The layers in the TCP/IP model

The TCP/IP model comprises four layers, as shown in the following diagram. Each layer uses a different set of protocols allocated to it. Every protocol has specific designated roles, and all of them are designed in such a way that they comply with industry standards.

The first layer is the Application Layer that directly interacts with users and other network-level protocols; it is primarily concerned with the representation of the data in an understandable format to the user. The Application layer also keeps track of user web sessions, which users are connected, and uses a set of protocols, which helps the application layer interface to the other layers in the TCP/IP model. Some popular protocols that we will cover in this book are as follows:

• The Hyper Text Transfer Protocol (HTTP)
• The File Transfer Protocol (FTP)
• The Simple Network Management Protocol (SNMP)
• The Simple Mail Transfer Protocol (SMTP)

The second layer is the Transport Layer. The sole purpose of this layer is to create sockets over which the two hosts can communicate (you might already know about the importance of network sockets) which is essential to create an individual connection between two devices.

There can be more than one connection between two hosts at the same instance. IP addresses and port numbers together make this possible. An IP address is required when we talk about WAN-based communication (in LAN-based communication, the actual data transfer happens over MAC addresses), and these days, a single system can communicate with more than one device over multiple channels which is possible with the help of port numbers. Apart from the restricted range of port numbers, every system is free to designate a random port for their communication.

This layer also serves as a backbone to the communication between two hosts. The most common protocols that work in this layer are TCP and UDP, which are explained as follows:

• TCP: This is a connection-oriented protocol, often called a reliable protocol. Here, firstly, a dedicated channel is created between two hosts and then data is transferred. Then, the sender sends equally partitioned chunks, over the dedicated channel, and then, the receiver sends the acknowledgement for every chunk received. Most commonly, the sender waits for a particular time after which it sends the same chunk again for assurance. For example, if you are downloading something, TCP is the one that takes care and makes sure that every bit is transferred successfully.
• UDP: This is a connection-less protocol and is often termed an unreliable form of communication. It is simple though because there is no dedicated channel created, and the sender is just concerned with sending chunks of data to the destination, whether it is received or not. This form of communication actually does not hamper the communication quality; the sole purpose of transferring the bits from a sender to receiver is fulfilled. For example, if you are playing a LAN-based game, the loss of a few bytes is not going to disrupt your gaming experience, and as a result, the user experience is not harmed.

The third layer is the Internet Layer, which is concerned with the back and forth movement of data. The primary protocol that works is the IP (Internet Protocol) protocol, and it is the most important protocol of this layer. The IP provides the routing functionality due to which a certain packet can get to it's destination. Other protocols included in this layer are ICMP and IGMP.

The last layer is the Link Layer (often termed as the Network Interface Layer) that is close to the network hardware. There are no protocols specified in this layer by TCP/IP; however, several protocols are implemented, such as Address Resolution Protocol (ARP) and Point to Point (PPP). This layer is concerned with how a bit of information travels inside the real wires. It establishes and terminates the connection and also converts signals from analog to digital and vice versa. Devices such as bridges and switches operate in this layer.

The combination of an IP address and a MAC address for both the client and server is the core of the communication process, where the IP address is assigned to the device by the gateway or assigned statically, and the MAC address comes from the Network Interface Card (NIC), which should be present in every device that communicates with other hosts. As data progresses from the Application layer to the Link Layer, several bits of information are attached to the data bits in the form of headers or footers, which allow different layers of the TCP/IP model to coordinate with each other. The process of adding these extra bits is called data encapsulation, and in this process, a Protocol data unit (PDU) is created at the end of the networking model.

It consists of the information being sent along with the different protocol information that gets attached as part of the header or footer. By the time PDU reaches the bottom-most layer, it is embedded with all the required information required for the real transfer. Once it reaches the destination, the embedded header and footer PDU elements are ripped off one by one as it passes through each and every layer of the TCP/IP model as it progresses upward in the model.

The following figure depicts the process of encapsulation:

Figure 1.1: Data encapsulation