Ethernet Switching 359
Figure 6-30 Transmitting Data to a Known Station
The original transmission indicated where that MAC address came from, enabling the
switch to deliver traffic in the network more efficiently.
Microsegmentation
As technology improved, it became possible to build bridges with more capability.
Naturally, the ultimate goal is to have only one node on each port of a bridge. This
would reduce the collision domains so that collisions would be nearly nonexistent. A
switch does exactly that and is simply a bridge with many ports. These small physical
segments are called microsegments.
Microsegmentation facilitates the creation of a dedicated segment and provides dedi-
cated bandwidth to each user on the network. Each user receives instant access to the
full bandwidth and does not have to contend for available bandwidth with other users.
This means that pairs of devices on the same switch can communicate in parallel with
a minimum number of collisions. Microsegmentation reduces collisions in a network
and effectively increases the capacity for each station connected to the network.
In addition to faster microprocessors and memory, two other technological advances
made switches possible. Content Addressable Memory (CAM) is memory that essentially
works backward compared to conventional memory. Entering data into the memory
returns the associated address. Using CAM allows a switch to directly find the port
that is associated with a MAC address without using search algorithms. An application-
specific integrated circuit (ASIC) is a device consisting of undedicated logic gates that
can be programmed to perform functions at logic speeds. Operations that might have
been done in software now can be done in hardware using an ASIC. The use of these
technologies greatly reduced the delays caused by software processing and enabled a
switch to keep pace with the data demands of many microsegments and high bit rates.
1102.book Page 359 Tuesday, May 20, 2003 2:53 PM
360 Chapter 6: Ethernet Technologies and Ethernet Switching
Full-Duplex Transmission
Another function of LAN switching that dramatically improves bandwidth is full-
duplex transmission, which effectively doubles the amount of bandwidth between

nodes. Full-duplex transmission between stations is achieved by using point-to-point
Ethernet connections. This feature can be important, for example, between high-band-
width consumers, such as a connection between a switch and a server. Full-duplex
transmission provides a collision-free transmission environment. Because both nodes
can transmit and receive at the same time, there are no negotiations for bandwidth.
In 10-Mbps connections, for example, full-duplex transmission provides 10 Mb of
transmit capacity and 10 Mb of receive capacity, for effectively 20 Mb of capacity on
a single connection. Likewise, a 100-Mbps connection offers effectively 200 Mbps of
throughput, as illustrated in Figure 6-31. Full-duplex communication also supports
two data transmission paths, with speeds up to 1 Gbps.
Figure 6-31 Switching Technology: Full Duplex
Latency
Latency, sometimes called propagation delay, is the time that a frame, or packet, of
data takes to travel from the source station or node to its final destination on the net-
work. A wide variety of conditions can cause delays as a frame travels from source to
destination:
■ Media delays caused by the finite speed that signals can travel through the physi-
cal media.
■ Circuit delays caused by the electronics that process the signal along the path.
■ Software delays caused by the decisions that software must make to implement
switching and protocols.
■ Delays caused by the content of the frame and where in the frame switching deci-
sions can be made. For example, a device cannot route a frame to a destination
until the destination MAC address has been read.
Latency is the time delay between when a frame first starts to leave the source device
and when the first part of the frame reaches its destination.
1102.book Page 360 Tuesday, May 20, 2003 2:53 PM
Ethernet Switching 361
Switch Modes
How the content of a frame is switched to the destination port is a trade-off of latency

and reliability. The three modes of switching—store-and-forward, cut-through, and
fragment-free switching—offer different performance and latency.
Store-and-Forward Switching
In store-and-forward switching, the switch reads the entire frame of data, checks the
frame for errors, decides where it needs to go, and then sends it on its way. Figure 6-32
illustrates the operation of store-and-forward switching. The obvious trade-off here
is that it takes the switch longer to read the entire frame. As it reads the entire frame,
however, it detects any errors on that frame. If the frame is in error, the frame is not
forwarded and is discarded. Although cut-through switching is faster, it offers no error
detection. The latency introduced by store-and-forward switching is usually not a sig-
nificant issue.
Figure 6-32 Store-and-Forward Switching
Cut-Through Switching
In cut-through switching, the switch reads the beginning of the frame up to the destina-
tion MAC address as the traffic flows through the switch and “cuts through” to its des-
tination without continuing to read the rest of the frame, as illustrated in Figure 6-33.
Cut-through switching decreases the latency of the transmission. However, cut-through
switching has no error detection.
Source Destination
Checked the
frame. It is
good. I am to
forward the
frame now.
1102.book Page 361 Tuesday, May 20, 2003 2:53 PM
362 Chapter 6: Ethernet Technologies and Ethernet Switching
Figure 6-33 Cut-Through Switching
Fragment-Free Switching
Fragment-free switching is a modified form of cut-through switching. Fragment-free
switching filters out collision fragments, which are the majority of packet errors, before

forwarding begins. Fragment-free switching waits until the received packet has been
determined not to be a collision fragment before forwarding the packet.
When to Use Each Switching Mode
When using cut-through and fragment-free modes, both the source port and the desti-
nation port must be operating at the same bit rate to keep the frame intact. This is
called synchronous switching. If the bit rates are not the same, the frame must be stored
at one bit rate before it is sent out at the other bit rate. This is known as asynchronous
switching. Store-and-forward mode must be used for asynchronous switching. Asym-
metric switching provides switched connections between ports of unlike bandwidths,
such as a combination of 100 Mbps and 1000 Mbps. Asymmetric switching is opti-
mized for client/server traffic flows in which multiple clients simultaneously communi-
cate with a server, requiring more bandwidth dedicated to the server port to prevent a
bottleneck at that port.
Introduction to the Spanning Tree Protocol
When multiple switches are connected, there is a possibility of creating a loop where
there is no clear path from source to destination. If switches are arranged in a simple
hierarchical tree, no loops will occur, as shown in Figure 6-34.
Source Destination
I have the
destination
MAC
address. I
can transmit
now.
1102.book Page 362 Tuesday, May 20, 2003 2:53 PM
Ethernet Switching 363
Figure 6-34 STP Reducing Routing Loops
However, when extra switches and bridges are added to provide redundant paths for
reliability and fault tolerance, loops can occur, as shown in Figure 6-35.
Figure 6-35 Broadcast Storms

Hub
Cat-6
Cat-7
Cat-1
Cat-3
Cat-2
Cat-4
Cat-5
Wiring Closet Backbone
Server Farm
1102.book Page 363 Tuesday, May 20, 2003 2:53 PM
364 Chapter 6: Ethernet Technologies and Ethernet Switching
In Figure 6-35, the following steps are occurring:
1. Host 1 broadcasts.
2. SWA and SWB receive the frame.
3. SWA forwards to SWB.
4. SWB forwards to SWA.
5. Each switch now begins to see multiple frames of the same broadcast, thus
resulting in routing loops.
To counteract the possibility of loops, switches are provided with a protocol for them
to talk with each other to resolve the condition. A switch sends special messages called
bridge protocol data units (BPDUs) out all its ports to let other switches know of its
existence, as shown in Figures 6-36 and 6-37. The switches use a spanning tree algo-
rithm (STA) to resolve and shut down the redundant paths. The process of shutting
down a port is called blocking. The result of resolving and eliminating the loops is a
logical hierarchical tree created with no loops. However, the alternate paths are still
available, in case they are needed. The protocol used to resolve and eliminate loops is
known as the Spanning Tree Protocol (STP). This creates another switch and bridge
operation mode known as loop-avoidance mode.
Figure 6-36 BPDU Communication

LAN Switch 1
Bridge ID 12345111
Port 2
LAN Switch 3
Bridge ID 12345556
Port 1
AA-11
Switch 1’s Bridge ID is
lower and wins the
comparison.
BPDU
Priority: 1
Root: 12345555
Path Cost: 0
Bridge ID: 12345555
Port ID: 8001
BPDU
Priority: 1
Root: 12345111
Path Cost: 0
Bridge ID: 12345111
Port ID: 8002
100 Mbps Fast
Ethernet
Segment
All Other
Segments Are
10 Mbps
1102.book Page 364 Tuesday, May 20, 2003 2:53 PM
Ethernet Switching 365

Figure 6-37 BPDU Protocol Layout
Consequently, switches have five operating modes:
■ Blocking—A port in blocking state sends and listens to BPDUs but does not for-
ward frames. By default, all ports are in blocking state when the switch is turned on.
■ Listening—In listening state, a port listens to the BPDUs to make sure there are
no loops on the network. No frames are forwarded in this state.
■ Learning—In this state, a port learns MAC addresses and builds a address table,
but it does not forward frames.
■ Forwarding—A port in the forwarding state forwards frames. BPDUs are sent
and listened to.
■ Disabled—A port in the disabled state does not participate in the operation of
STP. Therefore, it does not listen to BPDUs or forward frames.
Figure 6-38 illustrates some of the port states and operating modes in a switched net-
work using STP.
Lab Activity Introduction to Fluke Network Inspector
This lab is a tutorial demonstrating how to use the Fluke Networks Network
Inspector (NI) to discover and analyze network devices within a broadcast
domain. This lab demonstrates the key features of the tool that can be incorpo-
rated into various troubleshooting efforts in the remaining labs.
1102.book Page 365 Tuesday, May 20, 2003 2:53 PM
366 Chapter 6: Ethernet Technologies and Ethernet Switching
Figure 6-38 Port States
Summary
In this chapter, you learned the following key points:
■ Several types of Ethernet exist: Ethernet, Fast Ethernet, Gigabit Ethernet, and
10-Gb Ethernet. Each type is associated with a different transfer rate.
■ Ethernet uses carrier sense multiple access collision detect (CSMA/CD).
■ 10-Mbps Ethernet operates within the timing limits offered by a series of no
more than five segments separated by no more than four repeaters.
■ Switches can provide dedicated access to improve shared LAN technologies.

■ A switch segments a LAN into microsegments. Microsegmentation reduces the
number of collisions to a minimum and increases the effective bandwidth.
■ Switches achieve high-speed transfer by reading the destination Layer 2 MAC
address of the frame packet.
Lab Activity Introduction to Fluke Protocol Inspector
This lab is a tutorial demonstrating how to use the Fluke Networks Protocol
Inspector to analyze network traffic and data frames. This lab demonstrates
key features of the tool that can be incorporated into various troubleshooting
efforts in the remaining labs.
1102.book Page 366 Tuesday, May 20, 2003 2:53 PM
Summary 367
■ Full-duplex communication allows two devices to communicate with each other
simultaneously and effectively doubles the throughput that the LAN switch can
translate.
■ Switches can support multiple simultaneous conversations in a network.
■ Three switching modes can be used to forward frames through a switch: store-
and-forward, cut-through, and fragment-free switching.
■ The main task of the Spanning Tree Protocol (STP) is to prevent the occurrence
of network loops on a Layer 2 network.
■ The ports on a bridge or switch using STP exist in one of the following five
states: blocking, listening, learning, forwarding, or disabled.
To supplement all that you’ve learned in this chapter, refer to the chapter-specific Videos,
PhotoZooms, and e-Lab Activities on the CD-ROM accompanying this book.
1102.book Page 367 Tuesday, May 20, 2003 2:53 PM
368 Chapter 6: Ethernet Technologies and Ethernet Switching
Key Terms
10BASE2 10-Mbps baseband Ethernet specification using 50-ohm thin coaxial cable.
10BASE2, which is part of the IEEE 802.3 specification, has a distance limit of 185m
(606 ft.) per segment.
10BASE5 10-Mbps baseband Ethernet specification using standard (thick) 50-ohm

baseband coaxial cable. 10BASE5, which is part of the IEEE 802.3 baseband physical
layer specification, has a distance limit of 500m (1640 ft.) per segment.
10BASE-T 10-Mbps baseband Ethernet specification using two pairs of twisted-pair
cabling (Category 3, 4, or 5): one pair for transmitting data and the other for receiving
data. 10BASE-T, which is part of the IEEE 802.3 specification, has a distance limit of
approximately 100m (328 ft.) per segment.
100BASE-FX 100-Mbps baseband Fast Ethernet specification using two strands of
multimode fiber-optic cable per link. To guarantee proper signal timing, a 100BASE-
FX link cannot exceed 400m (1312 ft.) in length. It is based on the IEEE 802.3 standard.
100BASE-TX 100-Mbps baseband Fast Ethernet specification using two pairs of
either UTP or STP wiring. The first pair of wires is used to receive data; the second is
used to transmit. To guarantee proper signal timing, a 100BASE-TX segment cannot
exceed 100m (328 ft.) in length. It is based on the IEEE 802.3 standard.
1000BASE-T 1000-Mbps baseband Gigabit Ethernet specification using four pairs
of Category 5 UTP cable for a maximum length of 100m (328 ft.).
1000BASE-SX 1000-Mbps baseband Gigabit Ethernet specification using a short
laser wavelength on multimode fiber-optic cable for a maximum length of 550m
(1804.5 ft.).
1000BASE-LX 1000-Mbps baseband Gigabit Ethernet specification using a long wave-
length for a long-haul fiber-optic cable for a maximum length of 10,000 (32808.4 ft.).
4D-PAM5 The symbol-encoding method used in 1000BASE-T. The four-dimensional
quinary symbols (4D) received from the 8B1Q4 data encoding are transmitted using
five voltage levels (PAM5). Four symbols are transmitted in parallel each symbol
period.
8B1Q4 For IEEE802.3, the data-encoding technique used by 1000BASE-T when
converting GMII data (8B-8 bits) to four quinary symbols (Q4) that are transmitted
during one clock (1Q4).
BPDU (bridge protocol data unit) Spanning Tree Protocol hello packet that is sent
out at configurable intervals to exchange information among bridges in the network.
1102.book Page 368 Tuesday, May 20, 2003 2:53 PM