ARTICLE: Computer Networks: An Overview

In the computer networking field, the order of presentation of topics is not well agreed. It means that there are several interconnections in this field and all the topics involve “forward references” to the following topics and chapters. So, you would find largely bottomed ordering all the concepts and the chapters of computer networking. There is a series of articles in which you would find the difficult concepts explained in easy to comprehend language.

In this blog read about the layers, packet-based switching, routing, and the other important material. Let us begin studying about the layers in the computer networks.

Layers of computer networks

The layers in this context constitute of the three topics; these are the LAN, IP, and TCP. In these, the composition of every topic is done by the Link Layer, Transport Layer, and the Internet Layer. When the layer in which the software functioning takes place includes in this, it becomes a four-layered model. The name of that layer which modifies the layers of the computer networks is “application layer.” This layer is the one who is responsible for corresponding to the library and programming interface idea. In addition to this, it is the layer in which we understand that the communication of each layer is done with only two layers at a time. These are the layers which are in direct contact with that particular layer. Yes, the one which is above it and the second which is below it. The process begins with handing off the chunk of data to the TCP library by an application. Then the application makes calls to the IP library, which in further turn into calling the LAN layer. It is the layer which results in the actual delivery of the signals from one layer to the second. But, remember one thing that the application never makes direct communication with the LAN and IP. Not at all.

Computer Networking layer which is the in charge of the task of making the actual delivery of the packets is the LAN layer. It uses the LAN-layer supplied an address for the delivery. There is a conceptual subdivision of LAN layer. It is in the form of a physical and logical layer.

Physical Layer

It is the layer which deals with the analogue electrical, radio or optical signals and the involved mechanisms. It is the layer which has the direct concern with the designing the hardware of the LAN. The hardware consists of the kernel software interface to the LAN. The logical LAN layer is corresponding to the physical layer.

Logical Layer

It is the layer above the physical layer in the LAN layer. It deals with the remaining all the digital, non-analogue for the operations on packets.

With the subdivision of the LAN layer, the computer networking discipline gets a five-layer model. It is more of an ad hoc classification of the model, rather than being a formal hierarchy.

Let us see how we reach to the seven-layer model of the computer networking layers.

Data Rate

The data rate is the rate which measures the number of bits transmitted per unit time. Every computer networking connection has a data rate.

Throughput

The overall and the effective rate of the transmission packets which considers the protocol inefficiencies, transmission overhead, and the competing traffic as well is referred to as the throughput. The layer in which the throughput is measured is higher than the layer measuring the data rate.

Bandwidth

There can be two definitions of the bandwidth. But in this context, we would use the word as a data rate synonym. The field from which the term has come from is the radio transmission. In the radio transmission, the available band frequency is proportional. The rest of the parameters are equal to the achievable data rate.

Goodput

It is a term used in the discussion of TCP layer. The other name of the term can be “application-layer throughput.” It is the amount of the data which has been used and delivered to the receiving application. The data which has been retransmitted is calculated by goodput once. But when throughput makes some interpretations, it gets counted twice.

Units

The units of the data rates are the kilobits per second (Kbps). Mbps, i.e. megabits can be used to measure the amount of data transferred. Here is a confusion which most of the students have. It is the value of the bits in the context of data rates. It is 10^3 here in this case, which the students think that it is 2^10 as in the other concepts of binary forms. The units like KB and MB are used for the denotation of data volumes. The value of these is 2^10 and 2^20 in the context of volume. Here, a byte is the full form of B in this acronym.

There is an introduction of new terms, or we can say units, these are KiB and MiB. But these can unfold many confusions.

Packets

The modest-sized buffers of data are called as packets. These are transmitted as the unit of sharing the set of links. The header which contains the delivering information prefix the packets. The destination address in the common case known as the datagram forwarding is with the header. In the virtual-circuit forwarding network, the header contains an identifier for the connection. The networking we are using, and the one which we will be using is packet-based. For the voice-telephony, the circuit-based switching options are used.

The imposition of the buffer and address at the top of the low-level series lines is the view in which the packets can be seen. The additional layers impose the additional structure. The informal name of the packets is framed in the LAN layer. Whereas in the Transport layer, segments are the name which is given to the packets.

Size of the packets

Every LAN has an intrinsic attribute which denotes the maximum size of the packet which a given LAN supports. In the Ethernet, 2500 bytes of data is the maximum value of data. But if we talk about the TCP and IP packets, 512 bytes is the original data value held. While 4KB is the size of Token Ring packets. There are several proponents which contain data more than 64KB. The protocol, ATM, Asynchronous Transfer Mode uses 48 bytes of data per packet. The size of the packet believes the modest packet sizes.

Every layer has its header. The Ethernet headers are of 14 bytes; TCP headers are of 20 bytes. The IP headers are of 20 bytes. The size of the TCP connection is sending 512 bytes’ data per packet. A not-unreasonable overhead, the headers are amounting to more 10% of the total data.

The address of the destination and the other delivery information are the components of the data-forwarding packets. The names of the nodes which are linked internally are called the routers and switches. These ensure that there is the delivery of packets to the requested destination.

Paul Baran is the pioneer in the field and is the one who has used the concepts of packets and switching of packets in 1962. He had talked about the node failure and the network survivability. In addition to this, his prime concern was the central switching of the existing protocol which was vulnerable to central failure. The term, the packet was used by Davies in 1964.

The standard and the universal size of the packets have become 8-bit bytes. Now the term octet was used by the internet specifications. The packets are the sequences of non-octet-oriented hosts, and octets.