NET 222: Introduction to Routers and Routing

Chapter 1: Internetworking

Objectives:
 

This chapter introduces many concepts, the following among them:

  1. Internetworking basics
  2. Network segmentation
  3. How bridges, switches, and routers are used to physically segment a network
  4. How routers are employed to create an internetwork
Concepts:

Before beginning the first chapter, it is recommended that the student skim over the introduction to the text. It covers some history of the Cisco company, some reasons for studying Cisco material, and an overview of various Cisco certifications. This course deals with the first half of the text, which is roughly half of the material you will need to know to pass the Cisco Certified Network Associate exam. The introduction ends with a 40 question test over material from the entire text. Students are not expected to be able to pass this test without studying the material, which is the point of studying the text.

Chapter 1 begins with a discussion of using network hardware to segment a network. In this sense, network segmentation is breaking the network into parts to reduce congestion. Each subnetwork still communicates with the rest of the network as needed, and each subnetwork functions better than it would in a network that is too busy.

A LAN can be congested by many causes. The text offers a list of possible causes:

  • Too many hosts in a broadcast domain
  • Broadcast storms
  • Multicasting
  • Low bandwidth
  • Adding hubs for connectivity to the network
  • A large amount of ARP or IPX traffic

As the author does throughout the text, terms are introduced, then defined. He explains that a broadcast domain is "the set of all devices on a network segment that hear all broadcasts sent on that segment". Why is that a bad thing? If every device on a network were in the same broadcast domain, each device would have to "listen" to every broadcast transmission from every device. This puts an unnecessary load on the processing power of each device, and on the bandwidth of your network. If you break a network into subnetworks, connected by routers, those routers do not pass broadcasts from one network to the next (unless you configure them to do so). This increases the number of broadcast domains, and reduces the amount of broadcast traffic on each of them.

The next term that is defined is collision domain. A collision occurs when two or more signals are sent on a channel meant to carry one at a time. A collision domain is any media segment in which devices contend for access. This is why the author tells us that it is typically an Ethernet problem: Ethernet is one of the most popular network methods, and it works by contention: every device is allowed to attempt to use the network medium at any time it is not already in use. Only one device at a time may transmit within a collision domain.

A hub is a device used to connect several network nodes together. This is good. However, every signal that goes into a port on a hub comes out every other port on the hub. This puts all devices plugged into that hub in one collision domain.

Bridges and switches are superior to hubs because they can determine where a signal is bound, and pass it only where needed. A bridge with two ports creates three collision domains. Assume nodes 1-5 are plugged into a hub connected to bridge port 1. Nodes 6-10 are plugged into a hub connected to bridge port 2. Each of the hubs, and the nodes plugged into them represent one of the collision domains. If traffic originates in one of these domains and is meant for another device in that domain, the bridge does not pass that traffic across. However, if traffic originates in one domain, and is meant for a device in the other domain, that effectively creates a third domain which is the sum of both. In this case, all the devices are still on one network, and are part of one broadcast domain: every device must receive every broadcast transmission.

Switches are used in place of hubs. A switch filters traffic by MAC addresses (the hardware address burned into a network card), and only sends output to the ports that need it, which means that it allows the other ports to be used independently. This has the effect of increasing the bandwidth of the LAN because more users can use it simultaneously. Each device plugged into a switch is on its own collision domain. However, all of those devices are still on the same broadcast domain. A broadcast signal sent into a switch will be sent out all ports.

The way a network works can be understood in terms of a model of a network that was created by the International Organization for Standardization, called the ISO for short. (No it isn't an acronym, it is from the Greek word isos, meaning equal.) Their model is called the Open Systems Interconnection (OSI) Reference Model, hence the ISO-OSI model.

Once you understand this model, you will have a general, powerful reference for examining and comparing networks. In the Novell curriculum, we spent nine chapters on this model. I will present an overview here, in keeping with your text. For those who would like to understand the subject better, without the Cisco spin, look over the notes for course CNE 250.

The seven layers of the model are usually written in a list, numbering the top as layer seven and the bottom as layer one.

  • 7 - Application
  • 6 - Presentation
  • 5 - Session
  • 4 - Transport
  • 3 - Network
  • 2 - Data Link
  • 1 - Physical

Several mnemonic sentences exist to help us remember the proper order. I recommend "Please Do Not Throw Sausage Pizza Away", because this is in the correct numeric order (bottom to top). If you must have one that goes from top to bottom, try "All People Studying This Need Drastic Psychotherapy".  Warning: you MUST remember the correct order, the correct numbers, and the correct details for each layer.

The processes that happen in each layer communicate with the next layer. Which way is next, up or down? It depends whether data is being passed out of the stack (down) or into it (up). Typically, a computer generates a request starting at the top layer, and working down. The request is passed across the network (probably to a server) and the received request is passed up the layers. When a response is generated, the process reverses.

The Novell texts for this subject provide another metaphor for the model: think of each layer of the model as being like a shelf in a bookcase. There are seven shelves, and each shelf has several books on it. The books represent the topics that we discuss in our study of that layer. In fact, in each book are discussions of the various methods that relate to that topic. A method is a way of implementing a task the topic covers. For instance, the Physical layer includes the Connection Type topic (a task), and we will see that there are two classic methods for making connections (ways to do it).

Traditionally, the layers are described from the bottom up. Your text explains them from the top down. This is the order in which they would be encountered by a signal leaving your computer, bound for the network. As a request is passed down the layers, a header is added to the request at each new layer. The header holds information that is added to the request by processes at that layer. So, by the time the request leaves the computer, it has had six more parts added to it. When the request is received by the server, the receiving layers strip off the information added to the request by their counterpart layers. The request increases in size as it leaves the requester, and decreases in size as it is processed by the receiver.

  • Layers 5, 6 and 7: Session, Presentation and Application- Some networks group all these functions into one group of Upper Layer Protocols. The Session layer functions control dialog between nodes, the Presentation layer functions format data and bits, and the Application layer functions share network services.
  • Layer 7, Application - The Application Layer is concerned with all the network services, as well as service advertisement, and service use
  • Layer 6, Presentation - The Presentation Layer is concerned with translation of signals into formats network entities can understand. This includes translating network protocols, characters, and file standards. Your text identifies several file formats that it associates with this layer: PICT, TIFF, JPEG, MPEG, MIDI, RTF, and QuickTime. (Later in the chapter, it refers to them as protocols. They are not.)
  • Layer 5, Session -The Session Layer is concerned with communications between service requesters and providers. Dialogs are set up, maintained, and terminated at this layer. The text tells us that Cisco associates the following with this layer:
    • Network File System (NFS) - used on UNIX systems to provide access to network storage.
    • Structured Query Language (SQL - used to retrieve data from most standard databases
    • Remote Procedure Call (RPC) - a protocol that one program can use to request a service from a program located in another computer in a network
    • X Window - a GUI environment for operating programs on a UNIX server
    • AppleTalk Apple Session Protocol (ASP) - a protocol for creating remote sessions on Apple networks
    The text mentions three types of sessions (simplex, half duplex, and duplex) but does not define them. (How odd. Usually, this author uses a term and then defines it three paragraphs later. He forgot this time?)
    • Simplex - this is communication in one direction, more like a monolog than a dialog. It is like a public speech or a television transmission.
    • Half-Duplex - this is a dialog that can flow both directions, but only one direction at a time. After one side transmits, the channel has to be "reversed" for the other side to transmit. It is like CB or Ham radio, using only one frequency at a time and taking turns.
    • Full-Duplex - this is a dialog in which both sides can transmit and receive at the same time. It is like a telephone conversation, in which both sides have a live earphone and microphone.
  • Layer 4, Transport - The Transport Layer is concerned with segment development (building message units) and moving data to specific processes or services in a reliable way. Associate this layer with the words "reliable" and "dependable". End-to-end control and error checking. In some networks, this may be the last or next-to-last layer.
    • We are told that TCP and UDP are two protocols that operate on this layer, but TCP is the only reliable one. This is to give developers a choice of connection-oriented (reliable) or connectionless (unreliable) service.
    • Flow Control is listed as a Transport layer feature, and described as including segment numbering, acknowledgement of received segments, retransmission of unacknowledged segments, reassembly of segments, and features to manage the rate of data flow. The three name methods for flow control are buffering, windowing, and congestion avoidance.
    • Acknowledgement may be sent for every segment received, but this is slow. It is faster to send several segments, then acknowledge them all. This number of segments that are sent before an acknowledgement is the window size. A window is usually measured in bytes.
    • The text illustrates one method of windowing. The sending station, assuming a window size of 3, sends segments 1-3. The receiving station sends the signal ACK 4, which means it is acknowledging receipt, and wants to continue with segment 4. If a segment is lost, the receiving station sends an ACK with that segment's number, meaning it received everything up to that segment.
    • Flow control can be achieved by having a sufficient number of buffers, temporary storage areas, to hold incoming traffic until it can be processed. If the buffers are full, the receiving device will have to send a signal to the transmitting device, telling it to stop sending until the buffers are emptied.
  • Layer 3, Network - Datagram packets are routed to other networks. Connection services are introduced. The Network Layer is concerned with moving data to specific locations across networks. The Network layer includes addresses that uniquely identify networks.
    • Data packets - Two classic protocols for this layer are IP and IPX. Both are used to send data from one network to another.
    • Route update packets - Routers also exchange information about the networks they are connected to. Some classic protocols used to send these packets are RIP, OSPF, and EIGRP.
    • Routers are the devices associated with this layer. Routers build routing tables, lists that are held in memory of paths to networks. Each routing table entry may include a label for a network, the port to send packets out that are bound for that network, and a metric for that path. A metric may be any of several measures of how far away the destination is on that path, such as hop count (number of routers to cross), tick count (time to reach the destination in ticks. One tick is 1/18th of a second), or bandwidth.
    The author presents a list of features of routers that he suggests should be memorized:
    • Routers, by default, will not forward any broadcast or multicast packets.
    • Routers use the logical address in a Network layer header to determine the next router to forward the packet to.
    • Routers can use access lists, created by an administrator, to control security on the types of packets that are allowed to enter or exit an interface.
    • Routers can provide layer 2 bridging functions if needed and can simultaneously route through the same interface.
    • Layer 3 devices (routers in this case) provide connections between virtual LANs (VLANs).
    • Routers can provide quality of service (QoS) for specific types of network traffic.
  • Layer 2, Data Link - Bits are formed into frames, headers give address information. This layer tells us how the network is physically accessed, how some errors are handled, how data flow is handled, and how entities on the network are addressed.
    Addresses are written in the header of this layer every time a frame is passed from one router to another. How come? What happens is like this
    1. The sender's network address and node ID are saved in the Network Layer header, along with the network address and node ID for the recipient. This data stays in the Network Layer header until the intended recipient breaks down the header.
    2. Since the intended recipient is not on the sender's network, the sending station sets the Data Link Layer address of the recipient to the address of the router on his network, and sends a frame
    3. The router on the sender's network gets the frame, erases the sender and recipient data in the Data Link Layer, and decides on a route to the recipient's network. The next router in a logical chain is selected. The next router's address is written in the Data Link Layer header as the "recipient", and the current router's address is written to the Data Link Layer header as the "sender". This works because routers are special. The are members of their own networks, but they are also members of other networks. (Think of a series of links in a chain. Each link is like a network, and routers are like the points where the links touch.)
    4. The process in the step above is repeated until a router on the intended recipient's network gets the frame. Then, information matching the original information in the Network Layer header is written to the Data Link Layer header, and the frame is delivered.
    Sub-layers:
    • Media Access Control (MAC) - Rules to access the media, logical topologies, hardware addressing. There are three classic media access methods: contention, token passing, and polling. The author discusses MAC addresses (hardware addresses) and explains that bridges and switches can filter traffic based on these addresses. MAC addresses are six bytes long, usually written out as two hexadecimal digits per byte. The first three bytes of an address are assigned by the IEEE to each NIC manufacturer. The manufacturers, in turn, assign the remaining three bytes of each address as NICs are created. The first three bytes are called the organizationally unique identifier (OUI). The text calls the last three bytes the globally administered address.
    • Logical Link Control (LLC) - Frame synchronization, connection services and error control. The text explains that Data Link layer frames can contain information about which protocol is being used in the Network layer. This enables the receiver to hand the frame to the proper protocol.
  • Layer 1, Physical - Protocols define structure, physical specifications for media, rules for transmitting bits. Most network models do not specify what happens on this layer, making it possible to use many kinds of wiring for different kinds of networks. This layer tells us how the network is physically set up. This layer does not use addressing. Hubs belong on this layer, because they do not care about addresses either.

The chapter changes focus to describe Ethernet networks. Ethernet is a networking method invented by Bob Metcalf at a place called Xerox PARC (Palo Alto Research Center). As stated above, it is a popular method that uses contention.

Contention systems work by letting each device try to send a message on the net as needed, contending (competing) with all the other devices for the bandwidth. Ethernet uses CSMA/CD (Carrier Sensing Multiple Access, with Collision Detection). Carrier Sensing means listening to the medium, and only transmitting when it is not in use. Multiple Access means all devices can attempt to access the medium as needed. A collision occurs when two signals collide on the medium, causing signal loss. In a CSMA/CD system, collisions are detected and the devices that cause them each wait a random number of nanoseconds before sending again. The text refers to this waiting period as "backoff". (Sounds like a test question, doesn't it?) This usually results in one device going ahead of the other, avoiding a new collision.

The text explains that Cisco thinks of Ethernet as running half-duplex or full-duplex (duplex). The author describes half-duplex as using one circuit, and transmissions flow both directions. This does not mention the part about the transmissions only flowing one direction at a time. Cisco describes a hub passing information to a switch in half-duplex mode, because the stations would have to watch for collisions. We are further told that such a system would only run at 30 to 40 per cent of the theoretical bandwidth.

If every device were connected through switches instead, each device would be on its own collision domain, which Cisco says allows the network to run in full-duplex mode, using two circuits for each connection. This system, according to the text, would use those two circuits at 100% efficiency, taking us to 200% of the theoretical bandwidth. The author cautions that this is theory, and it may not occur in practice. Three kinds of connections are listed as candidates for full-duplex connections: host to switch, switch to switch, and host to host using a crossover cable (which, of course, is defined later in the chapter).

It is explained that the Data Link layer creates frames to hold data passed to it from the Network layer. There are several types of frames, four of which are listed and described:

  • Ethernet_II
  • IEEE 802.3
  • IEEE 802.2
  • SNAP (Sub-Network Access Protocol)

The author digresses for a bit into what he calls Ethernet addressing. Think of it this way: MAC addresses are unique identifiers for individual machines. If you did not need to send traffic from one network to another, MAC addresses could be the only addresses you might need. This is why we are told that Ethernet addressing uses MAC addresses. We are also told that MAC addresses are used as the host (individual device) identifier on networks that use the IPX/SPX protocol suite. (These would be older Novell networks, from before the time that Novell built in support for the TCP/IP protocol suite.)

Back to frames... all frames have some common characteristics. Frames have fields that contain different types of information. Ethernet_II and 802.3 frames both contain 8 bytes of preamble (like a warning that data is coming), 6 bytes of Destination Address (recipient's MAC address), and 6 bytes of Source Address (sender's MAC address). I told you above how theses two addresses are constantly rewritten when routers pass data between them. These fields are followed by Type OR Length fields (see below), a data field of up to 1500 bytes, and a Frame Check Sequence (FCS) field holding error checking data. (Other fields exist, but these are the important ones.)

Ethernet_II frames have a Type field next, that tells the type of protocol to hand the data to on the Network layer. The text tells us the Type 08-00 stands for IP.

802.3 frames do not have a Type field. They have a Length field. The designers of this frame type created it for Novell networks that only used IPX/SPX, so they did not see a need for a Type field. Darn. To overcome this lack of information the other two frame types are like 802.3, with this information included.

Ethernet_II 802.3
Preamble - 8 bytes Preamble - 8 bytes
Destination Address - 6 bytes Destination Address - 6 bytes
Source Address - 6 bytes Source Address - 6 bytes
Type - 2 bytes Length - 2 bytes
DATA DATA
FCS FCS

 

802.2 frames include Destination SAP and Source SAP fields. These fields hold data to identify the upper layer protocol to hand off to. The example in the text shows the value 0xe0, which stands for NetWare, which means IPX/SPX.

SNAP frames have Destination SAP and Source SAP fields, as well, but they always hold the value 0xAA, which means they are SNAP frames. The protocol information is stored in a Protocol field. The author warns us that Cisco uses a SNAP frame with their proprietary protocol Cisco Discovery Protocol (CDP).

The author uses the acronym IEEE several times, but I did not see his definition of it. The Institute of Electrical and Electronic Engineers (IEEE) is a standards organization that has specified many of the rules used in building networks. Their standards are often referred to by number. In this chapter, we discuss some of the IEEE 802.x standards. LAN protocols that support the IEEE 802 standards can be referred to as 802.x protocols.

  • 802.2 - specifies the use of headers and frames, supporting the LLC sublayer of the Data-Link layer.
  • 802.3 - specifies the CSMA/CD access method, so this is often thought to be the Ethernet standard.
    802.3 was based on Ethernet, and made a more general standard. Ethernet can be thought of as one implementation of the 802.3 standard. This standard fits on the Physical layer and the MAC sublayer of the Data-Link layer. You should be aware of the parsing (breaking into understandable parts) of the names of LAN types using this standard. For instance, 10BASE5 means 10 Mbps, baseband transmission and a limit of 500 meters per segment (thick Ethernet, using coaxial cable). 10BASET means 10 Mbps, baseband transmission and using UTP cable. (Parse: to resolve into component parts. This is a word used in English grammar classes.) Base may be shown in lower or upper case in the name of a LAN type.
  • 802.3u - Fast Ethernet
  • 802.3z and 802.3ab - Gigabit Ethernet. Your text lists other variations.
  • 802.5 - specifies a token passing system based on IBM's token ring standard. IBM's standard specifies a physical ring, but 802.5 does not, so we often see physical stars (constructed like a star) that are logical rings (work like a ring) by this standard.
  • 802.11 - specifies how wireless LANs work, like spread spectrum, infrared, and short range Gigahertz radio
  • 802.12 - 100VG-AnyLAN, is a standard that may authorize a hub to decide which of two contenders for the bandwidth should have it, based on priority. Not used much.

Traditionally, networks have been built with cable of various types. Some classic types of cable media:

  • twisted pair - come in two types:
    • unshielded - UTP does not have an EMI resistant sheath
    • shielded - STP has an EMI resistant sheath
  • coaxial - Coax similar to that used for cable TV, or a much thicker variety sometimes called "Etherhose"
  • fiber optic - glass or plastic channels that conduct light, often red laser light. This is the only one that is immune to electromagnetic interference.

For images of cable media, I recommend following this link to online catalog of the Black Box company.

Your text discusses one cable type in detail: Unshielded Twisted Pair (UTP). The graphic on this page shows several twisted pairs of wires. Each wire is covered with an insulator, and the two wires in each pair are meant to be used as a circuit. These wires suffer from crosstalk, leakage of signal from one circuit to another. The twists help cancel out such leaks. The graphic shows a UTP cable with eight wires in it, making four pairs.

This type of cable comes in several varieties: two pair, three pair and four pair are common. Also, each variety may be available in grades, such as CAT 1 (Category 1) through CAT 5 (Category 5). There are several categories, and a major difference between them is the number of twists per foot in each pair. CAT 1 will have less than 5 twists per foot, CAT 5 will have 25 or more twists per foot (so it is better, and costs more). Note that the better the class of cable, the more bits per second can be passed across it.

Connecting a system with twisted pair wiring is easy. You may use either hubs or switches, but switches are preferred, for the reasons noted above.

  • Cost - inexpensive
  • Installation - cheap and easy
  • Capacity - 1 to 100 Megabits per second (Mbps), but 10 Mbps is common (100 Mbps, if Fast Ethernet).
  • Attenuation - nothing is perfect, so this is high (poor)
  • Immunity from Electromagnetic interference (EMI) - also poor. Recommendation: run UTP lines perpendicular (not parallel) to fluorescent lights.

UTP cables are usually connected to devices with RJ-45 connectors. Your text does not show an RJ-45 connector (or any other) very well. In the enlarged picture on the right, note the eight gold-colored connections for the eight wires usually found in UTP cables. The wires are used in pairs to form circuits. See the Networking Technologies notes for more information on this sort of connector.
The text lists three ways to wire a UTP cable to an RJ-45 connector:

  • Straight-through (standard) cable - to connect a device to a standard hub or switch (concentrator)
  • Crossover cable - a crossover cable is used on UTP networks to bypass a concentrator and connect two NICs directly to each other. You can't use a standard UTP cable for this: one end of the cable must have the live circuits reversed.
  • Rolled cable, also called rollover cable - to connect a workstation to Cisco router

The tables below describe a straight-through cable, a crossover cable, and a rollover cable.

Standard RJ-45 Pin Assignments
Pin Color Color Signal Circuit
1 Orange/White
Orange White
TX data + Orange Circuit
2 Orange Orange TX data - Orange Circuit
3 Green/White
Green White
RX data + Green Circuit
4 Blue Blue unused Blue Circuit
5 Blue/White
Blue White
unused Blue Circuit
6 Green Green RX data - Green Circuit
7 Brown/White
Brown White
unused Brown Circuit
8 Brown Brown unused Brown Circuit

The insulation shown in the graphics above should NOT be stripped back on these wires.
Straight-through (standard) cable
If you are making a straight-through cable (to run from a workstation to a hub or switch) connect both ends as listed above and shown on the right. Insert the wires into the RJ-45 connector, then crimp with the crimping tool. (There will be no spaces between the wires when they are inserted into the RJ-45 connector. Space is used here to make the color pattern more readable.)
End 1
End 2
Crossover cable
If making a crossover cable (to run directly from one NIC to another) swap the orange and green circuits on one end only: put orange/white on 3, orange on 6, green/white on 1, and green on 2. Insert the wires as shown on the right, then crimp. (This second configuration is actually EIA/TIA 568A.)
End 1
End 2
Rollover cable
Now, for something completely different, if you are making a rollover cable (to run from a workstation to an older Cisco router), prepare the cable like a standard cable, both ends in the same configuration.
Before
crimping the second end, roll the cable (or the RJ-45 connector) over, 180 degrees. The will make pin 1 on one end of the cable connect to pin 8 on the other end, pin 2 to pin 7, pin 3 to pin 6, and pin 4 to pin 5. If you don't want to think about rolling anything over, insert the wires as shown on the right, then crimp. This cable is used with an adapter to connect to a Cisco router's console port.
End 1
End 2

Your text suggests that making a rolled cable is easy, since you just cut the end off a good straight-through cable, and crimp a new end on it that has been rolled over. In the real world, the crimping part is not easy until you are practiced at it.

The text describes configuring Hyperterminal to use as your connection to the Cisco console. It does not mention how to start Hyperterminal. On my XP workstation, I clicked Start | Programs | Accessories | Communications | Hyperterminal. The text also specifies that the default port settings in Hyperterminal will not work for a connection to the Cisco console. The recommended port settings are:

  • Bits per second - 9600
  • Data bits - 8
  • Parity - None
  • Stop bits - 1
  • Flow control - None

The text continues with a few statements about wireless networking. This topic includes several technologies. At this time, the most popular seem to be Broadband Personal Communication Service (PCS) and Wi-Fi (IEEE 802.11b and g). You should be aware of the usual bandwidth limits for Wi-Fi: 802.11b is limited to 11 Mbps (Megabits per second), 802.11g is limited to 54 Mbps. These are theoretical limits: actual bandwidth may be less.

The text returns to the OSI model by telling us that as data packets are passed down from one layer to another, they are wrapped in an envelope by each new layer. That envelope includes information from each protocol that acts on the data at that layer. The process of wrapping new information around the message is called encapsulation. Each message passed down the layers has a series of wrappers placed on it. Notably, the Transport layer adds a Segment wrapper, the Network layer adds a Packet wrapper, the Data Link layer adds a Frame wrapper, and the Physical layer adds a Bit wrapper.

Encapsulation is also used to send data across networks that do not support specific protocols. I once worked on a system that supported IPX/SPX, but not TCP/IP. The problem was that we had to pass data to distant IPX networks across IP networks that would not recognize our IPX traffic. The solution was to encapsulate our IPX traffic in IP wrappers, pass it across the IP networks, and remove the wrappers at the receiving systems.

The text discusses the specific name given to protocol data units (globs of data) at each layer. All layers involve the term packets, and most have other terms specific to them as well:

  • 7 - Application: Messages and packets
  • 6 - Presentation: Packets
  • 5 - Session: Packets
  • 4 - Transport: Segments and packets
  • 3 - Network: Datagrams and packets
  • 2 - Data Link: Frames and packets
  • 1 - Physical: Bits and packets

The last major topic in the chapter is the Cisco hierarchical model for networks. It describes three layers and the relationship between devices found in those layers. This is a logical model, so devices may perform functions on more than one layer. The hierarchy is a structure of functions, not of physical layout of the network.

  • Core layer - backbone functions, high bandwidth data paths and connectivity devices
  • Distribution layer - routing functions, routers and switches that connect LANs and VLANs together
  • Access layer - switching functions, collision domains and switches that create LANs