1.5 Connect Wired Devices

Background and Theory

Introduction

Isn't it time you set up a network? It's time we talked about connecting devices together to create one. The two devices we are going to talk about are hubs and switches, and you'll want to remember what we talked about in the Cabling lesson above. They are called wired devices, well... because they need cables to connect to each other for them to transfer data. If it was wireless, it wouldn't be called wired, now would it?

Before we start, you should look at the following vocabulary words (hover your mouse over each word for the definition). I will be using them throughout the lesson, and you need to know them.

Layers, layers, and more layers...

Hubs operate on the level 1 layer, the physical layer. Switches operate on the level 2 layer, the data link layer. You're probably asking, what does that mean? The layers come from the OSI model (Open Systems Interconnection) and there are seven layers. The OSI model is used to divide the protocols into layers, with the higher levels building from the lower ones. I'm not going to go over the OSI model that much because it is discussed into further detail in The OSI model. But perhaps we can talk about how the first two layers relate to switches and hubs...

What is a hub?

A hub is similar to a switch, except it operates on Layer 1 of the OSI model. It repeats the signal received from one port, and sends it to every other port. An analogy would be to compare the hub to a power strip. The power strip is connected to the power outlet. It provides many sockets for other devices to plug in. Using this method, many devices can be operated by using one power strip. Ethernet hubs are the most common, but there are also USB hubs. Hubs do not read any of the data that goes through them and not aware of the source or destination. They are unsophisticated, unrefined, and unintelligent. Three types of hubs exist:

• Passive - used in a LAN environment. A special form of repeater, where multiple nodes can be wired to a central place and share the same media. Does not repeat the signal.
• Active - Allow multiple nodes to be wired to a central place and share the same media. But they are different from passive hubs in that they do repeat the signal.
• Intelligent - Permit star wiring with UTP cable and provides Simple Network Management Protocol (SNMP) management and configuration from a remote location.

Screenshot of a hub

There are a variety of sizes and features for Hubs. Small hubs have 4 to 24 ports, medium hubs have up to 60 ports, and larger designs have up to 360 ports.

Layer 1: Physical Layer

• Repeater
• Hub
• Concentrator
• MAU

This is the most basic of the layers, and only provides the means of transmitting raw bits. It is used to physically attach users to the network (i.e. with wires). It performs the services of the layer 2, Data link layer. The Physical layer determines how the cable is connected to the network adapter. It also determines the technique for transmitting data over the cable. It sends out bits (binary numbers (0s and 1s)) to the ends it is connected to, and it defines the connector and interface specification. It includes cabling system components, wireless system components, hub/repeater specifications, connector design and pin assignments.

What is a switch?

A network switch is a device that performs transport bridging using wires. It has many ports, unlike a PC based bridge. It is used to connect packet switched network segments into a heterogeneous network, following the protocols of OSI layer 2. That's a mouthful and you'll probably need some more explanations to understand it.

"They interpret and make switching decisions on the LAN hardware adapter address contained in the data link header of MAC frames. They forward frames only to the destination hardware address contained in the frame" - Internet and Network Communications

Network switches are slowly replacing hubs. Switches usually follow a star topology. All the ports in the switch are isolated, until one transmits data. Unlike hubs, they use microsegmentation to create collision domains. They are easy to install, as you'll see in the lab. Today's switches can support 10, 100, and 1,000 Mbps segments.

There are four methods that a switch can forward information:

1. Store and forward
2. Cut through
3. Fragment free
4. Adaptive switching

Of Note

Each method have advantages and disadvantages. It depends on the situation and environment to determine which one is best for you

Store and forward - Information is sent to a station, where it is kept, and sent at a later time to the final destination or another station. The information is verified before forwarding it though. This technique is used in sporadic networks; it is used if there are long transmission rates, high error rates, or if direct contact is not available.

Cut through - The switch starts forwarding that frame (or packet) before the whole frame has been received, normally as soon as the destination address is processed. This technique reduces latency through the switch, but decreases reliability.

Fragment free - This technique ensures that enough bytes are read from a source to prevent collision.

Adaptive Switching - a method automatically switching between the other three modes.

Layer 2: Data link layer

The second layer responds requests coming from the third layer (network layer), and sends requests to the first layer (physical layer). It provides the means of transferring data between network entities and correct errors in the physical layer. It identifies the MAC address and sends the data to the right one. It connects two or more networks using MAC addressing, reduces latency by processing frames using the cut-through method, and makes intelligent forwarding decisions.

"The Data Link layer checks for physical transmission errors and packages bits into frames. The Data Link layer also manages physical addressing schemes such as MAC addresses for Ethernet networks, controlling access of any various network devices to the physical medium. Because the Data Link layer is the single most complex layer in the OSI model, it is often divided into two parts, the Media Access Control sublayer and the Logical Link Control sublayer." -About: Data Link Layer

Comparison between the two

Of Note

Hubs and switches are completely different things! Don't mistake one for the other.

"Technically speaking, hubs operate using a broadcast model and switches operate using a virtual circuit model." What is the difference between a hub and a switch?

Hubs, compared to a switch, are fairly unsophisticated broadcasting devices. Anything that enters the hub is automatically transmitted to every single port. It does not manage the traffic that goes through the ports, and thus results in packet collisions. This disturbs the smooth traffic flow.

A switch, on the other hand, isolates all the ports, so that it is transmitted to only the port where the target is found. Since it intelligently transfers the data, it does not disturb the flow of traffic, and it is generally smoother than that of hubs. The more expensive switches have more functions.

Generally, it depends on the situation to decide which one to pick. If there are fewer ports, then the traffic flow of hubs and switches are minor. However, the more ports there are, the more chances of packet collisions occur with a hub. You would need a switch for a network with many ports.

Topology

• Star Topology
• Extended star Topology

Topolo- what?! Huh? I'm confused... (maybe it's just me) What is a network topology? It is used to determine in what path the computers take to communicate with each other. But it doesn't have to be just computers. It can be any node. A node, by the way, is a device connected in the network. It includes PDAs, cell phones, routers, etc. There are at least seven basic topologies, but the one's we're going to be talking about is the star topology and the extended topology.