Transcript Week 1 and 2 - DePaul University

Wired LANs: Ethernet
13.1
Copyright © The McGraw-Hill Companies, Inc. Permission required for reproduction or display.
A Simple Network (LAN)
192.168.0.10
Crossover UTP
cable
Internet
192.168.0.1
140.192.40.5
2
Typical
LAN
9th floor
2nd floor
1st floor
Internet
Data center
3
Campus Area Network (CAN)
CDM Building
DePaul
Center
Administration
Building
Lewis
Center
4
Metropolitan Area Network (MAN)
Lake Forest
Lincoln Park
Rolling Meadow
Loop
Naperville
5
Wide Area Network (WAN)
WAN
6
Success of Ethernet




Easy to understand, implement, manage,
and maintain
Low-cost network implementations
Extensive topological flexibility for network
installation
Interoperability and operation of
standards-compliant products, regardless
of manufacturer
7
LAN Services


Users do not see the network, and they
are interested in the services.
What are the essential LAN services?






Printer sharing
File sharing
Application sharing
Internet Surfing
Intranet Web
Database






Backup
Fax Service
Telephony
Conferencing
Management
Miscellaneous
8
Network Components



Cables
Network Interface Cards (NIC)
Network hardware








Hubs
Switches
Routers
Gateways
Printers
Storage devices (NAS and SAN)
Network Operating System (NOS)
Software
9
Network Operating System
Software programs that control resources shared over the network.
Application
Application
Protocols
Protocols
Drivers
Drivers
Clients
Network
Server
Example: Windows, NetWare, UNIX, Linux, MacOS
10
History of Ethernet
McGraw-Hill
11
©The McGraw-Hill Companies, Inc., 2000
ALOHA Network (’68-’72)
Radio frequency w/ speed = 4.8-9.6K bps
terminal
IBM 360
Inventor:
Norman Abramson
Outbound channel: one-to-many transmission
In-bound: same channel frequency with contention
Slotted Aloha followed after Aloha. Throughput jumped from 18% to 37%.
12
First Ethernet at Xerox





Inventor: Bob Metcalfe
Location: Xerox Palo Alto Research Lab
Time: May 22, 1973, The first Local Area
Network for personal computer (called
ALTO) and printer (called EARS).
Speed: 2.94Mbps
Importance: carrier sense (listen first
before transmitting)
13
Standardization of Ethernet








Intel – Chip technology
DEC – System engineering and hardware
supplier
Xerox – Ethernet technology
1980 - The Ethernet blue book (DIX)
Speed – 10Mbps
1981 – IEEE 802.3 subcommittee
1983 – IEEE 10Base5
1989 – ISO 88023
14
Figure 13.1 IEEE standard for LANs
13.15
Figure 13.2 HDLC frame compared with LLC and MAC frames
13.16
Figure 13.3 Ethernet evolution through four generations
13.17
Figure 13.4 802.3 MAC frame
13.18
Figure 13.5 Minimum and maximum lengths
13.19
Note
Frame length:
Minimum: 64 bytes (512 bits)
Maximum: 1518 bytes (12,144 bits)
13.20
Figure 13.6 Example of an Ethernet address in hexadecimal notation
13.21
Figure 13.7 Unicast and multicast addresses
13.22
Note
The least significant bit of the first byte
defines the type of address.
If the bit is 0, the address is unicast;
otherwise, it is multicast.
13.23
Note
The broadcast destination address is a
special case of the multicast address in
which all bits are 1s.
13.24
Example 13.1
Define the type of the following destination addresses:
a. 4A:30:10:21:10:1A
b. 47:20:1B:2E:08:EE
c. FF:FF:FF:FF:FF:FF
Solution
To find the type of the address, we need to look at the
second hexadecimal digit from the left. If it is even, the
address is unicast. If it is odd, the address is multicast. If
all digits are F’s, the address is broadcast. Therefore, we
have the following:
a. This is a unicast address because A in binary is 1010.
b. This is a multicast address because 7 in binary is 0111.
c. This is a broadcast address because all digits are F’s.
13.25
Example 13.2
Show how the address 47:20:1B:2E:08:EE is sent out on
line.
Solution
The address is sent left-to-right, byte by byte; for each
byte, it is sent right-to-left, bit by bit, as shown below:
13.26
Figure 13.8 Categories of Standard Ethernet
13.27
Figure 13.9 Encoding in a Standard Ethernet implementation
13.28
Figure 13.10 10Base5 implementation
13.29
Figure 13.11 10Base2 implementation
13.30
StarLAN (1BaseT)

Problem with 10Base5 and 10Base2: poor
cabling plant





What is the problem?
Need: a cabling plant like telephone system
1983: AT&T and NCR, running Ethernet on
UTP
Speed: 1M bps (This is the major issue)
Importance: a market failure but a technology
milestone

We have a structured cabling plant now
31
StarLAN
Ethernet hub or repeater
• Network topology: now a star topology
• A structured cabling plant similar to telephone network.
Q: Why is star topology better than bus topology?
32
Figure 13.12 10Base-T implementation
13.33
Ethernet Encoding Schemes

Manchester (10BaseT)


0-bit = + voltage, - voltage
1-bit = - voltage, + voltage
Figure 13.13 10Base-F implementation
13.35
NRZ-I Encoding (100Base-FX)
0: no change 1: inverse
CSMA/CD Basic Algorithm



13.37
Medium idle? Transmit. Not ready? Wait
until it becomes ready then wait
interframe gap (9.6 microseconds) or use
p-persistent algorithm
Collision detected? Continue transmission
until min packet time reached (jam signal)
Increment retransmission counter
CSMA/CD Procedure
Station is
ready to send
Sense
Channel
try again
channel
busy
Wait
(backoff
strategy)
Send
jam signal
Transmit data and
sense channel
collision detected
successful
transmission
38
802.3 Parameters (Table 3.2)
39
Slot time
512 bit times (51.2 s)
InterFrameGap
96 bit times (9.6 s)
AttemptLimit
16 (tries)
BackoffLimit
10 (exponent)
JamSize
32 bits
MaxFrameSize
1518 bytes
MinFrameSize
64 bytes
AddressSize
48 bits (6 bytes)
Ethernet Performance
Why do large frames have better performance?
40
CSMA/CD Basic Algorithm





13.41
Collisions are the killer!!
Network throughput typically 20-30%,
40% was max!
Need a way to eliminate collisions
What about token ring and token bus?
What about bridges and switches?
CSMA/CD Collisions





13.42
How long does it take to hear a collision?
Signals travel at 2/3 speed of light thru
copper
Propagation time = distance / velocity
How many bits can you transmit during
that time?
Data transfer time = length of frame (bits)
/ data rate (bps)
Table 13.1 Summary of Standard Ethernet implementations
13.43
13-3 CHANGES IN THE STANDARD
The 10-Mbps Standard Ethernet has gone through
several changes before moving to the higher data
rates. These changes actually opened the road to the
evolution of the Ethernet to become compatible with
other high-data-rate LANs.
Topics discussed in this section:
Bridged Ethernet
Switched Ethernet
Full-Duplex Ethernet
13.44
Figure 13.14 Sharing bandwidth
13.45
Figure 13.15 A network with and without a bridge
13.46
Figure 13.16 Collision domains in an unbridged network and a bridged network
13.47
Figure 13.17 Switched Ethernet
13.48
How Switches Learn Host Addresses
and Locations
MAC forwarding table
A
0260.8c01.1111
C
0260.8c01.2222
B
E0
E1
E2
E3
0260.8c01.3333
D
0260.8c01.4444
• Initial MAC forwarding table is empty
How Switches Learn Host Addresses
and Locations
MAC forwarding table
E0: 0260.8c01.1111
A
0260.8c01.1111
C
B
E0
E1
E2
E3
0260.8c01.2222
0260.8c01.3333
D
0260.8c01.4444
• Station A sends a frame to Station C
• Switch caches station A MAC address to port E0 by learning the
source address of data frames
• The frame from station A to station C is flooded out to all ports
except port E0 (unknown unicasts are flooded)
How Switches Filter Frames
MAC forwarding table
A
0260.8c01.1111
C
0260.8c01.2222


E0:
E2:
E1:
E3:
0260.8c01.1111
0260.8c01.2222
0260.8c01.3333
0260.8c01.4444
E0
E1
E2
E3
B
0260.8c01.3333
D
0260.8c01.4444
Station A sends a frame to station C
Destination is known, frame is not flooded
Broadcast Frames
A
0260.8c01.1111
C
E0:
E2:
E1:
E3:
0260.8c01.1111
0260.8c01.2222
0260.8c01.3333
0260.8c01.4444
E0
E1
E2
E3
0260.8c01.2222
B
0260.8c01.3333
D
0260.8c01.4444
• Station D sends a broadcast frame
• Broadcast frames are flooded to all ports other than the
originating port
MAC Forwarding Table



Unlike IP address, MAC address is local.
MAC forwarding table is also local, within the
single broadcast domain (which is an IP subnet).
MAC address learning stops at the router, and a
switch does not learn the MAC addresses at the
other side of the router.

The switch learns one MAC address of the router.
Figure 13.18 Full-duplex switched Ethernet
13.54
5-4-3 Rule (10BaseX)
5 segments, 4 repeaters, 3 populated
segments
hub/repeater
55
13-4 FAST ETHERNET
Fast Ethernet was designed to compete with LAN
protocols such as FDDI or Fiber Channel. IEEE
created Fast Ethernet under the name 802.3u. Fast
Ethernet is backward-compatible with Standard
Ethernet, but it can transmit data 10 times faster at a
rate of 100 Mbps.
Topics discussed in this section:
MAC Sublayer
Physical Layer
13.56
Figure 13.19 Fast Ethernet topology
13.57
Figure 13.20 Fast Ethernet implementations
13.58
Figure 13.21 Encoding for Fast Ethernet implementation
13.59
MLT-3 Signal (100Base-TX)
three levels: +1, 0, -1
transition (at 1’s only): +1 => 0 = -1 => 0 => +1
Table 13.2 Summary of Fast Ethernet implementations
13.61
13-5 GIGABIT ETHERNET
The need for an even higher data rate resulted in the
design of the Gigabit Ethernet protocol (1000 Mbps).
The IEEE committee calls the standard 802.3z.
Topics discussed in this section:
MAC Sublayer
Physical Layer
Ten-Gigabit Ethernet
13.62
Note
In the full-duplex mode of Gigabit
Ethernet, there is no collision;
the maximum length of the cable is
determined by the signal attenuation
in the cable.
13.63
Figure 13.22 Topologies of Gigabit Ethernet
13.64
Figure 13.23 Gigabit Ethernet implementations
13.65
Figure 13.24 Encoding in Gigabit Ethernet implementations
13.66
Table 13.3 Summary of Gigabit Ethernet implementations
13.67
1G to 10G (802.3ae)
10GbE




Full-duplex only (no CSMA/CD)
Fiber only (802.3ae)
WAN (SONET-friendly) PHY
Mapping to OC-192 carrier



New line coding (64b/66b)
Standard for copper 10GBase-CX (802.3ak) - 2004



Rate adaptation to SONET payload capacity
dual coax cable
<20m (for data center use only)
Standard for copper 10GBase-T (802.3an) - 2006

UTP cable (~50m for Cat-6 and ~100m for Cat-6a)
IEEE P802.3ba
100G physical layer specifications for operation up to:

40 km on single-mode fiber (SMF) using wavelength division
multiplexing (WDM)

10 km on SMF using WDM

100 m on parallel OM3 multimode fiber (MMF)

10 m over a copper cable assembly
40G physical layer specifications for operation up to:

10 km on SMF using WDM

100 m on parallel OM3 MMF

10 m over a copper cable assembly
70
Other High Speed LAN
Technologies




Token Ring
Fiber Distributed Data Interface
(FDDI)
100VG-AnyLAN
ATM LAN
71
Continuous Improvement






10M => 100M (802.3u) => 1G (802.3z and
802.3ab)
10G (802.3ae, 802.3ak, 802.3an)
Now 40 G and 100G (approved in 2010)
Virtual LAN (802.1Q)
Quality of Service (802.1p)
Fault tolerance





Spanning Tree Algorithm and Protocol 802.1D
Rapid STP 802.1w and per VLAN STP 802.1s
Link aggregation 802.1ad
Port based network access control: 802.1X
Wireless LAN: 802.11
72
Power Over Ethernet (PoE)
Power over Ethernet (PoE)
PoE delivers low-voltage power over Cat 5 cable.
Ideal for applications such as wireless access points,
access-card readers, IP telephones, IP video cameras,
and may even be useful with cell phones, PDAs, and laptops.
Once a proprietary technology, now it is an IEEE standard
802.3af.
For small scale systems, installing PoE may be as simple
as installing a new line card in a switch or hub.
13.73
Power Over Ethernet (PoE)
Power over Ethernet (PoE)
Installing PoE in a large network may create physical
power infrastructure problems.
For example, each powered device (PD) can draw up to
15 watts from the switch/hub (power sourcing equipment,
or PSE). If you have a 100-port hub, that’s 1500 watts
to support the PDs, plus another 1000 watts normally
needed to drive the switch. The total is 2500 watts, or
roughly 25 amps. Does your wiring closet have a 25 amp
circuit? What if the switch has 300 ports?
PoE can deliver this 48 volt 15 watt power over the
unused spare pair of wires in an Ethernet connection,
or over the transmit/receive pairs.
13.74