Overview
26.1 Introduction
to Ethernet/802.3 LANs
26.1.1
Ethernet/802.3 LAN development
26.1.2 Factors
that impact network performance
26.1.3 Elements
of Ethernet/802.3 networks
26.1.4 Half-duplex
networks
26.1.5 Network
congestion
26.1.6 Network
latency
26.1.7 Ethernet
10BASE-T transmission time
26.1.8 The
benefits of using repeaters
26.1.9
Full-duplex transmitting
26.2 Introduction
to LAN Switching
26.2.1 LAN
segmentation
26.2.2 LAN
segmentation with bridges
26.2.3 LAN
segmentation with routers
26.2.4 LAN
segmentation with switches
26.2.5 Basic
operations of a switch
26.2.6 Ethernet
switch latency
26.2.7 Layer 2
and Layer 3 switching
26.2.8 Symmetric
and asymmetric switching
26.2.9 Memory
buffering
26.2.10 Two
switching methods
26.3 Switch
Operation
26.3.1 Functions
of Ethernet switches
26.3.2 Frame
transmission modes
26.3.3 How
switches and bridges learn addresses
26.3.4 How
switches and bridges filter frames
26.3.5 Why
segment LANs?
26.3.6
Microsegmentation implementation
26.3.7 Switches
and collision domains
26.3.8 Switches
and broadcast domains
26.3.9
Communication between switches and workstations
Summary
Overview
LAN design has
evolved. Network designers until very recently used hubs and bridges to build
networks. Now switches and routers are the key components in LAN design, and
the capabilities and performance of these devices continue to improve.
This module
describes the roots of modern Ethernet LANs with an emphasis on the evolution
of Ethernet/802.3, the most commonly deployed LAN architecture. A look at the
historical context of LAN development and various network devices that can be
utilized at different layers of the OSI model will help students better
understand the reasons why network devices have evolved as they have.
Until recently,
repeaters were used in most Ethernet networks. Network performance suffered as
too many devices shared the same segment. Network engineers then added bridges
to create multiple collision domains. As networks grew in size and complexity,
the bridge evolved into the modern switch which allows microsegmentation of the
network. Modern networks are now built with switches and routers, often with both
functionalities in one device.
Many modern
switches are capable of performing varied and complex tasks in the network.
This module will provide an introduction to network segmentation and will
describe the basics of switch operation.
Switches and
bridges perform much of the heavy work in LANs where they make nearly
instantaneous decisions when frames are received. This module describes in
detail how switches learn the physical addresses of nodes, and how switches
transmit and filter frames. This module also describes the principles of LAN
segmentation and collision domains.
Switches are
Layer 2 devices that are used to increase available bandwidth and reduce
network congestion. A switch can segment a LAN into microsegments, which are
segments with only a single host. Microsegmentation creates multiple
collision-free domains from one large domain. As a Layer 2 device, the LAN
switch increases the number of collision domains, but all hosts connected to
the switch are still part of the same broadcast domain.
This module
covers some of the objectives for the CCNA 640-801 and ICND 640-811 exams.
Students who
complete this module should be able to perform the following tasks:
- Describe the history and
function of shared, or half-duplex Ethernet
- Define collision as it relates
to Ethernet networks
- Define CSMA/CD
- Describe some of the key
elements that affect network performance
- Describe the function of
repeaters
- Define network latency
- Define transmission time
- Define network segmentation
with routers, switches, and bridges
- Define Ethernet switch latency
- Explain the differences between
Layer 2 and Layer 3 switching
- Define symmetric and asymmetric
switching
- Define memory buffering
- Compare and contrast
store-and-forward and cut-through switching
- Understand the differences
between hubs, bridges, and switches
- Describe the main functions of
switches
- List the major switch frame
transmission modes
- Describe the process by which
switches learn addresses
- Identify and define forwarding
modes
- Define LAN segmentation
- Define microsegmentation with
the use of switches
- Describe the frame-filtering
process
- Compare and contrast collision
and broadcast domains
- Identify the cables needed to
connect switches to workstations
- Identify the cables needed to
connect switches to other switches
26.1
Introduction to Ethernet/802.3 LANs
26.1.1
Ethernet/802.3 LAN development
This page will
review the devices that are found on a network.
The earliest LAN
technologies used either thick Ethernet or thin Ethernet infrastructures. It is
important to understand the limitations of these infrastructures, as shown in
Figure , in order to understand the advancements in LAN switching.
The addition of
hubs or concentrators into the network offered an improvement on thick and thin
Ethernet technology. A hub is a Layer 1 device and is sometimes referred to as
an Ethernet concentrator or a multi-port repeater. Hubs allow better access to
the network for more users. Hubs regenerate data signals which allows networks to
be extended to greater distances. Hubs do not make any decisions when data
signals are received. Hubs simply regenerate and amplify the data signals to
all connected devices, except for the device that originally sent the signal.
Ethernet is
fundamentally a shared technology where all users on a given LAN segment
compete for the same available bandwidth. This situation is analogous to a
number of cars that try to access a one-lane road at the same time. Since the
road has only one lane, only one car can access it at a time. As hubs were
added to the network, more users competed for the same bandwidth.
Collisions are a
by-product of Ethernet networks. If two or more devices try to transmit at the
same time, a collision occurs. This situation is analogous to two cars that try
to merge into a single lane and cause a collision. Traffic is backed up until
the collision can be cleared. Excessive collisions in a network result in slow
network response times. This indicates that the network is too congested or has
too many users who need to access the network at the same time.
Layer 2 devices
are more intelligent than Layer 1 devices. Layer 2 devices make forwarding
decisions based on Media Access Control (MAC) addresses contained within the
headers of transmitted data frames.
A bridge is a
Layer 2 device used to divide, or segment, a network. Bridges collect and
selectively pass data frames between two network segments. In order to do this,
bridges learn the MAC address of devices on each connected segment. With this
information, the bridge builds a bridging table and forwards or blocks traffic
based on that table. This results in smaller collision domains and greater
network efficiency. Bridges do not
restrict broadcast traffic. However, they do provide greater traffic control
within a network.
A switch is also
a Layer 2 device and may be referred to as a multi-port bridge. Switches make
forwarding decisions based on MAC addresses contained within transmitted data
frames. Switches learn the MAC addresses of devices connected to each port and
this information is entered into a switching table.
Switches create a
virtual circuit between two connected devices that want to communicate. When
the virtual circuit is created, a dedicated communication path is established
between the two devices. The implementation of a switch on the network provides
microsegmentation. This creates a collision free environment between the source
and destination, which allows maximum utilization of the available bandwidth.
Switches are able to facilitate multiple, simultaneous virtual circuit
connections. This is analogous to a highway that is divided into multiple lanes
and each car has its own dedicated lane.
The disadvantage
of Layer 2 devices is that they forward broadcast frames to all connected
devices on the network. Excessive broadcasts in a network result in slow
network response times.
A router is a
Layer 3 device. Routers make decisions based on groups of network addresses, or
classes, as opposed to individual MAC addresses. Routers use routing tables to
record the Layer 3 addresses of the networks that are directly connected to the
local interfaces and network paths learned from neighbor routers.
The following are
functions of a router:
- Examine inbound packets of
Layer 3 data
- Choose the best path for the
data through the network
- Route the data to the proper
outbound port
Routers do not
forward broadcasts unless they are programmed to do so. Therefore, routers
reduce the size of both the collision domains and the broadcast domains in a
network. Routers are the most important devices to regulate traffic on large
networks. Routers enable communication between two computers regardless of
location or operating system.
LANs typically
employ a combination of Layer 1, Layer 2, and Layer 3 devices. Implementation
of these devices depends on factors that are specific to the particular needs
of the organization.
26.1
Introduction to Ethernet/802.3 LANs
26.1.2
Factors that impact network performance
This page will describe
some factors that cause LANs to become congested and overburdened. In addition
to a large number of network users, several other factors have combined to test
the limits of traditional LANs:
- The multitasking environment
present in current desktop operating systems such as Windows, Unix/Linux,
and Mac OS X allows for simultaneous network transactions. This increased
capability has lead to an increased demand for network resources.
- The use of network intensive
applications such as the World Wide Web has increased. Client/server
applications allow administrators to centralize information and make it
easier to maintain and protect information.
- Client/server applications do
not require workstations to maintain information or provide hard disk
space to store it. Given the cost benefit of client/server applications,
such applications are likely to become even more widely used in the
future.
26.1 Introduction to Ethernet/802.3 LANs
26.1.3
Elements of Ethernet/802.3 networks
This page will
describe some factors that can have a negative impact on the performance of an
Ethernet network.
Ethernet is a
broadcast transmission technology. Therefore network devices such as computers,
printers, and file servers communicate with one another over a shared network
medium. The performance of a shared medium Ethernet/802.3 LAN can be negatively
affected by several factors:
- The data frame delivery of
Ethernet/802.3 LANs is of a broadcast nature.
- The carrier sense multiple
access/collision detect (CSMA/CD) method allows only one station to
transmit at a time.
- Multimedia applications with
higher bandwidth demand such as video and the Internet, coupled with the
broadcast nature of Ethernet, can create network congestion.
- Normal latency occurs as frames
travel across the network medium and through network devices.
Ethernet uses
CSMA/CD and can support fast transmission rates. Fast Ethernet, or 100BASE-T,
provides transmission speeds up to 100 Mbps. Gigabit Ethernet provides
transmission speeds up to 1000 Mbps and 10-Gigabit Ethernet provides
transmission speeds up to 10,000 Mbps. The goal of Ethernet is to provide a
best-effort delivery service and allow all devices on the shared medium to
transmit on an equal basis. Collisions are a natural occurrence on Ethernet
networks and can become a major problem.
26.1
Introduction to Ethernet/802.3 LANs
26.1.4
Half-duplex networks
This page will
explain how collisions occur on a half-duplex network.
Originally
Ethernet was a half-duplex technology. Half-duplex allows hosts to either
transmit or receive at one time, but not both. Each host checks the network to
see whether data is being transmitted before it transmits additional data. If
the network is already in use, the transmission is delayed. Despite
transmission deferral, two or more hosts could transmit at the same time. This
results in a collision. When a collision occurs, the host that detects the
collision first, sends out a jam signal to the other hosts. When a jam signal
is received, each host stops data transmission, then waits for a random period
of time to retransmit the data. The back-off algorithm generates this random
delay. As more hosts are added to the network, collisions are more likely to
occur.
Ethernet LANs
become saturated because users run network intensive software, such as
client/server applications, which cause hosts to transmit more often and for
longer periods of time. The network interface card (NIC), used by LAN devices,
provides several circuits so that communication among devices can occur.
26.1
Introduction to Ethernet/802.3 LANs
26.1.5
Network congestion
This page will
discuss some factors that create a need for more bandwidth on a network.
Advances in
technology produce faster and more intelligent desktop computers and
workstations. The combination of more powerful workstations and network
intensive applications has created a need for greater network capacity, or
bandwidth. All these factors place a
strain on networks with 10 Mbps of available bandwidth and that is why many
networks now provide 100 Mbps bandwidth on their LANs.
The following are
types of media that have increased in transmission over networks:
- Large graphics files
- Full-motion video
- Multimedia applications
There is also an
increase in the number of users on a network. As more people utilize networks
to share larger files, access file servers, and connect to the Internet,
network congestion occurs. This results in slower response times, longer file
transfers, and less productive network users. To relieve network congestion,
either more bandwidth is needed or the available bandwidth must be used more
efficiently.
26.1
Introduction to Ethernet/802.3 LANs
26.1.6
Network latency
This page will
help students understand the factors that increase network latency.
Latency, or
delay, is the time a frame or a packet takes to travel from the source station
to the final destination. It is important to quantify the total latency of the
path between the source and the destination for LANs and WANs. In the specific
case of an Ethernet LAN, it is important to understand latency and its effect
on network timing as it is used to determine if CSMA/CD will work properly.
Latency has at
least three sources:
- First, there is the time it
takes the source NIC to place voltage pulses on the wire and the time it
takes the destination NIC to interpret these pulses. This is sometimes
called NIC delay, typically around 1 microsecond for a 10BASE-T NIC.
- Second, there is the actual
propagation delay as the signal takes time to travel through the cable.
Typically, this is about 0.556 microseconds per 100 m for Cat 5 UTP.
Longer cable and slower nominal velocity of propagation (NVP) results in
more propagation delay.
- Third, latency is added based
on network devices that are in the path between two computers. These are
either Layer 1, Layer 2, or Layer 3 devices.
Latency does not
depend solely on distance and number of devices. For example, if three properly
configured switches separate two workstations, the workstations may experience
less latency than if two properly configured routers separated them. This is
because routers conduct more complex and time-intensive functions. A router
must analyze Layer 3 data.
26.1
Introduction to Ethernet/802.3 LANs
26.1.7
Ethernet 10BASE-T transmission time
This page will
explain how transmission time is determined for 10BASE-T.
All networks have
what is called bit time or slot time. Many LAN technologies, such as Ethernet,
define bit time as the basic unit of time in which one bit can be sent. In
order for the electronic or optical devices to recognize a binary one or zero,
there must be some minimum duration during which the bit is on or off.
Transmission time
equals the number of bits to be sent times the bit time for a given technology.
Another way to think about transmission time is the interval between the start
and end of a frame transmission, or between the start of a frame transmission
and a collision. Small frames take a shorter amount of time. Large frames take
a longer amount of time.
Each 10-Mbps
Ethernet bit has a 100 ns transmission window. This is the bit time. A byte
equals eight bits. Therefore, 1 byte takes a minimum of 800 ns to transmit. A
64-byte frame, which is the smallest 10BASE-T frame that allows CSMA/CD to
function properly, has a transmission time of 51,200 ns or 51.2 microseconds.
Transmission of an entire 1000-byte frame from the source requires 800
microseconds. The time at which the frame actually arrives at the destination
station depends on the additional latency introduced by the network. This
latency can be due to a variety of delays including all of the following:
- NIC delays
- Propagation delays
- Layer 1, Layer 2, or Layer 3
device delays
26.1
Introduction to Ethernet/802.3 LANs
26.1.8
The benefits of using repeaters
This page will
explain how a repeater can be used to extend the distance of a LAN.
The distance that
a LAN can cover is limited due to attenuation. Attenuation means that the
signal weakens as it travels through the network. The resistance in the cable
or medium through which the signal travels causes the loss of signal strength.
An Ethernet repeater is a physical layer device on the network that boosts or
regenerates the signal on an Ethernet LAN. When a repeater is used to extend
the distance of a LAN, a single network can cover a greater distance and more
users can share that same network. However, the use of repeaters and hubs adds
to problems associated with broadcasts and collisions. It also has a negative
effect on the overall performance of the shared media LAN.
26.1
Introduction to Ethernet/802.3 LANs
26.1.9
Full-duplex transmitting
This page will
explain how full-duplex Ethernet allows the transmission of a packet and the
reception of a different packet at the same time. This simultaneous
transmission and reception requires the use of two pairs of wires in the cable
and a switched connection between each node. This connection is considered point-to-point
and is collision free. Because both nodes can transmit and receive at the same
time, there are no negotiations for bandwidth. Full-duplex Ethernet can use a
cable infrastructure already in place, as long as the medium meets the minimum
Ethernet standards.
To transmit and
receive simultaneously, a dedicated switch port is required for each node.
Full-duplex connections can use 10BASE-T, 100BASE-TX, or 100BASE-FX media to
create point-to-point connections. The NICs on all connected devices must have
full-duplex capabilities.
The full-duplex
Ethernet switch takes advantage of the two pairs of wires in the cable and
creates a direct connection between the transmit (TX) at one end of the circuit
and the receive (RX) at the other end. With the two stations connected in this
manner a collision free environment is created as the transmission and receipt
of data occurs on separate non-competitive circuits.
Ethernet can
usually only use 50 to 60 percent of the available 10 Mbps of bandwidth because
of collisions and latency. Full-duplex Ethernet offers 100 percent of the
bandwidth in both directions. This produces a potential 20 Mbps throughput,
which results from 10 Mbps TX and 10 Mbps RX.
26.2
Introduction to LAN Switching
26.2.1
LAN segmentation
This page will
explain LAN segmentation.
A network can be
divided into smaller units called segments. Figure shows an example of a segmented Ethernet
network. The entire network has fifteen computers. Of the fifteen computers,
six are servers and nine are workstations. Each segment uses the CSMA/CD access
method and maintains traffic between users on the segment. Each segment is its
own collision domain.
Segmentation
allows network congestion to be significantly reduced within each segment. When
data is transmitted within a segment, the devices within that segment share the
total available bandwidth. Data that is passed between segments is transmitted
over the backbone of the network through a bridge, router, or switch.
26.2
Introduction to LAN Switching
26.2.2
LAN segmentation with bridges
This page will
describe the main functions of a bridge in a LAN.
Bridges are Layer
2 devices that forward data frames based on the MAC address. Bridges read the
source MAC address of the data packets to discover the devices that are on each
segment. The MAC addresses are then used to build a bridging table. This allows
bridges to block packets that do not need to be forwarded from the local
segment.
Although bridges
are transparent to other network devices, the latency on a network increases by
ten to thirty percent when a bridge is used. The increased latency is because
of the decisions that bridges make before the packets are forwarded. A bridge
is considered a store-and-forward device. Bridges examine the destination
address field and calculate the cyclic redundancy check (CRC) in the Frame
Check Sequence field before the frame is forwarded. If the destination port is
busy, bridges temporarily store the frame until that port is available.
26.2
Introduction to LAN Switching
26.2.3
LAN segmentation with routers
This page will
explain how routers are used to segment a LAN.
Routers provide
network segmentation which adds a latency factor of twenty to thirty percent over
a switched network. The increased latency is because routers operate at the
network layer and use the IP address to determine the best path to the
destination node. Figure shows a Cisco
router.
Bridges and
switches provide segmentation within a single network or subnetwork. Routers
provide connectivity between networks and subnetworks.
Routers do not
forward broadcasts while switches and bridges must forward broadcast frames.
The Interactive
Media Activities will help students become more familiar with the Cisco 2621
and 3640 routers.
26.2
Introduction to LAN Switching
26.2.4
LAN segmentation with switches
This page will
explain how switches are used to segment a LAN.
Switches decrease
bandwidth shortages and network bottlenecks, such as those between several
workstations and a remote file server. Figure
shows a Cisco switch. Switches segment LANs into microsegments which
decreases the size of collision domains. However, all hosts connected to a
switch are still in the same broadcast domain.
In a completely
switched Ethernet LAN, the source and destination nodes function as if they are
the only nodes on the network. When these two nodes establish a link, or
virtual circuit, they have access to the maximum available bandwidth. These links
provide significantly more throughput than Ethernet LANs connected by bridges
or hubs. This virtual network circuit is
established within the switch and exists only when the nodes need to
communicate.
26.2
Introduction to LAN Switching
26.2.5
Basic operations of a switch
This page will
discuss the basic functions of a switch in a LAN.
Switching is a
technology that decreases congestion in Ethernet, Token Ring, and Fiber
Distributed Data Interface (FDDI) LANs. Switches use microsegmentation to
reduce collision domains and network traffic. This reduction results in more
efficient use of bandwidth and increased throughput. LAN switches often replace
shared hubs and are designed to work with cable infrastructures already in
place.
The following are
the two basic operations that switches perform:
- Switch data frames - The
process of receiving a frame on a switch interface, selecting the correct
forwarding switch port(s), and forwarding the frame.
- Maintain switch operations -
Switches build and maintain forwarding tables. Switches also construct and
maintain a loop-free topology across the LAN.
Figures
through show the basic operations
of a switch.
26.2
Introduction to LAN Switching
26.2.6
Ethernet switch latency
This page will
explain how Ethernet switches contribute to latency.
Switch latency is
the period of time when a frame enters a switch to the time it takes the frame
to exit the switch. Latency is directly related to the configured switching
process and volume of traffic.
Latency is
measured in fractions of a second. Network devices operate at incredibly high
speeds so every additional nanosecond of latency adversely affects network
performance.
26.2
Introduction to LAN Switching
26.2.7
Layer 2 and Layer 3 switching
This page will
show students how switching occurs at the data link and the network layer.
There are two
methods of switching data frames, Layer 2 switching and Layer 3 switching.
Routers and Layer 3 switches use Layer 3 switching to switch packets. Layer 2
switches and bridges use Layer 2 switching to forward frames.
The difference
between Layer 2 and Layer 3 switching is the type of information inside the
frame that is used to determine the correct output interface. Layer 2 switching
is based on MAC address information. Layer 3 switching is based on network
layer addresses, or IP addresses. The features and functionality of Layer 3
switches and routers have numerous similarities. The only major difference
between the packet switching operation of a router and a Layer 3 switch is the
physical implementation. In general-purpose routers, packet switching takes
place in software, using microprocessor-based engines, whereas a Layer 3 switch
performs packet forwarding using application specific integrated circuit (ASIC)
hardware.
Layer 2 switching
looks at a destination MAC address in the frame header and forwards the frame
to the appropriate interface or port based on the MAC address in the switching
table. The switching table is contained
in Content Addressable Memory (CAM). If the Layer 2 switch does not know where
to send the frame, it broadcasts the frame out all ports to the network. When a
reply is returned, the switch records the new address in the CAM.
Layer 3 switching
is a function of the network layer. The Layer 3 header information is examined
and the packet is forwarded based on the IP address.
Traffic flow in a
switched or flat network is inherently different from the traffic flow in a
routed or hierarchical network. Hierarchical networks offer more flexible
traffic flow than flat networks.
26.2
Introduction to LAN Switching
26.2.8 Symmetric
and asymmetric switching
This page will
explain the difference between symmetric and asymmetric switching.
LAN switching may
be classified as symmetric or asymmetric based on the way in which bandwidth is
allocated to the switch ports. A symmetric switch provides switched connections
between ports with the same bandwidth.
An asymmetric LAN switch provides switched connections between ports of
unlike bandwidth, such as a combination of 10-Mbps and 100-Mbps ports.
Asymmetric
switching enables more bandwidth to be dedicated to the server switch port in
order to prevent a bottleneck. This allows smoother traffic flows where
multiple clients are communicating with a server at the same time. Memory
buffering is required on an asymmetric switch. The use of buffers keeps the
frames contiguous between different data rate ports.
26.2
Introduction to LAN Switching
26.2.9
Memory buffering
This page will
explain what a memory buffer is and how it is used.
An Ethernet
switch may use a buffering technique to store and forward frames. Buffering may
also be used when the destination port is busy. The area of memory where the
switch stores the data is called the memory buffer. This memory buffer can use
two methods for forwarding frames, port-based memory buffering and shared
memory buffering.
In port-based
memory buffering frames are stored in queues that are linked to specific
incoming ports. A frame is transmitted to the outgoing port only when all the
frames ahead of it in the queue have been successfully transmitted. It is
possible for a single frame to delay the transmission of all the frames in
memory because of a busy destination port. This delay occurs even if the other
frames could be transmitted to open destination ports.
Shared memory
buffering deposits all frames into a common memory buffer which all the ports
on the switch share. The amount of buffer memory required by a port is
dynamically allocated. The frames in the buffer are linked dynamically to the
destination port. This allows the packet to be received on one port and then
transmitted on another port, without moving it to a different queue.
The switch keeps
a map of frame to port links showing where a packet needs to be transmitted.
The map link is cleared after the frame has been successfully transmitted. The
memory buffer is shared. The number of frames stored in the buffer is
restricted by the size of the entire memory buffer, and not limited to a single
port buffer. This permits larger frames to be transmitted with fewer dropped
frames. This is important to asymmetric switching, where frames are being
exchanged between different rate ports.
26.2
Introduction to LAN Switching
26.2.10
Two switching methods
This page will
introduce store-and-forward and cut-through switching.
The following two
switching modes are available to forward frames:
- Store-and-forward - The entire
frame is received before any forwarding takes place. The destination and
source addresses are read and filters are applied before the frame is
forwarded. Latency occurs while the frame is being received. Latency is
greater with larger frames because the entire frame must be received
before the switching process begins. The switch is able to check the
entire frame for errors, which allows more error detection.
- Cut-through - The frame is
forwarded through the switch before the entire frame is received. At a
minimum the frame destination address must be read before the frame can be
forwarded. This mode decreases the latency of the transmission, but also
reduces error detection.
The following are
two forms of cut-through switching:
- Fast-forward - Fast-forward
switching offers the lowest level of latency. Fast-forward switching
immediately forwards a packet after reading the destination address.
Because fast-forward switching starts forwarding before the entire packet
is received, there may be times when packets are relayed with errors.
Although this occurs infrequently and the destination network adapter will
discard the faulty packet upon receipt. In fast-forward mode, latency is
measured from the first bit received to the first bit transmitted.
- Fragment-free - Fragment-free
switching filters out collision fragments before forwarding begins.
Collision fragments are the majority of packet errors. In a properly
functioning network, collision fragments must be smaller than 64 bytes.
Anything greater than 64 bytes is a valid packet and is usually received
without error. Fragment-free switching waits until the packet is
determined not to be a collision fragment before forwarding. In
fragment-free mode, latency is also measured from the first bit received
to the first bit transmitted.
The latency of
each switching mode depends on how the switch forwards the frames. To
accomplish faster frame forwarding, the switch reduces the time for error
checking. However, reducing the error checking time can lead to a higher number
of retransmissions.
26.3
Switch Operation
26.3.1
Functions of Ethernet switches
This page will
review the functions of an Ethernet switch.
A switch is a
device that connects LAN segments using a table of MAC addresses to determine
the segment on which a frame needs to be transmitted. Both switches and bridges
operate at Layer 2 of the OSI model.
Switches are
sometimes called multiport bridges or switching hubs. Switches make decisions
based on MAC addresses and therefore, are Layer 2 devices. In contrast, hubs regenerate the Layer 1
signals out of all ports without making any decisions. Since a switch has the
capacity to make path selection decisions, the LAN becomes much more efficient.
Usually, in an Ethernet network the workstations are connected directly to the
switch. Switches learn which hosts are connected to a port by reading the
source MAC address in frames. The switch opens a virtual circuit between the
source and destination nodes only. This confines communication to those two
ports without affecting traffic on other ports. In contrast, a hub forwards
data out all of its ports so that all hosts see the data and must process it,
even if that data is not intended for it.
High-performance LANs are usually fully switched:
- A switch concentrates
connectivity, making data transmission more efficient. Frames are switched
from incoming ports to outgoing ports. Each port or interface can provide
the full bandwidth of the connection to the host.
- On a typical Ethernet hub, all
ports connect to a common backplane or physical connection within the hub,
and all devices attached to the hub share the bandwidth of the network. If
two stations establish a session that uses a significant level of
bandwidth, the network performance of all other stations attached to the
hub is degraded.
- To reduce degradation, the
switch treats each interface as an individual segment. When stations on
different interfaces need to communicate, the switch forwards frames at
wire speed from one interface to the other, to ensure that each session
receives full bandwidth.
To efficiently
switch frames between interfaces, the switch maintains an address table. When a
frame enters the switch, it associates the MAC address of the sending station
with the interface on which it was received.
The main features
of Ethernet switches are:
- Isolate traffic among segments
- Achieve greater amount of
bandwidth per user by creating smaller collision domains
The first
feature, isolate traffic among segments, provides for greater security for
hosts on the network. Each segment uses the CSMA/CD access method to maintain
data traffic flow among the users on that segment. Such segmentation allows
multiple users to send information at the same time on the different segments
without slowing down the network.
By using the
segments in the network fewer users and/or devices are sharing the same
bandwidth when communicating with one another. Each segment has its own
collision domain. Ethernet switches
filter the traffic by redirecting the datagrams to the correct port or ports,
which are based on Layer 2 MAC addresses.
The second
feature is called microsegmentation. Microsegmentation allows the creation of
dedicated network segments with one host per segment. Each hosts receives
access to the full bandwidth and does not have to compete for available
bandwidth with other hosts. Popular servers can then be placed on individual
100-Mbps links. Often in networks of today, a Fast Ethernet switch will act as
the backbone of the LAN, with Ethernet hubs, Ethernet switches, or Fast
Ethernet hubs providing the desktop connections in workgroups. As demanding new
applications such as desktop multimedia or video conferencing become more
popular, certain individual desktop computers will have dedicated 100-Mbps links
to the network.
26.3
Switch Operation
26.3.2
Frame transmission modes
This page will
describe the three main frame transmission modes:
·
Cut-through - A switch that performs
cut-through switching only reads the destination address when receiving the
frame. The switch begins to forward the frame before the entire frame arrives.
This mode decreases the latency of the transmission, but has poor error
detection. There are two forms of cut-through switching:
1.
Fast-forward switching - This type of switching offers the lowest level of
latency by immediately forwarding a packet after receiving the destination
address. Latency is measured from the first bit received to the first bit
transmitted, or first in first out (FIFO). This mode has poor LAN switching
error detection.
2.
Fragment-free switching - This type of switching filters out collision
fragments, with are the majority of packet errors, before forwarding begins.
Usually, collision fragments are smaller than 64 bytes. Fragment-free switching
waits until the received packet has been determined not to be a collision
fragment before forwarding the packet. Latency is also measured as FIFO.
·
Store-and-forward - The entire frame
is received before any forwarding takes place. The destination and source
addresses are read and filters are applied before the frame is forwarded.
Latency occurs while the frame is being received. Latency is greater with
larger frames because the entire frame must be received before the switching
process begins. The switch has time available to check for errors, which allows
more error detection.
·
Adaptive cut-through - This
transmission mode is a hybrid mode that is a combination of cut-through and
store-and-forward. In this mode, the switch uses cut-through until it detects a
given number of errors. Once the error threshold is reached, the switch changes
to store-and-forward mode.
26.3
Switch Operation
26.3.3
How switches and bridges learn addresses
This page will
explain how bridges and switches learn addresses and forward frames.
Bridges and
switches only forward frames that need to travel from one LAN segment to
another. To accomplish this task, they must learn which devices are connected
to which LAN segment.
A bridge is
considered an intelligent device because it can make decisions based on MAC
addresses. To do this, a bridge refers to an address table. When a bridge is
turned on, broadcast messages are transmitted asking all the stations on the
local segment of the network to respond. As the stations return the broadcast
message, the bridge builds a table of local addresses. This process is called
learning.
Bridges and
switches learn in the following ways:
Reading the
source MAC address of each received frame or datagram
Recording the port
on which the MAC address was received
In this way, the
bridge or switch learns which addresses belong to the devices connected to each
port.
The learned
addresses and associated port or interface are stored in the addressing table.
The bridge examines the destination address of all received frames. The bridge
then scans the address table searching for the destination address.
The switching
table is stored using Content Addressable Memory (CAM). CAM is used in switch
applications to perform the following functions:
To take out and
process the address information from incoming data packets
To compare the
destination address with a table of addresses stored within it
The CAM stores
host MAC addresses and associated port numbers. The CAM compares the received
destination MAC address against the CAM table contents. If the comparison
yields a match, the port is provided, and the switch forwards the packet to the
correct port and address.
An Ethernet
switch can learn the address of each device on the network by reading the
source address of each frame transmitted and noting the port where the frame
entered the switch. The switch then adds this information to its forwarding
database. Addresses are learned dynamically. This means that as new addresses
are read, they are learned and stored in CAM. When a source address is not
found in CAM, it is learned and stored for future use.
Each time an
address is stored, it is time stamped. This allows for addresses to be stored
for a set period of time. Each time an address is referenced or found in CAM,
it receives a new time stamp. Addresses that are not referenced during a set
period of time are removed from the list. By removing aged or old addresses,
CAM maintains an accurate and functional forwarding database.
The processes
followed by the CAM are as follows:
If the address is
not found, the bridge forwards the frame out all ports except the port on which
it was received. This process is called flooding. The address may also have
been deleted by the bridge because the bridge software was recently restarted,
ran short of address entries in the address table, or deleted the address
because it was too old. Since the bridge does not know which port to use to
forward the frame, it will send it to out all ports, except the one from which
it was received. It is clearly unnecessary to send it back to the same cable
segment from which it was received, since any other computer or bridges on this
cable must already have received the packet.
If the address is
found in an address table and the address is associated with the port on which
it was received, the frame is discarded. It must already have been received by
the destination.
If the address is
found in an address table and the address is not associated with the port on which
it was received, the bridge forwards the frame to the port associated with the
address.
If the address is
found in an address table and the address is not associated with the port on
which it was received, the bridge forwards the frame to the port associated
with the address.
26.3
Switch Operation
26.3.4
How switches and bridges filter frames
This page will
explain how switches and bridges filter frames. In this discussion, the terms
“switch” and “bridge” are synonymous.
Most switches are
capable of filtering frames based on any Layer 2 frame field. For example, a
switch can be programmed to reject, not forward, all frames sourced from a
particular network. Because link layer information often includes a reference
to an upper-layer protocol, switches can usually filter on this parameter.
Furthermore, filters can be helpful in dealing with unnecessary broadcast and
multicast packets.
Once the switch
has built the local address table, it is ready to operate. When it receives a
frame, it examines the destination address. If the frame address is local, the
switch ignores it. If the frame is addressed for another LAN segment, the
switch copies the frame onto the second segment.
- Ignoring a frame is called
filtering.
- Copying the frame is called
forwarding.
Basic filtering
keeps local frames local and sends remote frames to another LAN segment.
Filtering on
specific source and destination addresses performs the following actions:
- Stopping one station from
sending frames outside of its local LAN segment
- Stopping all
"outside" frames destined for a particular station, thereby
restricting the other stations with which it can communicate
Both types of
filtering provide some control over internetwork traffic and can offer improved
security.
Most Ethernet
switches can now filter broadcast and multicast frames. Bridges and switches
that can filter frames based on MAC addresses can also be used to filter
Ethernet frames by multicast and broadcast addresses. This filtering is
achieved through the implementation of virtual local-area networks or VLANs.
VLANs allow network administrators to prevent the transmission of unnecessary
multicast and broadcast messages throughout a network. Occasionally, a device
will malfunction and continually send out broadcast frames, which are copied
around the network. This is called a broadcast storm and it can significantly
reduce network performance. A switch that can filter broadcast frames makes a
broadcast storm less harmful.
Today, switches
are also able to filter according to the network-layer protocol. This blurs the
demarcation between switches and routers. A router operates on the network
layer using a routing protocol to direct traffic around the network. A switch
that implements advanced filtering techniques is usually called a brouter.
Brouters filter by looking at network layer information but they do not use a
routing protocol.
26.3
Switch Operation
26.3.5
Why segment LANs?
This page will
explain the two main reasons to segment a LAN.
There are two
primary reasons for segmenting a LAN. The first is to isolate traffic between
segments. The second reason is to achieve more bandwidth per user by creating
smaller collision domains.
Without LAN
segmentation, LANs larger than a small workgroup could quickly become clogged
with traffic and collisions.
LAN segmentation
can be implemented through the utilization of bridges, switches, and routers.
Each of these devices has particular pros and cons.
With the addition
of devices like bridges, switches, and routers the LAN is segmented into a
number of smaller collision domains. In the example shown, four collision
domains have been created.
By dividing large
networks into self-contained units, bridges and switches provide several
advantages. Bridges and switches will diminish the traffic experienced by
devices on all connected segments, because only a certain percentage of traffic
is forwarded. Bridges and switches reduce the collision domain but not the
broadcast domain.
Each interface on
the router connects to a separate network. Therefore the insertion of the
router into a LAN will create smaller collision domains and smaller broadcast
domains. This occurs because routers do not forward broadcasts unless
programmed to do so.
A switch employs
"microsegmentation" to reduce the collision domain on a LAN. The
switch does this by creating dedicated network segments, or point-to-point
connections. The switch connects these segments in a virtual network within the
switch.
This virtual
network circuit exists only when two nodes need to communicate. This is called
a virtual circuit as it exists only when needed, and is established within the
switch.
26.3
Switch Operation
26.3.6
Microsegmentation implementation
This page will explain
the functions of a switch in a LAN due to microsegmentation.
LAN switches are
considered multi-port bridges with no collision domain, because of
microsegmentation. Data is exchanged at
high speeds by switching the frame to its destination. By reading the
destination MAC address Layer 2 information, switches can achieve high-speed
data transfers, much like a bridge does. This process leads to low latency
levels and a high rate of speed for frame forwarding.
Ethernet
switching increases the bandwidth available on a network. It does this by
creating dedicated network segments, or point-to-point connections, and
connecting these segments in a virtual network within the switch. This virtual
network circuit exists only when two nodes need to communicate. This is called
a virtual circuit because it exists only when needed, and is established within
the switch.
Even though the
LAN switch reduces the size of collision domains, all hosts connected to the
switch are still in the same broadcast domain. Therefore, a broadcast from one
node will still be seen by all the other nodes connected through the LAN
switch.
Switches are data
link layer devices that, like bridges, enable multiple physical LAN segments to
be interconnected into a single larger network. Similar to bridges, switches
forward and flood traffic based on MAC addresses. Because switching is
performed in hardware instead of in software, it is significantly faster. Each
switch port can be considered a micro-bridge acting as a separate bridge and gives
the full bandwidth of the medium to each host.
26.3
Switch Operation
26.3.7
Switches and collision domains
This page will
discuss collisions, which is a major disadvantage of Ethernet 802.3 networks.
A major
disadvantage of Ethernet 802.3 networks is collisions. Collisions occur when
two hosts transmit frames simultaneously. When a collision occurs, the
transmitted frames are corrupted or destroyed in the collision. The sending
hosts stop sending further transmissions for a random period of time, based on
the Ethernet 802.3 rules of CSMA/CD. Excessive collisions cause networks to be
unproductive.
The network area
where frames originate and collide is called the collision domain. All shared
media environments are collision domains.
When a host is connected to a switch port, the switch creates a
dedicated connection. This connection is considered to be an individual
collision domain. For example, if a twelve-port switch has a device connected
to each port then twelve collision domains are created.
A switch builds a
switching table by learning the MAC addresses of the hosts that are connected
to each switch port. When two connected
hosts want to communicate with each other, the switch looks up the switching table
and establishes a virtual connection between the ports. The virtual circuit is
maintained until the session is terminated.
In Figure , Host
B and Host C want to communicate with each other. The switch creates the
virtual connection which is referred to as a microsegment. The microsegment
behaves as if the network has only two hosts, one host sending and one
receiving providing maximum utilization of the available bandwidth.
Switches reduce
collisions and increase bandwidth on network segments because they provide
dedicated bandwidth to each network segment.
26.3
Switch Operation
26.3.8
Switches and broadcast domains
This page will
describe the three methods of data transmission that are used in a network.
Communication in
a network occurs in three ways. The most common way of communication is by
unicast transmissions. In a unicast transmission, one transmitter tries to
reach one receiver.
Another way to
communicate is known as a multicast transmission. Multicast transmission occurs
when one transmitter tries to reach only a subset, or a group, of the entire
segment.
The final way to
communicate is by broadcasting. Broadcasting is when one transmitter tries to
reach all the receivers in the network. The server station sends out one
message and everyone on that segment receives the message.
When a device
wants to send out a Layer 2 broadcast, the destination MAC address in the frame
is set to all ones. A MAC address of all ones is FF:FF:FF:FF:FF:FF in
hexadecimal. By setting the destination to this value, all the devices will
accept and process the broadcasted frame.
The broadcast
domain at Layer 2 in referred to as the MAC broadcast domain. The MAC broadcast
domain consists of all devices on the LAN that receive frame broadcasts by a
host to all other machines on the LAN.
A switch is a
Layer 2 device. When a switch receives a broadcast, it forwards it to each port
on the switch except the incoming port. Each attached device must process the
broadcast frame. This leads to reduced network efficiency, because available
bandwidth is used for broadcasting purposes.
When two switches
are connected, the broadcast domain is increased. In this example a broadcast
frame is forwarded to all connected ports on Switch 1. Switch 1 is connected to
Switch 2. The frame is propagated to all devices connected to Switch 2.
The overall
result is a reduction in available bandwidth. This happens because all devices
in the broadcast domain must receive and process the broadcast frame.
Routers are Layer
3 devices. Routers do not propagate broadcasts. Routers are used to segment
both collision and broadcast domains.
26.3
Switch Operation
26.3.9
Communication between switches and workstations
This page will
explain how switches learn about workstations in a LAN.
When a
workstation connects to a LAN, it is unconcerned about the other devices that
are connected to the LAN media. The workstation simply transmits data frames
using a NIC to the network medium.
The workstation
could be attached directly to another workstation using a crossover cable.
Cross-over cables are used to connect the following devices:
- Workstation to Workstation
- Switch to Switch
- Switch to hub
- Hub to hub
- Router to router
- Router to PC
Straight-through
cables are used to connect the following devices:
- Switch to router
- Switch to workstation or server
- Hub to workstation or server
Switches are
Layer 2 devices that use intelligence to learn the MAC addresses of the devices
that are attached to the ports of the switch. This data is entered into a switching
table. Once the table is complete, the switch can read the destination MAC
address of an incoming data frame on a port and immediately forward it. Until a device transmits, the switch does not
know its MAC address.
Switches provide
significant scalability on a network and may be directly connected. Figure illustrates one scenario of frame
transmission utilizing a multi-switch network.
Summary
This page
summarizes the topics discussed in this module.
Ethernet is the
most common LAN architecture and it is used to transport data between devices
on a network. Originally Ethernet was a half-duplex technology. Using
half-duplex, a host could either transmit or receive at one time, but not both.
When two or more Ethernet hosts transmit at the same time on a shared medium,
the result is a collision. The time a frame or a packet takes to travel from
the source station to the final destination is known as latency or delay. The
three sources of latency include NIC delay, actual propagation delay, and delay
due to specific network devices.
Bit or slot time
is the basic unit of time in which ONE bit can be sent. there must be some
minimum duration during which the bit is on or off in order for the device to
recognize a binary one or zero.
Attenuation means
that a signal will weaken at it travels through the network. This limits the
distance that a LAN can cover. A repeater can extend the distance of a LAN but
it also has a negative effect on the overall performance of a LAN.
Full-duplex
transmission between stations is achieved by using point-to-point Ethernet
connections. Full-duplex transmission provides a collision-free transmission
environment. Both stations can transmit and receive at the same time, and there
are no negotiations for bandwidth. The existing cable infrastructure can be
utilized as long as the medium meets the minimum Ethernet standards.
Segmentation
divides a network into smaller units to reduce network congestion and enhance
security. The CSMA/CD access method on each segment maintains traffic between
users. Segmentation with a Layer 2 bridge is transparent to other network
devices but latency is increased significantly. The more work done by a network
device, the more latency the device will introduce into the network. Routers
provide segmentation of networks but can add a latency factor of 20% to 30%
over a switched network. This increased latency is because a router operates at
the network layer and uses the IP address to determine the best path to the
destination node. A switch can segment a LAN into microsegments which decreases
the size of collision domains. However all hosts connected to the switch are
still in the same broadcast domain.
Switching is a
technology that decreases congestion in Ethernet, Token Ring, and Fiber
Distributed Data Interface (FDDI) LANs. Switching is the process of receiving
an incoming frame on one interface and delivering that frame out another
interface. Routers use Layer 3 switching to route a packet. Switches use Layer
2 switching to forward frames. A symmetric switch provides switched connections
between ports with the same bandwidth. An asymmetric LAN switch provides
switched connections between ports of unlike bandwidth, such as a combination
of 10-Mbps and 100-Mbps ports.
A memory buffer
is an area of memory where a switch stores data. It can use two methods for
forwarding frames including port-based memory buffering and shared memory
buffering.
There are two
modes used to forward frames. Store-and-forward receives the entire frame
before forwarding while cut-through forwards the frame as it is received
decreasing latency. Fast-forward and fragment-free are two types of cut-through
forwarding.
No comments:
Post a Comment