Transcript Semester 3 Chapter 2 - IIS Windows Server

Institute of Technology,
Sligo Dept of Computing
LAN Switching
Semester 3, Chapter 2
Table of Contents

LAN Communication Problems
• Full-Duplex, Fast Ethernet,
and Segmentation
• Switching and VLANs
• The Spanning-Tree Protocol
Institute of Technology,
Sligo Dept of Computing
LAN Communication
Problems
Network Performance

Network congestion has increased significantly
since the mid-90s due to:

Multitasking Operating Systems


Faster Processing Power


multiple simultaneous network transactions (e.g., ftp
download & browsing)
1980s: 1 MIPS; Today: over 75 MIPS
Network-intensive Applications

accessing network servers to use applications, files, etc.
Elements of Ethernet/802.3

Characteristics




Most common LAN architecture
Used to transport data between devices connected to
the same delivery medium
Uses a data frame broadcast method
Negative effects of a shared LAN



broadcast delivery of all frames
CSMA/CD: collisions are inherent
distance limitation requires using repeaters to extend
Half-Duplex Ethernet

Properties


Only one host can transmit at a time because
the NIC needs to listen for collisions
The NIC provides several circuits. Most
important are:


receive (RX), transmit (TX), and collision detection
bandwidth usage = 50% to 60%

Operation






CSMA/CD
Devices on shared media listen for a carrier before
transmitting
If no carrier is sensed for a specific period of time, a
device can transmit
If two devices transmit simultaneously, a collision occurs.
The NIC senses this because it is transmitting and
receiving at the same time
The first device to detect the collision will generate a jam
signal (colliding devices continue to transmit so that all
devices will hear the collision)
All devices calculate a backoff algorithm which will delay
transmission for a random length of time.
First device who’s delay time expires can attempt to
transmit data.
Network Congestion

Occurs as more people utilize a network to...




Share large files (e.g. databases, applications, etc.)
Access file servers
connect to the Internet
Relieving congestion requires


Increasing the amount of bandwidth and/or
Using available bandwidth more efficiently
Network Latency

Latency explained




Represents the time it takes a frame to travel from is
source device to its final destination on the network (also
know as propagation delay)
Latency can also be described as the delay between the
time a device requests access to a network and the time
it is granted permission to transmit
For switches and routers, latency is the amount of delay
between the time when the device receives the frame on
one interface and forwards that frame out another
interface
Routers have more inherent latency than a switch. Why?
Ethernet Transmission Time

Defined


Transmission time is the time necessary to move a
packet from the data link layer to the physical layer
10BaseT Transmission Time

Each bit has a 100ns window for transmission





ns-nanosecond (1 billionth of a second)
So each byte has what size window?
A 64 byte frame (the smallest allowed frame) requires
51,200 ns or 51.2 microseconds
Just to frame a 1000 byte packet requires 800
microseconds
Additional latency will be added propagating the
frame down the wire and by any additional devices
the frame has to go through before reaching the
destination
Using Repeaters

What is attenuation?


Benefits of Using a Repeater



Loss of signal strength as it travels through the
network; caused by resistance inherent in the medium
a layer 1 device that cleans up and boosts the signal
extends the coverage area of a LAN segment
Negative Effects of Using a Repeater



increases the collision domain size
increases the broadcast domain size
can’t filter traffic based on Layer 2 or 3 addressing
Institute of Technology,
Sligo Dept of Computing
Full-Duplex,
Fast Ethernet,
and Segmentation
Full-Duplex Ethernet

Simultaneous TX and RX





allows the transmission of a packet and the reception
of a different packet at the same time.
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.
100% of bandwidth is available: 10 Mbps increases to
20 Mbps of potential throughput (10 Mbps TX & 10
Mbps RX)
LAN Segmentation

Benefits of Segmenting the Network
By segmenting a LAN
fewer devices are
sharing the same
bandwidth, improving
performance of a
shared media LAN
Each segment is
considered its own
collision domain
How many broadcast
domains in graphic?
Segmenting with Bridges

Bridge Operation

Bridges “learn” a network’s segmentation by
building address tables that contain:


Bridge interface that will reach that device
Each device’s MAC address
Segmenting with Bridges

Generic Frame Format




Frame can be any length depending on
technology
Ethernet frame can be up to 1522 bytes long
Address section is 12 bytes (6 bytes for each
MAC)
FCS contain the CRC to check frame for
errors
Start
Frame
Address
Type/
Length
Data
FCS
Stop
Frame
Segmenting with Bridges

Bridge Performance



adds 10% to 30% latency due to decision-making
process
considered a store-and-forward device because it
must calculate the CRC at the end of the frame to
check it for errors before forwarding
if the bridge does not have an entry for the destination
MAC, it...



adds the source MAC to its bridging table
forwards the frame out all interfaces except the one it was
received on
when a reply returns, it adds the destination MAC to the table
Segmenting with Routers

Router Operation

Routers...




use layer 3 addressing (IP, IPX) and routing protocols (RIP,
IGRP) to determine the path and
switch the packet out the correct interface to the destination
because a router must open the packet to read Layer
3 addressing, it adds latency
In addition, protocols like TCP which require
acknowledgments of every packet can increase
latency, reducing throughput from 20% to 40%

Segmenting
with
Routers
Router Benefits



Like switches,
routers segment
collision domains.
However, since a
router will not
forward broadcasts,
it also segments
broadcast domains.
Each router interface
represents its own
broadcast domain.
Segmenting with Switches

Switching Benefits





a switch is simply a multi-port bridge, making
forwarding decisions based on MAC addresses
so, like a bridge, segmenting a LAN with a switch
creates more collision domains
replacing hubs with switches therefore decreases
congestion and increases available bandwidth.
a switch can microsegment a LAN creating collisionfree domains but still be in the same broadcast
domain.
switch creates a virtual circuits, allowing many users
to communicate in parallel.
Institute of Technology,
Sligo Dept of Computing
Switching and VLANs
Switch Operation

Switches perform two basic functions:



Building and maintaining switching tables (similar to a
bridge table) based on MAC addresses
Switching frames out the interface to the destination
Differences between switches & bridges



Switches operate at higher speeds
Switches are capable of creating virtual LANs
(VLANs) through microsegmentation
Bridges switch using software; switches typically
switch using hardware (called the “switch fabric”)
Switch Latency (Delay)



A switch adds 21 microseconds of latency.
This can be reduced by using a different
switching method
As opposed to store-and-forward, the switch can
use cut-through switching which switches the
packet as soon as the destination MAC is read.
How a LAN Switch Learns
Addresses





MAC addresses are learned dynamically and are
stored in CAM (content-addressable memory)
Each time a switch stores an address entry in the
table, it is time-stamped.
The time-stamp is updated each time a frame is
received
Addresses whose time-stamp expires are deleted
from the table
This keeps switching tables small
Benefits of LAN Switching





Cost-effective;
switches only cost 3
to 5 times that of a
hub
Allows the creation
of virtual circuits
More flexibility in
managing the
network
Reduces number of
collisions
Works with existing
802.3 cabling
Symmetric Switching
potential bottlenecks


symmetric switching
provides switched
connections between
ports with the same
bandwidth (10/10
Mbps or 100/100
Mbps)
can cause
bottlenecks as users
try to access servers
on other segments.
Asymmetric Switching


asymmetric switching
reduces the
likelihood of a
potential bottleneck
at the server by
attaching the
segment with the
server to a higher
bandwidth port (100
Mbps)
asymmetric switching
requires memory
buffering in the
switch
Memory Buffering

Defined


Area of memory in a switch where destination and
transmission data are stored until it can be switched
out the correct port.
Two types

Port-based memory buffering



packets are stored in a queue on each port
possible for one packet to delay transmission of other
packets because of a busy destination port
Shared memory buffering


common memory buffering shared by all ports
allows packets to be RX on one port and TX out another port
without changing it to a different queue.
Two Switching Methods

Store-and-Forward


The switch receives the
entire frame, calculating the
CRC at the end, before
sending it to the destination
Cut-through


Fast forward switching--only
checks the destination MAC
before immediately
forwarding the frame
Fragment Free--reads the
first 64 bytes to reduce
errors before forwarding the
frame
VLANs (IEEE 802.1q)

Characteristics




A logical grouping of network devices or users that
are not restricted to a physical switch segment.
The devices or users in a VLAN can be grouped by
function, department, application, and so on,
regardless of their physical segment location.
A VLAN creates a single broadcast domain that is not
restricted to a physical segment and is treated like a
subnet.
VLAN setup is done in the switch by the network
administrator using the vendor’s software.
Institute of Technology,
Sligo Dept of Computing
The Spanning-Tree Protocol
Overview of STP

Elements of the Spanning Tree Protocol




Main function of STP is to allow redundant paths in a
switched/bridged network without incurring latency from
the effects of loops.
STP prevents loops by calculating a stable spanningtree network topology (similar to OSPF operation)
Spanning-tree frames (called bridge protocol data units-BPDUs) are sent and received by all switches in the
network and are used to determine the spanning-tree
topology
STP operation is covered in detail in Semester 7 of the
CCNP curriculum.
Five STP States



States are established by configuring each port
according to policy
Then the STP modifies the states based on traffic
patterns and potential loops
The default order of STP states are:





Blocking--no frames forwarded, BPDUs heard
Listening--no frames forwarded, listening for data frames
Learning--no frames forwarded, learning addresses
Forwarding--frames forwarded, learning addresses
Disabled--no frames forwarded, no BPDUs heard