When several computers are connected to one another it is known as a network. Networks enable data exchange between different devices, making shared resources available. Different network types are implemented depending on which transmission techniques and standards are used. These differ in terms of the number of connected systems and potential reach.
The computer is started, the browser is opened, but instead of the start page, all you see is an error page telling you that the internet connection couldn’t be established. Has this ever happened to you? Checking the router and Ethernet cable doesn’t turn up any problems, and only when you run an error diagnostic do you find the answer: At least one network protocol is missing on the computer. In some cases, running an automatic repair and rebooting will fix the error. But often the protocol is also missing afterward, or the error message appears again after a certain amount of time. Possible causes can be, for example, outdated drivers, or complications with the anti-virus software or firewall being used.
So why do network protocols play such an important role in establishing an internet connection? In order to understand this, it helps to more closely identify the different types of network protocols and their various functions.
Cheap domain names
Domains as original as your ideas.
What are you waiting for? Make your project a success with the perfect domain extension!
What is a network protocol?
To combine computers in a network, it seems like it would be sufficient to just use an Ethernet cable. But without help, the computer systems aren’t capable of exchanging data packets and so can’t establish a data connection. That task is instead carried out by the network protocols, which are connected to their respective protocol family on the so-called switching layer, or network layer (layer 3 according to the OSI model). They contain agreements for the data exchange and so regulate the conditions for its subsequent transport, addressing, routing, and error checking. For two computers to communicate with each other, they need the same network protocols. They agree on the following conditions for the transmission, which are either placed in front of the respective package as a header, or attached as a trailer:
- Size of the data package or packages
- Type of package
- Sender and addressee
- Other involved protocols
Why are there different network protocols?
Not every data connection between computer systems is knitted in the same pattern. As a result, it makes a difference whether you are connecting two computers in your home network or over the internet as part of a huge computer network and sending data to multiple addressees. Similarly, the communication hierarchy of the participants plays a big role. That’s why various network protocols exist for individual forms of communication. These protocols have the following possible application scenarios and distinguishing features:
- Number of communication partners: Different network protocols can be distinguished by the number of computers allowed to connect to them at any given time. If, for example, transmitted data is addressed to a single receiver, then it’s a unicast transmission. An exchange between more than two communication partners happens via multicast connections. Sending data packets to all network users is called broadcasting – this type of connection is obviously best known for radio and television.
- Path of the data flow: The direction in which the data moves is another characteristic that separates different network protocols from one another. Protocols with simplex transmission are one-sided communications, where one computer acts only as the transmitter and the other only as the receiver. For half-duplex transmissions, communication partners can exchange data alternately. Full-duplex transmissions allow data to be sent simultaneously and in both directions.
- Hierarchy of the communication partners: Certain types of connections, like the client-server model, are based on a clearly defined hierarchical structure. In that case, for example, different clients initiate the connection to a single server, which then processes the requests. Symmetrical communication, the counterpart to this example’s asymmetrical communication, is marked by either peer-to-peer or computer-to-computer connections. In this structure form, all computers have equal rights and so can both offer services and use them.
- Synchronisation of communication: Data exchange can also be differentiated by whether individual bits are synchronised between transmitter and receiver (synchronous communication) or not (asynchronous communication).
- Type of connection: Network protocols can also be divided into connection-oriented and connectionless protocols. Connection-oriented protocols require the connection between sender and receiver to exist for the entire duration of the transmission, and try to ensure that the data arrives in a certain order and is retransmitted in the event of a failure. Connectionless protocols do without the connection set-up and removal, so their data packets contain substantially less information. But the data can also arrive at the receiver in a random order, and is not automatically retransmitted in the event of incorrect transmission.
Apart from the technical background, the variety of network protocols also results from the fact that many manufacturers have developed their own protocols or protocol stacks for their devices.
Which network protocols are there?
For the network layer, as for all the other layers, there are a number of standardised but also proprietary protocols which are suitable for different application areas or are partially limited to specific operating systems or devices. Many of these protocols are no longer active today, mainly due to the increased distribution of the internet protocol family. These stacks with more than 500 protocols also contain the most important and well-known network protocol IP (Internet Protocol), which is the basis of the internet.
The internet protocol has the job of transporting data packets from a sender to a receiver over multiple networks. To do this, it defines the guidelines for addressing and routing, or the finding of data packets. IP is not only the standard network protocol for wide area networks (WANs) – the individual, worldwide networks that connect the internet to each other – but also for local networks. It’s supported by all manufacturers and operating systems, but also requires the necessary know-how in terms of configuration as well as the appropriate hardware (router).
The following table shows a historically important network protocol overview:
|Network protocol||License||Platforms||Description||Functions||Weak points|
|IP (Internet Protocol)||Free (RFC 791 / 2460)||Cross-platform||Connectionless; Address length: 128 Bit (IPv6) / 32 Bit (IPv4)||Routing, addressing||Very extensive protocol stack; Security functions not immediately implemented (IPv4)|
|ARP (Address Resolution Protocol)||Free (RFC 826)||Cross-platform||Link between layers 2 and 3 with its own cache function||Address mapping (MAC assignment to IP addresses) for IPv4||No connected assignment correction check: Risk of ARP spoofing|
|NDP (Neighbor Discovery Protocol)||Free (RFC 4861)||Cross-platform||Link between layers 2 and 3 with its own cache function||Address mapping for IPv6; Identification of double IP addresses||Protection against spoofing not standardly integrated (requires SEND extension)|
|ICMP (Internet Control Message Protocol)||Free (RFC 792)||Cross-platform||Independent component of IPv4||Exchange of information and error messages||Can be targeted/used for Dos/DDoS attacks|
|SNA (Systems Network Architecture)||Proprietary (IBM)||IBM devices||Former hierarchical network structure with different protocols||Connects computers and their resources to SNA networks||Connection to non-SNA networks difficult; Comparatively high costs|
|NBF (NetBIOS Frames Protocol)||Proprietary (Microsoft)||Windows (up to and including 2000)||Former protocol for Windows systems||Communication with display layer (NetBIOS) and fuse layer (LLC)||Not routable; Only suitable for small networks (up to 20 computers)|
|IPX (Internetwork Packet Exchange)||Proprietary (Novell)||NetWare (no long supported network operating system), Linux, Windows||Connectionless protocol, functionally similar to IP; Address length: 80 bits (48 bit host address, 32 bit network number)||Routing, addressing||Unsuitable for larger WAN networks|
|DDP (Datagram Delivery Protocol)||Proprietary (Apple)||AppleTalk-supporting devices (up to and including Mac OS X 10.6)||Part of discontinued AppleTalk protocol stack; Connectionless; Address length: maximum 13 byte header and 587 byte user data||Routing, addressing||Risk AppleTalk-Spoofing; weak data throughput due to small packet size|
|OSPF (Open Shortest Path First)||Free (RFC 2328)||Cross-platform||Network protocol based on Dijkstra algorithm, particularly suitable for large company networks||Optimises routing with regard to transmission costs; Dynamic load distribution||High processor utilisation and memory requirements; Complex configuration and maintenance|
Transmission protocols take over the transport of data packets
After the protocols of the switching layer have established the basis for communication, another protocol is needed so that the data packets reach the corresponding applications. With the OSI model, this forwarding is carried out on the transport layer (layer 4). Each stack also has its own protocols. For the internet protocol family, these are particularly
- TCP (Transmission Control Protocol)
- and UDP (User Datagram Protocol).
Since the great success of the internet, the first mentioned TCP is equal to IP as a standard for network connections. In most cases it builds directly on IP, which is why TCP/IP networks are often used. As a connection-oriented protocol, TCP requires an existing connection between the communication users for the transport of data packets. It guarantees reliable transport of the data and that all of the packets will arrive complete and in the correct order. To do this, the protocol adds additional information, such as a sequence number and proof sum to the data.
UDP is the TCP counterpart of the internet protocol family for the simple and quick transfer of smaller data packets without a connection. UDP connections don’t offer any security for a packet arriving at the addressee, but thanks to the low administration data (additional information in the header), there isn’t a clear speed advantage for data transfer where smaller transmission errors aren’t a problem. For this reason, the User Datagram Protocol is used for audio and video streaming, DNS queries, and VPN (Virtual Private Network) connections.
Like the internet protocol family, other protocol stacks also have specific transmission protocols based on their network protocols and largely similar to TCP. Novell networks, for example, wait in the transport layer with the protocol SPX (Sequenced Packet Exchange). With the AppleTalk stack, the data packets can be transported using the ATP (AppleTalk Transaction Protocol).