TCP/IP Reference Model

TCP/IP Reference Model

Hello welcome you all!! This is a quick short blog regarding the TCP/IP reference model in networking. Hope you guys will learn something new through this.

What is TCP/IP reference model????

Before giving an introduction on the TCP/IP model I may give you a brief history on how this architecture was emerged. This model emerged with ARPANET. ARPANET is a research network sponsored by the Department of Defence USA hence can briefly define as the beginning of internet. Through ARPANET it connected hundreds of universities and government installations via a leased line. But when other networking technologies like satellite and radio networks came in to use the existing protocol became trouble shooting. So, to overcome this problem TCP/IP reference model was introduced. TCP/IP reference model was capable of connecting multiple networks.

TCP/IP reference model is a 4 layered reference model.

Four layers of the model respectively are application layer, transport layer, internet layer and host to network layer. Protocols belong to this model are located in the top 3 layers of the model.

Application layer, which is the top most layer of the TCP/IP model gives the ability for applications to access services of the other layers and defines all the higher- level protocols. Below are the protocols defined in the application layer. TELNET- Virtual terminal protocol (protocol for accessing remote computers) HTTP- Hypertext transfer protocol (protocol used in file transfer which make up web pages in the world wide web) FTP- File transfer protocol SMTP- Simple mail transfer protocol (protocol used in electronic mail) DNS- Domain name system ( protocol used in conversion of ip addresses to domain names)   Transport Layer This layer is responsible of accepting data from application layer and splitting them in to smaller parts and passing to the lower level.  Transport layer also builds up end to end connection between the source and the destination. It means a conversation is carried out between  2 similar programs on the source and the destination points. The transport layer is defined with 2 end to end protocols as, TCP- Transaction Control Protocol UDP- Universal Datagram Protocol Both TCP and UDP protocols run on top of the IP (Internet Protocol) TCP TCP can be defined as a reliable connection oriented protocol.  Why is TCP known as a reliable protocol?  There are basically 2 reasons ·       It allows byte stream to reach the destination point without an error ·     It maintains flow control So let’s see how TCP allows the byte stream to reach the destination point without an error.  When a byte stream (message) flow from top layer it fragments them in to fixed chunks (datagrams- short packets of data) and pass them as a sequence  to the internet layer. When they reach the destination point it reassembles them making sure the message is received without an error in the proper order. And another big advantage of TCP is the flow control. This means  a fast sender can’t overload a slow receiver with more than the maximum number of messages it can handle. It dynamically sets the best rate to send. Now let’s move on to the 2nd type protocol in this layer. UDP Unlike TCP, UDP can be called as an unreliable, connection less protocol meaning that no guarantee for delivery. This protocol is primarily used to establish low latency.  UDP doesn’t provide with sequencing and flow controlling like TCP and hence gives a prompt delivery to the receiver. This protocol is mostly used to transmit speeches and videos. The below figure shows how data and header information is arranged  in a UDP datagram. An application which uses UDP is DNS(Domain Name System) .So let's see an example....  Suppose www.ucsc.lk sends a UDP packet containing the host name to the DNS server. DNS server replies with a UDP packet containing the host’s IP address. Another fact is that in IPv4 UDP has a maximum payload of 65,507 bytes and in IPv6  65,527 bytes. Internet Layer This is the third layer in the tcp/ip model.  Protocols  mainly used in here are IP(Internet Protocol) and ICMP(Internet Control Message Protocol). This layer is mainly responsible of accepting data from the transport layer and transmitting them to the transport layer. Another fact is that this layer is also responsible of routing data in to the correct destination. If multiple paths are available to reach the destination it selects the shortest possible path. This shortest path which is selected by a datagram is known as the “route”. If the route is having any issues of  transmission it selects an alternative path. Host to Network Layer The bottom most layer is the Host to Network layer. This layer which gets the datagram from the internet layer is responsible of receiving and sending data from the network. It figures out whether the receiver is within the sub net or not. If it is not within its sub net it sends datagram to the router. There after the router is responsible of sending the datagram to the receiver Hope you gained some understanding on the tcp/ip reference model through this article. More to talk about the tcp/ip model and the ISO model.............KIT. :)