Network Topology: You Have to Know This Before Building a Network

NAKIVO
10 min readOct 12, 2021

--

If you’re looking into building a network for your business, you need to choose the right network topology. To do so, you should become familiar with the essentials of network topology and its types. The type of network topology you use affects such critical variables of your network as cost, scalability, and maintenance. This blog discusses the main types of network topologies.

Network Topology in a Nutshell

A network topology has physical and logical aspects. Physical topology has to do with the structure of the parts of a network that you can touch (overall). It determines how network devices, such as computers, routers, and stations connect to form a network. Logical topology is more concerned with getting data wherever it needs to go.

Physical topology involves:

  • Geometric configuration
  • Connections and interconnections
  • Location of devices
  • Number of network adapters used
  • Network adapter types
  • Cable types

Logical topology explains how data is transferred and received.

Logical topology covers:

  • Data movement from station to station
  • The transmission path and appropriate protocols
  • Cloud and virtual networks

Point-to-Point Network Topology or P2P

This type of topology is the simplest of all. In Point-to-Point topology, there are only two computers or network devices that interconnect.

A typical P2P connection scenario: two computers that have Ethernet network adapters with RJ-45 ports connect with the help of a twisted pair cable, such as:

  • UTP Cat 5e
  • FTP Cat 5e
  • STP Cat 5e

To connect two similar computers that use Ethernet cards, use the Ethernet crossover cable of category 5e. This cable has:

  • RJ-45 connectors on both ends
  • T568A wiring on one end
  • T568B on the other end

Note: T568A and T568B are color code termination standards that are used to place conductor wires correctly during installation. In a standard Ethernet cable, either T568A or T568B are used on both ends. However, the crossover cable uses the T568 standard on one end and the T568 standard on the other end (see the image below). An example of standard connection is a computer plugged into a wall. An example of a crossover connection is a computer connected directly to another computer (as it is in Point-to- Point Topology).

You should use the Ethernet crossover cable of category 5e to connect similar network devices such as two computers that use Ethernet cards. If you have a modern network card, you can use a patch cable without a crossover cable. This type of connection is possible due to Ethernet Auto MDI-X support. Use a patch cable to connect your network card to a switch and then connect switches. The ends of a patch cord are crimped with either T568B or T568A. However, the T568A standard is less common.

We have discussed the most simple type of topology, the Point-to-Point topology. Now let’s continue to review other types of network topologies, starting with the Bus Network Topology.

Bus Network Topology

A bus network has one main cable. You connect stations to the main cable via shorter cables known as drop lines and use a tap device to connect drop lines to the main cable. To build a bus network, use RG-58 coaxial cable.. Usually, instead of direct bare-wire connections, you use connectors such as a BNC (Bayonet Neill-Concelman) connectors for connecting parts of the network and a cable to a network card. To avoid signal reflection, install devices that absorb signals called terminators on both ends of the cable. Signal reflection should be avoided at all times because it can slow down the whole network.

A bus topology is cost-effective, and it utilizes less cable compared to other types. You tend to use a bus topology for small networks because its scalability is pretty low. The main cable has a limited length, which makes it impractical to connect more than a certain number of stations to the main cable. Overall installation difficulty for the bus topology is average. Moreover, if a dropline or one station becomes corrupted, it doesn’t affect the whole network. However, if there is an issue with the main cable, the whole network goes down. Data in the bus topology is sent only in one direction, in a half-duplex mode.

Note:

Half-duplex mode: stations can’t transmit and receive data simultaneously. When one station sends data, others can only receive it.

Full-duplex mode: allows simultaneous data transmission for both stations.

Therefore, when one station sends a packet to another station, the packet is sent to all stations. However, only the target station can receive it after the destination MAC address is verified. A network that works in a half-duplex mode is prone to congestion and can be impractical.

Token Bus

A Token Bus network utilizes a Token Ring protocol on a coaxial cable. A token is passed from station to station, and only the station that has the token can transmit frames in the network. A token is passed sequentially in a clockwise or counter-clockwise direction. The Bus Topology network employs IEEE 802.4 standard for building a logical token ring. IEEE 802.4 is currently disbanded by IEEE (SOURCE).

A bus network topology is not a great choice for networks with high traffic. This type of topology was popular in the 90s, and its maximum speed is 10 Mbit/s. Today there are much faster network topologies. Therefore, a bus network topology is not the best choice for building your network today.

Ring Network Topology

A ring network is similar to a bus network, but it has a circular structure — every station is connected to the other station on each side, and the first station is connected to the last. A ring network operates in a half-duplex mode, which means that data in your network travels sequentially in one direction. A ring network also uses coaxial cables similar to bus network topology but it is faster than the bus network.

Token ring

The token-based topology implements the IEEE 802.5 standard (currently disbanded by IEEE). The token topology was suggested by IBM back in 1984. In the token ring topology, a token passes around the ring. Only the node that has the token can transmit data. The token ring topology has a medium installation difficulty — you only need to change two links to add or remove a network device, and the installation process is affordable. However, the token ring topology has many disadvantages such as:

  • Each fragment of ring topology is a point of failure
  • The whole network fails if the signal can’t pass the point of failure
  • Data has to go through all nodes before reaching its destination
  • Adding new stations can slow down the network because all stations in the network share the same bandwidth.
  • If you need to add or remove the node, the whole station should stay online

The ring network was popular in the 90s. However, today it’s not used because of its fragility and low network speed of 4–16 Mbit/s.

Dual ring

In the dual ring topology, there is a second connection between the nodes in the ring, which enables data transfer in both directions. Data travels in a clockwise and counterclockwise direction, and the networks work in a full-duplex mode. If one ring fails, the other one can take over to avoid an interruption in data transmission.

Optical ring

The optical ring structure involves expensive equipment, such as optical cables and switches. You can use this type of topology to connect stations located in different localities (different districts or cities) to form a high-speed loop. The optical ring topology is used in modern networks by internet service providers (ISP) and managed service providers to create connections in wide area networks.

Star Network Topology

The star topology is one of the most convenient. It is used by many organizations today. The star topology has a switch which is a central unit of the network. All other devices connect to this switch via a network cable. The central switch has numerous ports that allow stations to interact. All communication goes through the switch. When two stations interact, the data protocol from a sender is sent to the switch, and this switch then sends the data to the network card of the intended station.

A star topology is convenient because it’s easy to scale. You can always get a switch with more ports or add a second switch to the existing one. A star network has a practical structure. If one of your connected units fails, all others remain functioning. However, if the switch becomes corrupted or goes offline, the entire network is affected. Your star network can operate in both full-duplex and half-duplex modes allowing for single-direction and simultaneous data flow. You can use twisted-pair cables or optical cables to establish connections between the stations and the switch.

Tree Network Topology

A tree topology is a more complex version of a star network. It allows you to bring together multiple star networks by connecting their switches. Every switch has a port, and you can connect several stations to those ports. Then you can connect those switches directly to the main switch.

A tree topology eliminates a single point of failure. If one of the switches malfunctions, only a particular segment goes offline. If the main switch has an issue, the connection between network branches becomes compromised. However, there is still a connection between individual machines within a branch.

A tree network is a good fit for a large company with multiple offices throughout a building. In this scenario, you can create a single main switch in a separate server room. Every department can have its own branch connected to a switch. All of the branch switches are connected to the main switch in the server room. For convenience and wireless connectivity, you can connect your main switch to a router. You can also create a subnet and assign a router for every single business department.

Mesh Network Topology

In a mesh network topology, every station is connected to the other station. There are partial mesh and full mesh networks. A partial mesh has at least two stations connected to other stations in the network. In a full mesh, every station is connected to all other stations. Mesh networks ensure a continuous data transfer. If one station fails, other stations can still transfer your data. Setting up a mesh network can be expensive. You would need to create multiple connections and purchase a lot of cable.

Mesh network topology is hard to scale. It’s not simple to add an extra station because you need to add a new station to all other stations in the network. And, you may also need to add additional network cards to the stations.

Generally, a mesh network requires having multiple cards per station to support all the connections. A mesh topology could utilize WiFi to supply extra network coverage. Such wireless networks are called wireless mesh networks.

Hybrid Networks

Hybrid networks are pretty popular nowadays. They involve a combination of two or more network topologies. But while hybrid networks can meet your unique needs, they have a few downsides. Hybrid networks are more complex and can be harder to install. They may also require extra cost. An example of a hybrid topology is a star-ring network. This network combines ring and star elements. This type of network implements high-quality fiber cables for making fast connections between buildings and cities. A fiber ring with numerous lines eliminates a single point of failure. Hybrid networks can be complex and hard to install, but they can be well worth it.

Conclusion

Since the 1970s, network topologies have evolved from simple P2P and bus architectures to complex hybrid systems. A star topology is a basic unit in modern hybrid networks. Building a network requires a lot of time and effort. You should estimate the current state of your company in order to choose a network topology that can best fit your needs. Before setting up a network of your choice, weigh all the pros and cons. Think about the size of your business and all related costs.

Having a well-functioning and cost-efficient network is one of the main steps toward success. Learn more about network topologies and good luck with building your network!

Once you decide on the type of network, you can connect your infrastructure, including your virtual machines (VMs), to that network. When you have your network and infrastructure up and running, don’t forget about protecting your data. You can secure your data by scheduling regular backups. NAKIVO Backup & Replication offers comprehensive all-in-one data protection for your physical and virtual environments.

--

--

NAKIVO

NAKIVO is a US-based corporation dedicated to developing the ultimate VM backup and site recovery solution: https://www.nakivo.com