Describe different network topology types (MPLS was covered in the previous chapter). This chapter covers in detail the different network topologies, including Point-to-Point, Point-to-Multipoint, Ring, Star, Mesh, Bus, Peer-to-Peer, and Client-Server. Different topologies have different types of configurations, as the traffic can pass from a source to a destination following multiple paths. Understanding these topologies and the data flows within the network can be useful in many phases, such as the following:
- Network design
- Network implementation
- Network troubleshooting
We cover network topology types in our CompTIA Network+ video course.
Network Topology Design Considerations
When deciding on the topology to use in the network, an effective approach involves a structured design that allows you to develop a complete system with the optimum design and the lowest cost, while meeting all of the following requirements, which are also illustrated in Figure 17.1 below:
- Performance
- Functionality
- Flexibility
- Capacity
- Availability
- Scalability
Figure 17.1 –Network Topology Design Solutions
The goal in this phase should be to develop a systematic approach that takes into consideration the business’s needs, the organizational goals, policies and procedures, technical goals and constraints, and the existing and future network infrastructure. This includes physical models, logical models, and functional models.
The top-down approach is the recommended methodology for a medium-sized network to a large enterprise campus design. Using this approach will give you the big picture and all the aspects of the design before getting down to the design details. This basically means beginning with Layer 7 of the OSI reference model and then moving down from the Application Layer to the Presentation Layer, Session Layer, Transport Layer, Network Layer, Data Link Layer, and, finally, the Physical Layer.
The network and the physical infrastructure should be adapted to the needs of the network applications and services. In other words, you shouldn’t choose your network devices or your hardware and software technologies until the application requirements have been fully analyzed and met.
The top-down approach is usually a very time-consuming process and is a bit more costly, but it is preferred over the bottom-up solutions, where the design is typically based on previous experience and you are just looking for a quick fix or solution. The problem with the bottom-up approach is that most likely you will get an inappropriate design in the medium-term to long-term and the organizational requirements and constraints are not included. This would cause trouble and possible process rollback at later project phases. Figure 17.2 below presents an example of the network design top-down approach methodology:
Figure 17.2 – Network Design Top-Down Approach
Figure 17.2 starts at the top with the applications and services, which include the Application, Presentation, and Session Layers. Based on the application requirements and needs and the way they map to the organizational goals, you will apply a network infrastructure design to meet the application and service requirements of the organization. This includes the data, the type of traffic, and services, and then what type of design and network services will meet the needs of those applications.
Once these goals are met, the network should be modularized, including the Data Center, the Server Farm, the Branch, the Access, Distribution, and Core Layers, and the Internet Connectivity modules and submodules. After the network is modularized, you can then apply the decisions made for infrastructure and services to different modular areas of the network, dealing with certain segments of the network at a time.
The next step is to take this modular implementation and create logical subdivisions that will be addressed on a project-by-project basis. If you look at them from a project management or steering committee standpoint, these will be logical subprojects. Different subprojects might exist for the following:
- Choosing the technology, acquisitioning, and provisioning
- Physical topology design (placing the design at different layers)
- Addressing design scheme, including NAT solutions
- Routing selection and design
- Quality of Service design (traffic management)
- Security design
- IP Multicast design (for video and audio streaming)
- IPv6 provisioning design
Point-to-Point
A Point-to-Point topology involves a direct link between two devices (this is also called a one-to-one connection). This type of connection is often utilized in WAN topologies, especially in some of the legacy Point-to-Point technologies like T1/E1 and T3/E3. Those types of connections go directly from one router to another router, as shown in Figure 17.3 below:
Figure 17.3 – Point-to-Point Topology
You will also see point-to-point connections between buildings in a campus, which are connected using fiber links that are directly attached to a network device on each side.
Point-to-Multipoint
Point-to-Multipoint is a type of network topology that is very popular today. It is often used in wireless networks and functions by having multiple stations communicating with a central device. The Point-to-Multipoint topology is shown in Figure 17.4 below:
Figure 17.4 – Point-to-Multipoint Topology
However, a Point-to-Multipoint topology does not imply that every device can actually communicate with every other device in the network. For example, in a wireless environment the end-stations are usually allowed to communicate with the access point but are not allowed to communicate with each other.
Ring
Ring topologies involve connecting all nodes in a ring fashion to ensure redundancy (see Figure 17.5 below). If one link between two nodes fails, traffic will still have a valid path to reach any devices on the other side of the ring.
Ring topologies were popular in LAN environments in the past, one solid example in this regard being Token Ring architectures. However, they are rarely (if at all) used in modern networks on the LAN side.
Figure 17.5 – Ring Topology
However, you will still find Ring topologies in large-scale networks, and in MAN (Metro Area Network) and WAN (Wide Area Network) architectures, mainly because of their redundancy features (using dual rings), which are critical in a WAN environment.
Star
A common LAN network topology used in modern networks is the Star topology. This design is used in all sizes of networks and involves all network devices having a direct connection to a central concentrator device that passes the data flows between all other nodes, as illustrated in Figure 17.6 below:
Figure 17.6 – Star Topology
Ethernet networks are a good example of this topology, as they use central switches to connect end-stations. If a station wants to communicate with another station on the network, it sends the data flow to the central device, which switches it out to the other station.
Bus
Bus networks are not really used in modern networks because they involve great complexity and are subject to errors. Bus topologies are usually built using coaxial cables and they use a single main link that aggregates all nodes on the network, as shown in Figure 17.7 below:
Figure 17.7 – Bus Topology
The main advantage of a Bus topology is that any station can easily connect to the network. However, if there is a break anywhere on the main bus, network communication will go down. This lack of redundancy is one of the main reasons this type of topology is no longer used.
WAN-Specific Topologies
WAN topologies are characterized by redundancy and are categorized as follows:
- Hub-and-spoke (remote nodes connect to a central device)
- Full-mesh (all nodes are connected to each other)
- Partial-mesh (hub-and-spoke, with additional redundant links between nodes)
Full-mesh topologies require a large number of nodes, which adds extra overhead. If you have “n” nodes you will need n*(n-1)/2 connections. For example, if you need to connect four nodes in a full-mesh topology, that will require six different connections. Full-mesh is the best option when considering availability and reliability because if there is any kind of failure, failover will occur on the other links/devices. The downside of the full-mesh topology is the extra overhead associated with building and maintaining all the connections, as well as the high costs required to install all the links.
Figure 17.8 – WAN Hub-and-Spoke Topology
The hub-and-spoke topology, illustrated in Figure 17.8 above, is one of the most popular WAN topologies. The hub router is usually located at the headquarters location and it offers connectivity to branch office routers. These various branch offices connect in hub-and-spoke fashion. The hub-and-spoke topology is not so great when it comes to redundancy and availability. The key location and the most usual point of failure is the hub device. In order to achieve some form of high availability, the hub device and/or the connections between the hub and the spokes should be duplicated. Hub-and-spoke topologies are less complex and less expensive than full-mesh topologies.
Note: In a hub-and-spoke topology, the minimum number of required connections equals the number of spokes. |
Another possible WAN topology is partial-mesh and this involves a combination of full-mesh and hub-and-spoke areas within a larger area. In terms of availability and costs, the partial-mesh topology offers a balance between the full-mesh and the hub-and-spoke topologies. This is useful when a high level of availability and redundancy is required only in some areas (e.g., for certain nodes).
Hybrid Topologies
Based on the business’s needs, most network topologies are not standardized to one single network topology. Instead, they use different topologies in different parts of the network. For example, a single infrastructure could combine Mesh, Bus, Star, and Point-to-Multipoint topologies, as shown in Figure 17.9 below. Such topology combinations grow in number as the size of the network grows.
Figure 17.9 – Hybrid Topology
Client-Server
Applications generally use two different technologies to operate:
- Client-Server
- Peer-to-Peer
The Client-Server architecture generally functions by having multiple clients communicate with a central server that offers different services on the network, as shown in Figure 17.10 below. In such an environment, you generally don’t have any type of client-to-client communication.
Figure 17.10 – Client-Server Topology
Examples of applications that function in a Client-Server architecture include:
- Web services (central Web server)
- FTP services (central FTP server)
- DHCP services (central DHCP server)
The advantages of such a topology include:
- Increased performance, as the central server is usually designed to operate at high speeds
- Single point of administration for the application, as everything is hosted by the central server
Some of the disadvantages of such a design are:
- High investment cost on the server side (performance comes with a cost)
- High maintenance cost for the server
- Increased complexity, especially in large environments.
Peer-to-Peer
A Peer-to-Peer topology works completely different from a Client-Server topology, as every device communicates with every other device and they all act as both servers and clients, as illustrated in Figure 17.11 below:
Figure 17.11 – Peer-to-Peer Topology
An example of an application that functions in a Peer-to-Peer topology is peer-to-peer file sharing (e.g., torrents), where each one of the devices can contain files and every device can start downloading different parts of that file from every other device. This kind of application usually spans across a large geographical area. The advantages of a Peer-to-Peer architecture are as follows:
- Easy to deploy
- Low cost of implementation
The disadvantages of such an implementation are as follows:
- Increased administration, as there is no single point of maintenance like with the Client-Server architecture
- Less secure than Client-Server topologies, as multiple points are offering services and they can become vulnerable
Summary
Different topologies have different types of configurations, as the traffic can pass from a source to a destination following multiple paths. Understanding these topologies and the data flows within the network can be useful in many phases, such as:
- Network design
- Network implementation
- Network troubleshooting
A Point-to-Point topology involves a direct link between two devices (this is also called a one-to-one connection). This type of connection is often utilized in WAN topologies, especially in some of the legacy Point-to-Point technologies like T1/E1 and T3/E3. Those types of connections go directly from one router to another router.
Point-to-Multipoint is a type of network topology that is very popular today. It is often used in wireless networks and functions by having multiple stations communicate with a central device.
Ring topologies involve connecting all nodes in a ring fashion to ensure redundancy. If one link between two nodes fails, traffic will still have a valid path to reach any devices on the other side of the ring.
A common LAN network topology used in modern networks is the Star topology. This design is used in all sizes of networks and involves all network devices having a direct connection to a central concentrator device that passes the data flows between all other nodes.
Bus networks are not really used in modern networks because they involve great complexity and are subject to errors. Bus topologies are usually built using coaxial cables and they use a single main link that aggregates all nodes on the network.
WAN topologies are characterized by redundancy and are categorized as follows:
- Hub-and-spoke (remote nodes connect to a central device)
- Full-mesh (all nodes are connected to each other)
- Partial-mesh (hub-and-spoke, with additional redundant links between nodes)
Applications generally use two different technologies to operate:
- Client-Server
- Peer-to-Peer
Omnisec network topology guide.
Configure network topologies in our 101 Labs – CompTIA Network+ book.