So the function can be done by routing host which has two or more networks connection.
Overview TCP / IP
As cited above, the TCP / IP is also developed by the Department of Defense (DOD). DOD has conducted research project to connect several networks designed by different vendors to be a networks of networks (internet). At first this is only successful because they provide basic services such as file transfer, electronic mail, remote logon. Some of the computers in a department can use TCP / IP (along with other protocols) in a single LAN. IP component provides routing from the Department to the enterprise network, then to regional networks and finally to the global Internet. This can make the network communication can be broken, so to overcome the DOD and then design a TCP / IP that can automatically improve when there is a node or phone lines failed. This design allows the results to build a very large network with a center bit. Because of the auto repair problems in the network is not checked and not repaired for a long time.
Like the other communication protocol, the TCP / IP also has several layers, layer-layer are:
IP (internet protocol) that play a role in pentransmisian packet of data from node to node. Precede each IP packet data based on 4 bytes (for version IPv4) destination address (IP number). Internet authorities to create a number range to a different organization. Organizations to create a group with the number for the department. IP on the machine work gateaway that move data from department to organization to region and then later to the world.
TCP (transmission transfer protocol) a role in improving the delivery of correct data from a client to the server. Data can be lost in the middle of the network. TCP can detect errors or lost data and then perform the transmission back up with the received data is correct and complete.
sockets is the name given to subrutin package that provides access to TCP / IP on most systems.
some important things in the TCP / IP
1. Bottom Requestor Network (Network of Lowest Bidders)IP was developed to create a network of networks (internet). Individual machine connected to the LAN (ethernet or token ring). TCP / IP LAN to share with other users (Novell file server, windows, etc.). One devais provide TCP / IP link between the LAN with the outside world.
To ensure that all types of systems from different vendors can communicate, the use of TCP / IP on the LAN distandarkan. With increasing speed mikroprossesor, fiber optics, and digital phone lines and has created some new technology that is among ISDN, frame relay, FDDI, Asynchronous Transfer Mode (ATM).
Original design of TCP / IP is a network of networks that match the use of this technology now. Data TCP / IP can be sent via a LAN, or can be taken with an internal corporate SNA network, or data can be connected to the TV cable. Furthermore, the machines that are connected on one network can communicate with other network gateways via the vendor network.
2. Problems address
In an SNA network, every machine has Logical Units with the network address of each. DECNET, Appletalk, Novell IPX, and have a design to make the number for each local network and for each workstation connected to the network.
In the main pengalamatan local network, TCP / IP to create a unique number for each workstation in the world. IP number is the value of 4 bytes (IPv4) with the convention each byte into a decimal number (0 to 255 for the IP that is used now) and separate each with a period of bytes. As an example such as 130.132.59.234.
An organization started by sending an electronic mail Hostmaster@INTERNIC.NET to ask for the number of networks. It is possible for almost everyone to get the number for the network "small class C" with the first 3 bytes and bytes meyatakan network states last individual computer. Organizations can gain greater network "Class B" with the first 2 bytes represent the network and the last 2 bytes represent states each workstation to achieve 64,000 individual workstation. Example Network Yale Class B is 130,132, so all the computers with the IP address is linked to 130,132 .*.* through Yale.
And organizations related to the intenet via a network of some of the regional or network-specific. vendor network are numbered customer networks and added to the routing configuration in each machine.
There is no mathematical formula to change the number 192.35.91 or 130,132 to "Yale University" or "New Haven". Machinery that regulates a large regional network or Internet routers can determine the location of the central network-network search for each number with the network in the table. Estimated to have a network of thousands of millions of class B and class C network Customers connected to the Internet, even large companies such as IBM does not need to maintain information on network-jatingan the other. They send all external data to the regional carrier that they langgan, and the regional carrier and maintain routing tables and doing the right.
3. Subnets
Although the individual customer does not need the network or provide a table of explicit routing, but for most class B network can be managed internally and so the smaller version of a network organization which is simpler. Usually divide the two byte internal assignment to one byte number and one byte Department Workstation ID.
Enterprise network is built on TCP / IP router in a commercial box. each router has a table with 255 entries to change the Department number one byte to be the choice destination ethernet that is connected to a router. For example, the message to the 130.132.59.234 network through National and regional New England under the number 130132. Arriving at Yale, 59 department ID select ethernet connector. 234 select a workstation on the LAN. Yale network must be updated as new ethernet and departments are added, but not affected by changes from outside the engine or switching in the department.
4. Channel-path is not necessarily
Each time a message arrives on an IP router, the router will then make a decision to which the next message will be sent. There is one concept at a particular time, with the preselected path for all traffic. Suppose a company with facilities in New York, Los angles, Chicago and Atlanta. Network can be made of the four phone lines form a loop (NY to Chicago to LA to NY to Atlanta). A message arrived at the router can go to NY to LA via Chicago or through Atlanta. answers can be returned to the other.
How a router can make a decision between the router with a router? not have the correct answers. Traffic can be mapped with the algorithm "clockwise" (to go to NY to Atlanta, LA to chicago). Router can determine, to send a message to the next and then to Atlanta to Chicago. Routing is a better measure of traffic patterns and send the data via links to the most is not busy.
If one phone line in a damaged network, the message can still reach the destination via another path. After the loss of channels from NY to Chicago, the data can be sent to Atlanta from NY to LA to Chicago. That way the path will continue despite the loss performance decreased.
Repairs such as this is an additional section on the design of IP.
5. Not a problem Reviewed (Undiagnosed Problem)
If there are any error occurs, the network was reported to authorities. Error must be justified or corrected. IP, is designed to be resistant and strong. Node or path loss is normal, but the road network must be fixed. So IP menkonfigurasi automatically reset itself if there is something wrong. If a lot of redundancy built into the system, the komuniksi still in progress and security. TCP is designed to restore the node that failed or a channel where propagasi routing table changes to all router nodes. Because the updating process requires a long time, TCP is rather slow to initiate recovery.
6. Tax on IP
Every large company or university that is connected to the Internet must have a level intermediet network. some routers may be configured to deal with bebarapa department LAN. All traffic outside the organization associated with a single connection to regional network providers.
So, the end user can install TCP / IP on the PC without having to know the regional network. Three-part information required:
IP address is created on the PC
The IP address (subnet mask) that distinguishes other machines in the same LAN (messages can be sent directly) with the machines in other departments atao anywhere in the world (which is sent to the router machine)
IP address of the router machine that connects the LAN with the world outside.
7. Stack TCP / IP protocol
Internet initially designed with two main criteria. Two criteria influence and shape this hardware and software that is used now. Criteria: The network must make the communication between researchers in different parts of the world, enabling They can share and communicate about their research to one another. Unfortunately, the research requires a variety of different computer platforms and network architectures for different scientific purposes. So it is necessary for the protocol suite to be able to deal with the various hardware Platforms different systems and even different networks. Furthermore, the network must be a communication network that has a strong ability to survive attacks from nuclear. This design memebawa to the decentralized network that consists of a network of separate, smaller, networks that have the ability diisolasi automatically when needed.
Layer level abstrsaksi to provide software and increase the ability to reuse and platform independence. Layer-layer is meant to truly separate from one another and also independent. Layer does not rely on detailed information from the other layer. Architecture design to make this easier to do the maintenance because the layer can be developed or redesigned without damaging the integrity of the protocol stack.
TCP / IP protocol suite consists of 4 layers: Application, Transport, Internetwork, and network interface. Layer can be seen as a hierarchy such as below:
Application Layer is an application that sends data to the transport layer. For example, FTP, email programs and web browsers.
Transport Layer is responsible for communication between applications. This layer and set the context information may provide error checking. Data is divided into several packages sent to the internet layer with a header. Header contains destination address, source address, and checksum. Checksum engine checked by the recipient to see if the package is lost on the route.
Internetwork layer responsible for communication between machines. Layer this engcapsul-meg package from the transport layer in IP datagrams to and use the routing algorithm to determine where datagaram should be sent. Entrance datagram is processed and checked before
guilty through the Transport layer.
Interface layer networks is the bottom level of the order of TCP / IP. This layer is device driver that allows datagaram IP sent to or from the physical network. Dapaat the form of a network cable, Ethernet, frame relay, token ring, ISDN, ATM networks, radio, satellite or other tool that can transfer data from system to system. Network interface layer is abstraksi that facilitate communication between the multitude network architecture.
No comments:
Post a Comment