Download Layer 2 Fundamentals

Prof. Dr. R.Nitsch, FH Darmstadt
Module 4 – Layer 2 Fundamentals
Reiner Nitsch
FB Informatik
Office:
F16/15c
Phone:
9255
Email:
[email protected]
Layer 2
1
Layer 1 Limitations - Layer 2 Tasks
Prof. Dr. R.Nitsch, FH Darmstadt
• Layer 1 involves
– media, signals, bit streams
that travel on media,
– components that put signals
on media, and
– various topologies.
• Layer 2 tasks and services are
– Layer-2 protocols transmit Layer-3 PDUs over a single link between two network
nodes
– Communicate with the upper-level layer using Logical Link Control (LLC) Sublayer
– Organize or group the bits using a Layer 2 frame format (framing)
– Decide which computer will transmit binary data from a group that are all trying to
transmit at the same time using a system called Media Access Control (MAC).
– Reliable Data Transmission (Optional)
– Flow Control (Optional)
– Error detection
– Error recovery (optional)
– Identify computers using a Layer 2 addressing (or naming) process.
• All Layer-2 protocol tasks are implemented within the Network Interface Circuit
(NIC)
Layer 2
2
Error Detection and Error Correction Techniques
Prof. Dr. R.Nitsch, FH Darmstadt
• Layer-2 PDUs include Error Discovery and Correction (EDC) bits in order to
enable error detection at the receiving node
• Even with good EDC techniques bit errors may remain unrecognized. The
propability of such an event should be as small as possible.
• Techniques in common use are parity checks, cyclic redundancy check (CRC)
and forward error correction (FEC)
Paritätsprüfungen:
0010111100111010 1
Prüfung mit gerader 1-Bit-Parität
1-dimensionale gerade Parität
0010111100111010
0010100101001110
1101011100010100
1000100010010010
0101100111110010
1
1
0
1
1
2-dimensionale gerade Parität
0010111100111010
0010100001001110
1101011100010100
1000100010010010
0101100111110010
1
1
0
1
1
Korrigierbare Bitfehler: jeder Einzelbitfehler
Erkennbare Bitfehler: jede Kombination von 2 Fehlern
Layer 2
3
Error Correction AND Detection Scenario
Prof. Dr. R.Nitsch, FH Darmstadt
Datagram
Datagram
Y
all
Recompute EDC
from D' and compare bits in D'
ok?
with EDC'
D
EDC
D+Errors = D'
N
Discard
frame
EDC'
Bit error-prone link
Layer 2
4
Internet Checksum
Prof. Dr. R.Nitsch, FH Darmstadt
• The Internet Checksum is used by the protocols IP, TCP und UDP (RFC 1071)
0010010010101101 0100100101010110 1011010001011110 1010110111100010
0010010010101101
0100100101010110
0010010010101101
transmit
0100100101010110
1011010001011110
1011010001011110
1010110111100010
1010110111100010
even parity for each column
0111010001000111
1000101110111000
ones complement
Einerkomplement
( 0 and 1 swapped)
1000101110111000
1111111111111111
16-BitInternet-checksum
Transmitter
Layer 2
No errors!
Every 0 would
indicate an error
even
parity
each
column
Receiver
5
Cyclic Redundancy Check
Prof. Dr. R.Nitsch, FH Darmstadt
• Is based on powerful mathematics (endliche Felder, finite fields)
• Ethernet (IEEE 802.3) and Token Ring (IEEE 802.5) use the so called
CRC-32 Method, which produces a CRC value of 32 Bit length
• This Method allows to detect
– All single-bit errors
– All bit-pair errors
– All frames with an odd error count
• It's not possible to correct errors with the CRC information!
Message + "000"
T
T
T
XOR
At the end of processing the registers contain the error check bits.
Layer 2
6
Forward Error Correction
Prof. Dr. R.Nitsch, FH Darmstadt
• The ability of a receiver to detect and correct errors is known as forward
error correction (FEC)
• FEC is commonly used in audio and video storage and playback devices (Audio
CD, DVD), DSL, …
• FEC reduces the number of sender retransmissions required.
• This is important for real-time network application because they have not to
wait for the round-trip propagation delay associated with each
retransmission.
• FEC techniques demand for powerful mathematics.
Layer 2
7
Layer-2 (Data Link) Sublayers
Prof. Dr. R.Nitsch, FH Darmstadt




The IEEE divides the OSI data link
layer into two separate sublayers.
 Media Access Control (MAC)
(transitions down to media)
 Logical Link Control (LLC)
(transitions up to the network
layer)
The MAC sublayer is concerned with
the physical components that will be
used to communicate the information.
The LLC sublayer remains relatively
independent of the physical equipment
The LLC, as a sublayer, participates in
the encapsulation process.
Layer 2
8
MAC – Media Access Control Sublayer
Prof. Dr. R.Nitsch, FH Darmstadt
The Media Access Control (MAC) sublayer

deals with the protocols that a host
follows in order to access the physical
media.

is responsible for the actual framing
IP
APPLEIPX

builds the 1s and 0s to hand off to
Layer 3
TALK
the physical layer.
LLC

is responsible for media access: (later) Layer 2 - LLC

Random Access
Ethernet
Token
FDDI
MAC &Layer 1

Token Passing
Ring

Polling (Master sequentially asks all
slaves wether they have to transmit
data or not)
• is concerned with physical naming (addressing); defines MAC addressing;
• The NIC uses the MAC address to assess whether the message is
destined for that host and therefore should be passed onto the upper
layers of the OSI model.
• The NIC makes this assessment without using CPU processing time.
Layer 2
9
Media Access Control
• Two types of links exist in networks
– point-to-point links consist of a single
sender and a single receiver at its ends.
– broadcast links with multiple sender and
receivers connected to the same shared
transmission medium
• A well-known example is a class room with a
teacher, several pupils and the air as shared
transmission media.
• Communication in a class room follows these
rules:
– Give everyone a chance to speak
– Don't speak until you are spoken to.
– Don't monopolize the conversation
– Raise your hand if you have a question
– Don't interrupt when someone is speaking
– Don't fall asleep when someone else is
talking
Prof. Dr. R.Nitsch, FH Darmstadt
• Computer networks similarly have protocols, so called multiple access protocols
Layer 2
10
Multiple Access Protocols
Prof. Dr. R.Nitsch, FH Darmstadt
• For a broadcast channel of rate R bits/s we have the following desired
characteristics:
– Throughput R bps when only one node has data to send
– Average throughput R/M bps when M nodes have data to send
– Decentralized protocol; no master nodes as single point of failures
– Simple protocol that is inexpensive to implement
• Many different multiple access protocols have been implemented in various
link layer technologies. They all belong to one of the following categories:
– channel partitioning protocols
– random access protocols
– taking-turns protocols
Layer 2
11
Channel Partitioning Protocols
K2
K3
time slot
K1 K2 K3 K1 K2 K3
time
link (R bps)
TDM-DEMUX
K1
K2
K3
available frequency band
K1
K2
K3
K1
f1
K2
K3
f2 frequency
link
FDM-DEMUX
Layer 2
K1
FDM-MUX
• pros&cons:
– No collisions possible.
– each node has all the time a fixed
 R/N bps bandwidth available even
 when only one node has data to send.
time frame
TDM-MUX
• Time Division Multiple Access – TDMA
– divides time into time frames and further
divides each time frame into N time slots
– assigns each time slot to one of N nodes
– Examples: Data-Highways (SDH, PDH,…)
• Frequency Division Multiple Access – FDMA
– divides available frequency band into N
frequency slots
– Assigns each frequency to one of N nodes
– Examples: Radio, TV,...
in former times
Prof. Dr. R.Nitsch, FH Darmstadt
K1
K2
K3
12
Random Access Protocols
• Nodes transmit at the full channel bitrate
R bps
• When 2 nodes transmit at the same time,
a collision occurs
• Nodes repeatedly retransmit their
frames until they get transmitted without
collision
• Nodes wait a individually selected random
time before they start the retransmission
in order to avoid a new collision
• Dozens of random access protocols are
described in literature
(see: R.Rom,M.Sidi: Multiple Access
Protocols, Performance and Analysis)
• Most commonly in use are
– ALOHA protocols
– Carrier Sense Multiple Access
(CSMA) protocols
Layer 2
Prof. Dr. R.Nitsch, FH Darmstadt
15
Slotted ALOHA
• Assumptions
– All frames consists of exactly L bits
– Time is divided in slots of size L/R seconds
(=time to transmit a frame)
– Nodes start frame transmissions only at the
beginnings of the slots
– Nodes can detect collisions before the slot
ends
• Operation of slotted ALOHA
– Nodes, with frames to send, wait until the
beginning of the next slot and transmit the
entire frame (Nodes must be synchronized
to slot times)
– Without a collision the frame is successfully
transmitted
– If a collision is detected the node
retransmits the frame in each subsequent
slot with a probability of p until the frame is
transmitted successfully. The node
effectively tosses a coin to determine
wether retransmission will start in next slot.
Layer 2
Prof. Dr. R.Nitsch, FH Darmstadt
empty slots
successfull slots
collisions
Advantages
• full rate for active node
• highly decentralized
• extreme simple
16
Efficiency of slotted ALOHA
Definition:
• Efficiency is defined as the long-run
fraction of successfull slots in case of
large number of nodes (N)
Assumptions:
– Each node attemps to transmit a
fresh frame in each slot with
probability p (nodes allways have
frames to send)
– Collided frames are retransmitted in
the next slot with same probability p
Calculation of Efficiency
• Probability of a successfull node
– p(1-p)N-1
• Probability of successfull slot fraction
– Np(1-p)N-1 = Efficiency E
Prof. Dr. R.Nitsch, FH Darmstadt
0.4
1/e= 37%
0.4
0.3
E( p  250 )
E( p  500 )
0.2
E( p  750 )
E( p  1000 )
0.1
0
0
0
0
0.05
0.1
0.15
p 100
0.2
0.25
0.3
0.3
Disadvantages:
• Only 37 % of the slots do useful work
• Max. effective channel transmission rate
is 0,37·R bps
• A similar analysis shows:
•37% slot times go empty
•26% slot times have collisions
Layer 2
17
Carrier Sense Multiple Access
Reason for the bad efficiency of ALOHA
• In slotted ALOHA, decision to send is made
independent of the other nodes activities
– nodes send without regard to other
nodes sending activities
– node don't stop transmission when a
collision occurs
• Polite humans obey the following rules in
conversation
– Listen before speaking: In networking
world this is called "carrier sense"
– If someone else begins talking at the
same time, stop talking: In networking
world this is called "collision detection"
• These 2 rules are embedded in the family
of CSMA (carrier sense multiple access)
and CSMA/CD (CSMA with collision
detection)
• Why do collisions occur despite the carrier
sensing feature? ->see space-time diagram
Layer 2
Prof. Dr. R.Nitsch, FH Darmstadt
CSMA (without CD) space-time diagram:
4 nodes in space (A,B,C,D)
t0: node B senses idle channel and starts
sending; signal propagates in both
directions along the broadcast medium
t1: node C senses idle channel at time t1
and starts sending too. Signal from node
B has not yet reached node D because
limited light velocity
18
CSMA with Collision Detection
Prof. Dr. R.Nitsch, FH Darmstadt
• CSMA protocol family didn't perform collision
detection. They continue to transmit their
frames even in case of a collision.
• Collision detection and aborting the
transmission in case of a detected collision
will rise protocol performance
During this time bandwith is
wasted by CSMA protocols
Layer 2
19
Taking-Turns Protocols
• Review of desireable protocol
characteristics
– Throughput R bps when only one
node has data to send
– Average throughput R/M bps when
M nodes have data to send
• ALOHA has the first characteristic but
not the second! Taking-turns protocols
have both!
• Important taking-turns protocols are
– Polling Protocol
– Token-Passing Protocol
Prof. Dr. R.Nitsch, FH Darmstadt
Polling Protocol
• requires a master node
• master node polls each node in round
robin fashion
• Each polled node is authorized to send a
maximum number of frames.
• End of transmission is indicated by lack
of signal!
• Advantages:
– No collisions, no empty slots
 better efficiency
• Disadvantages
– polling delay (time needed to poll a
node). If only one node is active N-1
polling delays are wasted.
– Master node is single point of failure
Layer 2
20
Token-Passing Protocol
Prof. Dr. R.Nitsch, FH Darmstadt
• There is no master node
• A small, speacial-purpose frame known as a token is exchanged among the nodes in
some fixed order (Ex: Node 1  Node 2  …  Node N  Node 1; logical ring
topology)
• The node holding the token is authorized to send a maximum number of frames.
• After end of transmission or if nothing is to send, the token is immediately passed to
the next node.
Advantages
• Token passing is decentralized and has high efficiency
Problems
• Failure of one node can crash the entire channel
• If a faulty node doesn't release the token a recovery procedure has to get the token
back into circulation
• Over the years many token-passing products have been developed (Token Ring (IEEE
802.5), FDDI (Fiber Distributed Data Interface) All of these have become relatively
minor players in competition with Ethernet
Layer 2
21
Naming Computers with MAC Addresses
Prof. Dr. R.Nitsch, FH Darmstadt
• MAC addresses provide a way for computers (nodes) to identify themselves within LANs
(Ethernet, Token Ring, FDDI)
• MAC addresses are:
48-Bit-MAC-Address
– 48 bits in length
– Expressed as twelve hexadecimal digits.
– The first six hexadecimal digits, which are
administered by the IEEE "universally
administered address", identify the
manufacturer or vendor and thus comprise the
Organizational Unique Identifier (OUI).
– The remaining six hexadecimal digits comprise
the interface serial number, or another value
administered by the specific vendor.
• MAC addresses are sometimes referred to as
burned-in addresses (BIAs) because they are
burned into read-only memory (ROM)
• The PC software (in PROTOCOL.INI or NET.CFG)
can be configured to substitute a different address
number. When this option is used, it is called a
"locally administered address."
•IEEE OUI FAQs: http://standards.ieee.org/faqs/OUI.html
Layer 2
22
Introduction to Ethernet
• Ethernet developed in the 1970s
• success of Ethernet is due to the following factors:
– Simplicity and ease of maintenance
– Ability to incorporate new technologies
– Reliability
– Low cost of installation and upgrade
Prof. Dr. R.Nitsch, FH Darmstadt
A drawing of the first Ethernet system
by Bob Metcalfe
History
• Early 1970s: At the University of Hawaii a system called Alohanet was developed to
control access of various stations. This work formed the basis for the Ethernet access
method known as CSMA/CD.
• 1980: First Ethernet standard; published by Digital Equipment Company, Intel, and
Xerox (DIX); open standard; up to 10 Mbps; Thicknet; <= 2000 m
• 1985: 802.3 Ethernet standard is published by IEEE; complies to ISO/OSI model;
only small modifications to original (DIX)-Ethernet.
• Any Ethernet network interface card (NIC) can transmit and receive both Ethernet
and 802.3 frames.
• 1995: IEEE announced a standard for a 100-Mbps Ethernet (100BASE-T, 100m).
• 1998 and 1999: IEEE standards for Gigabit Ethernet (1000BASE-T, 100m).
• All the standards are essentially compatible with the original Ethernet standard.
• Many Ethernet standard supplements were added in order to use different
transmission media and higher transmission rates.
Layer 2
23
Review: Encapsulation Example
Prof. Dr. R.Nitsch, FH Darmstadt
data
APDU
(Data)
Transport
Header
data
TPDU
(Segment)
Network Transport
Header
Header
data
NPDU
(Packet)
Routers
Network Transport
Header
Header
data
LPDU
(Frame)
Switches, Bridges
Email
Link
Header
Link
Trailer
Bits
1010010100010101111101100010110110001
Hosts
Repeaters, Hubs,
Cables, etc.
Let us focus on the Layer 2, Data Link, Ethernet Frame for now.
Layer 2
24
Generic Data Link Frame
Prof. Dr. R.Nitsch, FH Darmstadt
•
•
•
•
•
•
•
Framing is the Layer 2 encapsulation process.
A frame is the Layer 2 protocol data unit (2-PDU).
Framing provides order, or structure, to the bitstream.
There are many different types of frames described by various standards.
A single generic frame has sections called fields
Eeach field is composed of bytes.
The names of the fields are as follows:
– Start frame field
– Address field
– Length / type field
– Data field
– Frame check sequence field (FCS)
• Logical link control (LLC) bytes are also included with the data field in the IEEE
standard frames.
• There are three common conventions for the format of the remainder of the frame:
– Ethernet II or DIX
– IEEE 802.3 and 802.2
– SNAP
Layer 2
25
IEEE 802.3 and 802.2
64 bis 1518
Bytes
Prof. Dr. R.Nitsch, FH Darmstadt
• The DIX standard did not need a
length field because the vendor
protocols that used it (XNS,
DECNET, IPX, IP) all had their own
length fields. However, the 802
committee needed a standard that
did not depend on the good behavior
of other programs. The 802.3
standard therefore replaced the
two byte type field with a two byte
length field.
FCS calculation
Byte
MAC
MAC
• Maximum size of Ethernet frame payload
Interpretation of Length/Type Field:
(data) is 1500 and Xerox did not assign type
•Value >= 0x600 (hex): Type field
(Ethernet II or DIX, Ethernet
values below 1500. This allows DIX and 802
802.3)
standards to overlap conflictlessly.
•Value < 0x600 (hex): Length
• The 802.2 header follows the 802.3 header
field (Ethernet 802.3)
(and also follows the comparable fields in a
Token Ring, FDDI, or other types of LAN).
Layer 2
27
Ethernets Type of Service
Prof. Dr. R.Nitsch, FH Darmstadt
• All of the Ethernet technologies provide connectionless service to the
network layer.
– No handshaking is done with receiving node
• All of the Ethernet technologies provide unrealiable service to the network
layer.
– Received frames are not acknowledged
– A frame that fails the CRC check is discarded without informing the
sender
• The lack of reliable service makes Ethernet simple and cheap
• Does the application at the receiving node see the gaps in the data stream?
– No, if the layer-3 or layer-4 services detect and recover from the gaps.
Example: If TCP is the layer-4 protocol in use then the gaps are detected
and retransmission is enabled because TCP offers connection-oriented
and reliable transport service to it's upper layers.
– Yes, if they do not. If IP is the layer-3 protocol and UDP is the layer-4
protocol the gaps will remain undetected because both protocols offer
only unreliable transport service to their upper layers .
Layer 2
32
Sending and receiving Ethernet frames on a bus
Prof. Dr. R.Nitsch, FH Darmstadt
• When an Ethernet frame is sent
out on the “bus” all devices on the
bus receive it.
What do they do with it?
• Each NIC card compares its own
MAC address with the Destination
MAC Address.
• If it matches, it copies in the rest
of the frame.
• If it does NOT match, it ignores
(filters) the rest of the frame …
– … unless you are running a
Sniffer program
• So, what happens when multiple
computers try to transmit at the
same time?
• They produce a collision. Collisions
are the most common error
condition on an Ethernet.
1111
2222
3333
Abbreviated
MAC
Addresses
nnnn
3333 1111
1111
2222
3333
nnnn
X
Collision!
Layer 2
33
CSMA/CD (Carrier Sense Multiple Access with Collision Detection)
Prof. Dr. R.Nitsch, FH Darmstadt
•Ethernet is a shared-media broadcast
technology.
•The access method CSMA/CD used in
Ethernet performs three functions:
– Transmitting and receiving data
packets.
– Decoding data packets and checking
them for valid addresses before
passing them to the upper layers of
the OSI model.
– Detecting errors within data packets
or on the network.
•Collision detection method: Monitor the
amplitude of the signal during
transmission. If the amplitude increases
on the networking media a collision has
occurred.
•In case of collision the nodes stop transmitting for a random period of time (backoff
time), which is (hopefully) different for each device.
•When a device's backoff delay period expires, that device can attempt to gain access
to the networking media.
Layer 2
34
CSMA/CD and Collisions
Prof. Dr. R.Nitsch, FH Darmstadt
CSMA/CD (Carrier Sense Multiple Access with Collision Detection)
• Listens to the network’s shared media to see if any other users is “on the
line” by trying to sense a neutral electrical signal or carrier.
• If no transmission is sensed, then multiple access allows anyone onto the
media without any further permission required.
• If two NICs detect a neutral signal and access the shared media at the
exact same time, a collision occurs and is detected.
• The PCs sense the collision by wether excess voltage is on the line
additional to their own transmission voltage.
• When a collision occurs, a 32-bit jamming signal is sent out by the first
NIC that detects the collision in order to enforce the collision .
• A random back-off scheme, is used to prevent colliding retransmissions.
• If collisions continue to occur, the NICs random interval is doubled,
lessening the chances of a collision.
• Late Collisions: In a proper functioning Ethernet network, a NIC may
experience collision within the first slot time (minimum frame period, 51.2
µS) after it starts transmission. This is the reason why an Ethernet NIC
monitors the CD signal during this time and use CSMA/CD. A faulty CD
circuit, or misbehaving NIC or transceiver may lead to a late collision (i.e.
after one slot time).
Layer 2
35
Interframe spacing and backoff
Prof. Dr. R.Nitsch, FH Darmstadt
Interframe spacing
• After a frame has been sent, all stations on a 10-Mbps Ethernet are required
to wait a minimum of 96 bit-times (9.6 microseconds) before any station may
legally transmit the next frame.
• This minimum spacing between two non-colliding frames is also called the
interframe gap. It is measured from the last bit of the FCS field of the
first frame to the first bit of the preamble of the second frame.
• This gap is intended to allow slow stations time to process the previous
frame and prepare for the next frame.
Backoff Time
• After a collision occurs and all stations wait the full interframe spacing
• The stations that collided must wait an additional backoff time before
retransmitting the collided frames.
• The waiting period is intentionally random so that two stations do not delay
for the same amount of time before retransmitting which would result in
more collisions.
• The waiting period is measured in increments of the parameter "slot time".
• If the MAC layer is unable to send the frame after sixteen attempts, it
gives up and generates an error to the network layer.
Layer 2
36
Exponential Back-off Algorithm
Prof. Dr. R.Nitsch, FH Darmstadt
• If all NICs attempted to retransmit immediately
following a collision, then this would certainly
result in another collision.
• Ethernet uses a random back-off period to
minimize the propability of this event.
• Each node selects a random number, multiplies
this by the slot time (minimum frame period,
51.2 µS) and waits for this random period before
attempting retransmission. The small Inter-Frame
Gap (IFG) (e.g., 9.6 microseconds) is also added.
• On a busy network, a retransmission may still collide with another retransmission (or
possibly new frames being sent for the first time by another NIC). The protocol
therefore counts the number of retransmission attempts (using a variable N in the
above figure) and attempts to retransmit the same frame up to 15 times.
• For each retransmission, the transmitter constructs a set of numbers:
{0, 1, 2, 3, 4, 5, ... L} where L is (2K-1) and where K=N; K<= 10;
• A random value R is picked from this set, and the transmitter waits for a period
R x (slot time) i.e. R x 51.2 μs
• For example, after two collisions, N=2, therefore K=2, and the set is {0, 1, 2, 3} giving a
one in four chance of collision. This corresponds to a wait selected from {0, 51.2, 102.4,
153.6} micro seconds.
Layer 2
37
What are collision domains?
• Host A senses no voltage on the circuit.
It begins transmitting a frame. The
signal voltage travels through the wire
with about light velocity. Repeaters and
hubs flood the incoming signals out any
other outputs
• Host B also wants to transmit a frame.
It senses a free line up to the time
when host A's first frame byte arrives
at host B. Immediately after Host B
starts sending it detects the excessive
voltage from host A's frame. It stops
sending and starts sending a 32-bit jam
signal
• Host A cannot detect the collision until
the collision fragments of Host B
arrives. It MUST continue sending his
frame until that time.
• When the collision fragments finally
reaches Host A, it also truncates the
current transmission and substitutes
the jam signal.
• A hub or series of hubs/repeaters is a
single collision domain.
Prof. Dr. R.Nitsch, FH Darmstadt
1 segment
1
32
24
16
8
1
1
42
32
16
32-bit jam signal
8
1
64 byte  200 m  Diameter of collision domain
64
32-bit jam signal
Layer 2
56
48
40
38
Limitations of half/duplex operation
Prof. Dr. R.Nitsch, FH Darmstadt
• In order to savely detect collisions at
the sender, it must keep on sending the
frame for a minimum time (slot time).
This requires the frames to have a
minimum length.
• This minimum time is equal to the
maximum round trip delay of the
network, which itself is a measure of
the distance between the most distant
nodes in the network
• Consequence: For a given minimum
frame length, the extent of a network
scales inversely with data rate.
10000 m
• Slot-time is calculated assuming
– maximum cable lengths on the
1000 m
largest legal network architecture,
– all hardware propagation delay
times are at the legal maximum and 100 m
– the 32-bit jam signal is used when
collisions are detected.
10 m
Layer 2
≈2800 m
≈205 m
≈20 m
10
100
1000 Mbps
39
Slot Times required
Prof. Dr. R.Nitsch, FH Darmstadt
• At 10 Mbps the standard demands a slot-time of 51,2us (64 Bytes, 512 bits).
This allows an ethernet LAN to extent up to 2-3 km.
• Fast Ethernet (100 Mbps) has a slot time of 5.12μs with the same minimum
frame length (512 bit). The network extent it limited to 200 m with twistedpair cable.
• To allow 1000-Mbps Ethernet to operate in half duplex the maximum network
extent needs to be restricted to about 20 m (not acceptable or the frames
needs to be extended. Therefore the extension field was added to expand the
slot-time to 4.096μs (512 bytes, 4096 bits) for small data payloads.
6
DA
6
2
SA Length/Type
46-1500
4
<512
Data
FCS
Extension
Byte
• This field is present only on 1000-Mbps, half-duplex links and allows minimumsized frames to be long enough to meet slot-time requirements. Extension bits
are discarded by the receiving station.
• Propagation velocity is about 200 m per microsecond in a UTP cable.
• It takes just 10 bit-times for a 10BASE-T, 100 bit-times for 100BaseT, and
1000 bit-times for 1000BaseT to travel this length (200 m) of UTP cable.
• Therefore half duplex is not permitted in 10-Gigabit Ethernet.
Layer 2
40
Sending and receiving Ethernet frames via a hub
Prof. Dr. R.Nitsch, FH Darmstadt
3333
1111
2222
Nope
Wasted
bandwidth
3333
For me!
4444
Nope
5555
Nope
1111
• So, what does a hub do when it
receives a frame?
• Remember, a hub is nothing more
than a multiport repeater acting as a
layer 1 device.
• The hub will flood it out all ports
except for the incoming port.
• A hub does NOT look at layer 2
addresses, so it is fast in
transmitting data.
• Disadvantage with hubs: A hub or
series of hubs is a single collision
domain.
• A collision will occur if any two or
more devices transmit at the same
time within the collision domain.
• More on this later.
• Another disadvantage with hubs is
that it takes up unnecessary
bandwidth on other links.
Nope: no operation
Layer 2
42
Sending and receiving Ethernet frames via a switch
Prof. Dr. R.Nitsch, FH Darmstadt
Source Address Table
Port Source MAC Add. Port Source MAC Add.
3333 1111
• Switches are also known as
learning bridges or learning
switches.
• A switch has a source address
table in cache (RAM) where it
stores source MAC address
after it learns about them.
• A switch that receives an
Ethernet frame searches the
source address table for the
Destination MAC address (DA).
– If it finds a match, it filters
the frame by only sending it
out that port.
– If there is not a match if
floods it out all ports.
switch
1111
3333
Abbreviated
MAC
addresses
2222
4444
Layer 2
43
No Destination Address in table, Flood
Prof. Dr. R.Nitsch, FH Darmstadt
Source Address Table
Port Source MAC Add. Port Source MAC Add.
1
1111
3333 1111
• Next, in our scenario, the switch will
flood the frame out all other ports,
because the DA is not in the source
address table.
switch
1111
• How does it learn Source MAC
Addresses (SA)?
– First, the switch will see if the SA
(1111) is in it’s table.
• If it is, it resets the associated
timer (more in a moment).
• If it is NOT in the table it adds
it with the port number.
3333
Abbreviated
MAC
addresses
2222
4444
Layer 2
44
Destination Address in table, Filter
Prof. Dr. R.Nitsch, FH Darmstadt
Source Address Table
Port Source MAC Add. Port Source MAC Add.
1
1111
6
3333
1111 3333
• Now 3333 sends data back to 1111.
• The switch sees if it has the SA
stored.
– It does NOT so it adds it. (This
will help next time 1111 sends to
3333.)
• Next, it checks the DA and in our case
it can filter the frame, by sending it
only out port 1.
switch
1111
3333
Abbreviated
MAC
addresses
2222
4444
Layer 2
45
Destination Address in table, Filter
Prof. Dr. R.Nitsch, FH Darmstadt
Source Address Table
Port Source MAC Add. Port Source MAC Add.
1
1111
6
3333
3333 1111
switch
1111 3333
• Now, because both MAC addresses are
in the switch’s table, any information
exchanged between 1111 and 3333 can
be sent (filtered) out the appropriate
port.
1111
• What happens when two devices send
to the same destination?
• What if this was a hub, not a
switch?
• Where is (are) the collision domain(s)
in this example?
3333
Abbreviated
MAC
addresses
2222
4444
Layer 2
46
No Collisions in Switch, Buffering
Prof. Dr. R.Nitsch, FH Darmstadt
Source Address Table
Port Source MAC Add. Port Source MAC Add.
1
1111
6
3333
9
4444
3333 1111
switch
3333 4444
1111
• Unlike a hub, a collision does NOT
occur within these segments, which
would cause the two PCs to have to
retransmit the frames.
• Instead the switch buffers the frames
and sends them out port #6 one at a
time.
• The sending PCs have no idea that
their was another PC wanting to send
to the same destination.
3333
Abbreviated
MAC
addresses
2222
4444
Layer 2
47
Collision Domains
Prof. Dr. R.Nitsch, FH Darmstadt
Source Address Table
Port Source MAC Add. Port Source MAC Add.
1
1111
6
3333
9
4444
3333 1111
Collision Domains
switch
3333 4444
1111
3333
Abbreviated
MAC
addresses
2222
4444
Layer 2
• When there is only one device on a
switch port, the collision domain is only
between the PC and the switch. (Cisco
curriculum is inaccurate on this point.)
• Having these small collision domains
with a switch is called "micro
segmentation"
• With a full-duplex PC and switch port,
there will be no collision, since the
devices send and receive at different
circuits at the same time.
• Full-duplex operation allows for larger
network architecture designs since the
timing restriction for collision
detection is removed.
48
Other Information
Prof. Dr. R.Nitsch, FH Darmstadt
Source Address Table
Port Source MAC Add. Port Source MAC Add.
1
1111
6
3333
9
4444
switch
1111
3333
Abbreviated
MAC
addresses
2222
• How long are addresses kept in the
Source Address Table?
– 5 minutes is common on most vendor
switches.
• How do computers know the
Destination MAC address?
• ARP Caches and ARP Requests
(later)
• How many addresses can be kept in
the table?
– Depends on the size of the cache,
but 1,024 addresses is common.
• What about Layer 2 broadcasts?
– Layer 2 broadcasts (DA = all 1’s) is
flooded out all ports.
4444
Layer 2
49
What happens here?
Prof. Dr. R.Nitsch, FH Darmstadt
Source Address Table
Port Source MAC Add. Port Source MAC Add.
1
1111
6
3333
1
2222
1
5555
1111 3333
• Notice the Source Address
Table has multiple entries
for port #1.
• The switch filters the
frame out port #1.
• But the hub is only a layer 1
device, so it floods it out all
ports.
Collision Domain
• Where is the collision
domain?
3333
1111 2222 5555
Layer 2
50
Switches
• 2.5.2006
• A switch is like a multiport bridge.
• Like bridges, switches learns the MAC addresses of active
devices connected to its ports. It is a OSI Layer-2 device.
• Switches use this information to build forwarding tables
(similar to bridge tables)
• Switching lessens congestion in Ethernet LANs by reducing the
traffic and increasing the bandwidth.
• Switches can easily replace hubs because switches work with
existing cable infrastructures.
• All switching equipment performs two basic operations:
1. Switching data frames: a process by which a frame is
received on an input medium and then transmitted to an
output medium.
2.Build and maintain switching tables and search for loops.
• Switches operate at much higher speeds than bridges and can
support new functionality, such as virtual LANs.
• an Ethernet switch allows many users to communicate in
parallel through the use of virtual circuits and dedicated
network segments in a virtually collision-free environment.
• This maximizes the bandwidth available on the shared medium.
Layer 2
Prof. Dr. R.Nitsch, FH Darmstadt
51
IEEE-802.11 Wireless LANs
• Portable PCs, PDAs, cameras,
automobiles, pets, security systems,
kitchen appliances, household plants are
going wireless connected to the
Internet
• Important wireless device classification
criateria are power, range and data rate.
• Bluetooth: low-power, short-range, lowrate cable replacement technology
• 802.11 family (WLAN): higher-power,
medium-range, higher-rate "access"technology
Prof. Dr. R.Nitsch, FH Darmstadt
802.11b
• transmits in an unlicensed radio spectrum at 2,4 GHz
• provides wireless Ethernet access at 11 Mbps
• defines the physical layer and Media Access Control (MAC) layer
• Physical layer uses DSSS which codes each bit into a chipping code. This is similar to
CDM, except now all mobile hosts use the same chipping code. Therefore DSSS is not
a CDM technique that controls multiple access to a shared medium.
• DSSS is instead a physical layer mechanism that spreads the signal energy over a
wider frequency range, thereby improving the receivers ability to recover the original
transmitted bits.
Layer 2
53
Wireless LAN Standards
Prof. Dr. R.Nitsch, FH Darmstadt
• IEEE defines the physical and Media Access Control (MAC) layer for wireless networks
• 802.11b may also be called Wi-Fi™ (wireless fidelity) or high-speed wireless Ethernet
and refers to systems that operate at 1, 2, 5.5 and 11 Mbps.
– Available in Europe; Supports roaming
– Operates in the unlicensed 2,4 GHz ISM-Band (Industrial, Scientific, and Medical
Band);
– provides wireless Ethernet access at 11 Mbps. In production networks, a more
realistic rating is 4-6 Mbps.
– Physical layer uses DSSS which codes each bit into a chipping code. This is similar
to CDM, except now all mobile hosts use the same chipping code. Therefore DSSS is
not a CDM technique that controls multiple access to a shared medium.
– DSSS is instead a physical layer mechanism that spreads the signal energy over a
wider frequency range, thereby improving the receivers ability to recover the
original transmitted bits.
• 802.11a WLAN devices
– Used in North Amerika and Japan
– operate in the 5-6 GHz transmission band.
– supply data throughput of 54 Mbps and with proprietary technology known as "rate
doubling" has achieved 108 Mbps. In production networks, a more realistic rating is
20-26 Mbps.
• 802.11g provides the same throughput as 802.11a (54 Mbps) but operates in the 2,4
GHz band with backward compatibility for 802.11b devices.
• All of the 802.11 standards have the same architecture and use the same MAC protocol
Layer 2
54
The global picture: a multiplicity of WLAN standards
Prof. Dr. R.Nitsch, FH Darmstadt
Features
802.11a
North America
Japan
802.11g
Europe
802.11b
Europe
Operating
Frequency Band
5 GHz Industrial,
Scientific and
Medical Band
2,4 GHz
2,4 GHz
Data Rates
6, 9, 12, 18, 24, 36,
48 and 54 Mbps
6, 9, 12, 18, 24, 36,
48 and 54 Mbps
1, 2, 5.5 and 11 Mbps
System
Orthogonal
Frequency Division
Multiplexing
(OFDM)
OFDM
DSSS (Direct
Sequence Spread
Spectrum)
>= 100 m
100m
Approximate Range 50 m
Layer 2
55
802.11 LAN Architecture
Prof. Dr. R.Nitsch, FH Darmstadt
• Cells are the fundamental building blocks of 802.11 LANs also known as the basic
service set (BSS)
• A BSS contains one or more mobile stations and one central base station known as
access point (AP)
• Multiple APs may be conntected to form a so called distribution system (DS)
• The DS appears to upper layer protocols (e.g. IP) as a single 802 network.
• Cell overlapping will permit roaming between cells, allowing for the disconnect and
reconnect activity to occur seamlessly without service interruption.
Infrastructure Mode:
Wireless end-points
communicate via a
wireless relaying
Access Point device
BSS
BSS
f1
f3
AP
BSS
AP
AP
f2
AP
AP
f2
Infrastructure Mode
f1
cells
Layer 2
Ad-hoc mode
Ad-hoc Mode:
Wireless end-points
communicate
directly with each
other. An Access
Point is not involved.
57
IEEE 802.11 Media Access Protocol - Hidden Terminal Problem
Prof. Dr. R.Nitsch, FH Darmstadt
•
IEEE 802.11 Media Access Protocol is carrier sense multiple access with collision
avoidance (CSMA/CA). It can not implement the collision detection method, because
of the hidden terminal problem which is common to all wireless communications.
Hidden terminal problem description
a) Physical obstructions may prevent A and C from detecting each others transmissions,
even though A's and C's transmissions are receiveable at destination B.
b) Collisions are also undetectable when A and C are placed that their signal strengths
are sufficient for destination B but not strong enough to detect each others
transmissions and hence the collision at B.
Layer 2
58
IEEE 802.11 Media Access Protocol
Prof. Dr. R.Nitsch, FH Darmstadt
• IEEE 802.11 frames contain a "Duration" field in which the sending station indicates the
length of time ( also called network allocation vector (NAV) ) that it's frame will keep the
channel busy.
• The NAV determines the minimum amount of time other stations should defer their
access
• CSMA protocol first senses the channel to
determine wether it is busy or not.
• If the channel is sensed idle for a time
minimum given by the Distributed Inter
Frame Space (DIFS), a station is allowed
to transmit.
• This frame will be successfully received if
no other station's transmission will
interfere.
• Each completely and correctly received
frame is acknowledged to the sender.
After a short period of time, known as the
Short Interframe Spacing (SIFS) the
receiver sends a short achnowledgement
frame back to the sender.
• This is necessary because wireless
stations are not able to detect collisions
reliable (Hidden terminal problem)
Layer 2
Data transmission and acknowledgement in 802.11
59
802.11 Backoff Mechanism
Prof. Dr. R.Nitsch, FH Darmstadt
• When the sender senses the channel to be busy it performs a backoff
procedure similar to that of Ethernet:
– The sender waits until the channel is idle
– The sender additionally waits for DIFS time.
– If the channel stayed idle, the sender computes an additional random
backoff time.
– When the backoff timer reaches zero, the node transmits the frame
– As in the case of Ethernet, the backoff intervall is doubled each time the
retransmitted frame experiences a collision
Layer 2
60
IEEE 802.11 Media Access Protocol
• 802.11 protocol can also use short frames, called
RTS (Request to send) and CTS (clear to send), to
reserve access to the channel.
• Senders willing to send, first send a RTS frame to
the receiver indicating the expected channel
allocation duration (data and ACK; NAV).
• The receiver responds with a CTS frame, giving
the sender explicit permission to send.
• All other stations hearing the RTS or CTS then
know about the pending transmission and can avoid
interfering with it.
• RTS/CTS helps avoid collisions in 2 important ways
– The CTS frame helps avoid the hidden station
problem.
– Collisions involving RTS/CTS frames will last
only for a short time.
• Note:
– After correctly transmitted RTS/CTS frames
there should be no collisions.
– Without any hidden nodes, then the use of
RTS/CTS will only increase the amount of
overhead, which reduces throughput.
• Each 802.11 sender can operate with or without
using the RTS/CTS frames
Layer 2
Prof. Dr. R.Nitsch, FH Darmstadt
Collision avoidance using
the RTS and CTS frames
NAV: Network Allocation Vector
61
Wireless devices and topologies
Prof. Dr. R.Nitsch, FH Darmstadt
• When a client is activated within the WLAN, it will start "listening"
for a compatible device with which to "associate".
• This is referred to as "scanning" and may be active or passive.
– Active scanning causes a request to be sent from the wireless
node seeking to join the network.
– The request will contain the Service Set Identifier (SSID) of
the network it wishes to join.
– When an AP with the same SSID is present, the AP will issue a
response.
• The authentication and association steps are completed.
• From webpedia.com
– SSID: a 32-character unique identifier attached to the header of
packets sent over a WLAN that acts as a "password" when a mobile
device tries to connect to the BSS. The SSID differentiates one WLAN
from another, so all access points and all devices attempting to connect to
a specific WLAN must use the same SSID. A device will not be permitted
to join the BSS unless it can provide the unique SSID. Because an SSID
can be sniffed in plain text from a packet it does not supply any security
to the network.
Layer 2
62
How wireless LANs communicate
Prof. Dr. R.Nitsch, FH Darmstadt
• Performance of the network will
also be affected by signal strength
and degradation in signal quality due
to distance or interference.
• As the signal becomes weaker,
Adaptive Rate Selection (ARS) may
be invoked.
• The transmitting unit will drop the
data rate from 11 Mbps to 5.5
Mbps, from 5.5 Mbps to 2 Mbps or
2 Mbps to 1 Mbps.
Layer 2
63
The radio wave and microwave spectrums
Prof. Dr. R.Nitsch, FH Darmstadt
• Computers send data signals
electronically.
• Radio transmitters convert these
electrical signals to radio waves.
• However, radio waves attenuate
as they move out from the
transmitting antenna.
• In a WLAN, a radio signal measured
at a distance of just 10 meters (30
feet) from the transmitting
antenna would be only 1/100th of
its original strength.
• As the signal becomes weaker,
Adaptive Rate Selection (ARS) may
be invoked.
• The transmitting unit will drop the
data rate from 11 Mbps to 5.5
Mbps, from 5.5 Mbps to 2 Mbps or
2 Mbps to 1 Mbps.
Layer 2
64
Radio Interference: Signals and noise on a WLAN
Prof. Dr. R.Nitsch, FH Darmstadt
• When using RF technology many kinds of
interference must be taken into
consideration.
• In homes and offices, a device that is
often overlooked as causing interference is
the standard microwave oven.
• Leakage from a microwave oven of as little
as one watt into the RF spectrum can cause
major network disruption.
• Wireless phones operating in the 2.4 GHz
spectrum can also cause network disorder.
• The RF signal can be affected by some
extreme weather conditions such as fog or
very high moisture conditions. Lightning can
also charge the atmosphere and alter the
path of a transmitted signal.
Layer 2
65
Methods of Authentication
Prof. Dr. R.Nitsch, FH Darmstadt
•
•
•
•
WLAN authentication occurs at Layer 2.
WLAN authentication authenticates the device not the user.
IEEE 802.11 lists two types of authentication processes.
The first authentication process is the "open system".
– This is an open connectivity standard in which only the SSID must match.
– This may be used in a secure or non-secure environment despite the
ability of low level network ‘sniffers’ to discover the SSID of the WLAN
is high.
• The second process is the "shared key".
– This process requires the use of Wireless Equivalency Protocol (WEP)
encryption.
– WEP is a fairly simple algorithm using 40 and 128 bit keys.
– The AP is configured with an encrypted key and nodes attempting to
access the network through the AP must have a matching key.
– Statically assigned WEP keys provide a higher level of security than the
open system but are definitely not hack proof.
• The problem of unauthorized entry into WLANs is being addressed by a
number of new security solution technologies.
Layer 2
66