Vous êtes sur la page 1sur 134

Datalink Layer Protocols

(Chapter 5)
(Ethernet)
Idawaty Ahmad, PhD
(SKR3309)

Idawaty Ahmad, PhD


1 18 March 2017
Network Protocols
Objectives

* Upon completion of this course, you will be able to:


* Differentiate between various types of LAN
technologies
* Describe how the Ethernet works
* Perform basic Ethernet port configuration

Idawaty Ahmad, PhD


18 MarchPage2
2017
Network Protocols
Contents

1. Datalink Layer Protocols


2. LAN Overview
3. Ethernet
4. Ethernet Device Principle
5. Ethernet Port
6. Ethernet Configuration

Idawaty Ahmad, PhD


3 18 March 2017
Network Protocols
TCP/IP and OSI Model

Idawaty Ahmad, PhD


4 18 March 2017
Network Protocols
Function of Datalink Layer

* Framing – divides the stream data to manageable data units –


frame
* Detect Physical addressing (MAC address) – adds a header to
the frame –to define the physical address of sender( source
address) and receiver (destination address)
* Flow control – to prevent overwhelming at the receiver
* Error control – provides reliability – to detect and retransmit
damaged or lost frames, also prevent duplication of frames -
trailer
* Access control – require a protocol to determine which device
has control over the link at any given time –> same link with two
or more devices connected.

Idawaty Ahmad, PhD


5 18 March 2017
Network Protocols
TCP/IP Protocol Stack
Provide access to network
HTTP, Telnet, FTP resources; provide syntax and
Application Layer
TFTP, SMTP semantics

Provide end-to-end message


Transport Layer TCP/UDP delivery

Provide internetworking: move data


Network Layer IP,RIP,OSPF,BGP etc from source to destination

Data Link Layer Ethernet (802.3), PPP Organize bits into frames; to
provide node-to-node delivery
Physical Layer Interface and cables
Provide mechanical and electrical
specification; transfer data in the
Idawaty Ahmad, PhD
6
form of bits 18 March 2017
Network Protocols
Protocols of the Data Link Layer
* Data link layer LAN, WAN protocols

LAN WAN

802.2 LLC
Data Link Dial
Frame
on SDLC HDLC PPP
(Frames) Relay
Demand
8 8 8
F
0 0 0
Physical D V.24
2 2 2
D G.703
(Bits, . . .
I
signals, 3 4 5 V.35 EIA/TIA-232
clocking) xDSL

l Data link layer device

[ Ethernet(802.3),, Wireless LAN(802.11)


Idawaty Ahmad, PhD
7 18 March 2017
Network Protocols
Data Link Layer – MAC Address
* MAC address is composed of 48 bits. The vendor ID of Huawei
products is 0x00e0fc.

24 bits 24 bits

Vender Code Serial Number

00e0.fc01.2345
00e0.fc01.2345
Rom

Idawaty Ahmad, PhD


8 18 March 2017
Network Protocols
MAC Address
48 bits

24 bits 24 bits

Vender Code
Serial Number

example:00e0.fc39.8034
00e0.fc— the vender code that IEEE assigns to manufacturers
39.8034— serial number assigned by the manufacturer
Idawaty Ahmad, PhD
9 18 March 2017
Network Protocols
MAC Address ( Ethernet address and
Wireless LAN (Wifi) Address

Idawaty Ahmad, PhD


10 18 March 2017
Network Protocols
Contents

1. Datalink Layer Protocols


2. LAN Overview
3. Ethernet
4. Ethernet Device Principle
5. Ethernet Port
6. VLAN
7. Configuration of VLAN

Idawaty Ahmad, PhD


11 18 March 2017
Network Protocols
LAN Overview

Idawaty Ahmad, PhD


12 18 March 2017
Network Protocols
LAN Technologies and OSI Reference
Model

Network Layer IP, IPX

LLC sub
802.2 LLC/SNAP
layer
Data Link Layer
MAC sub 802.3 (Ethernet)、802.11
layer (wireless LAN)

Coaxial cable, twisted pair, fiber


Physical Layer optic,
electromagnetic waves

OSI Reference Model LAN


LAN Topologies

Bus Star Tree

Ring Mesh
Bus Topology

* Uses a multi-drop transmission medium


* All nodes (computers) on the network share a
common bus and thus share communication
* Allow only one device to transmit at a time. Thus it
requires a multiple access technique for the nodes
* The bus topology is passive, i.e. the computers on the
bus only listen for a signal, but does not regenerate
the signal
* Simple and low cost
Idawaty Ahmad, PhD
15 18 March 2017
Network Protocols
Star Topology

* Each node is connected to a single point (sever or


switching hub)
* The central server switches data around the network
* Central control: single point of failure
* More cabling, hence higher cost

Idawaty Ahmad, PhD


16 18 March 2017
Network Protocols
Ring Topology

* Nodes are connected in a circle (ring) of point-to-


point connections
* Each node in the ring has equal access but only one
node can talk at a time
* To coordinate this, token passing is usually used
* A token is passed from one node to the next around the
ring
* A node can only transmit when it captures the token
* It is difficult to add/remove a node from the ring
Idawaty Ahmad, PhD
17 18 March 2017
Network Protocols
Mesh Topology

* Full mesh topology


* All nodes are connected directly to all other nodes
* High reliability as many redundant paths exist
* However, it requires extensive cabling: high cost
* Let N be the number of nodes, number of connections
(links) = N(N-1)/2

Idawaty Ahmad, PhD


18 18 March 2017
Network Protocols
Mesh Topology (Cont.)

* Partial mesh topology


* Contains a subset of the full mesh links
* More cost effective and provides redundancy
* Since not all nodes are directly connected, intermediate nodes are
used as relay for some source-destination pairs
* Consequently, routing algorithms are needed for communication

Idawaty Ahmad, PhD


19 18 March 2017
Network Protocols
Tree Topology

* A hierarchical topology with multiple levels, with the root as the


top of the tree
* Each node has one parent (except the root) and may have many
child nodes or none
* Tree provides loop-free structure – simplifies routing over the
tree
* However, when one node fails, the tree will be partitioned

Root Level 1

Level 2

Level 3
Idawaty Ahmad, PhD
20 18 March 2017
Network Protocols
Unicast, Broadcast, and Multicast

* Unicast
* Transmission from one point to another point, i.e., one sender
and one receiver
* Broadcast
* Transmission from one point to all other points
* Broadcast Address: FF FF FF FF FF FF
* Multicast
* Transmission from one point to some selected points, i.e., one
sender and multiple receivers
* The receivers are identified by a group address (or multicast
address)

Idawaty Ahmad, PhD


21 18 March 2017
Network Protocols
Unicast

Host A
Source

Host B

Host C

Data to A
Data to B

Idawaty Ahmad, PhD


22 18 March 2017
Network Protocols
Broadcast

Host A
Source

Host B

Host C

Send to all

Idawaty Ahmad, PhD


23 18 March 2017
Network Protocols
Multicast

Host A
Source

Host B

Host C

Send to multicast group


Host B and Host C
belong to the same
multicast group
Idawaty Ahmad, PhD
24 18 March 2017
Network Protocols
Collision Domains and Broadcast
Domains

* Collision domain:
* Refers to a set of LAN devices whose frames could
collide with one another
* This happens with hubs, repeaters, and wireless access
points as only one device can send and receive at a time
* Broadcast domain:
* Refers to a set of devices where if one device sends a
broadcast frame, all other devices will receive where a
broadcast frame sent by one device is received by all
other devices in the same broadcast domain
Idawaty Ahmad, PhD
25 18 March 2017
Network Protocols
Collision Domains and Broadcast
Domains (Cont.)

Idawaty Ahmad, PhD


Broadcast domain
26 18 March 2017
Network Protocols Collision domain
Collision Domain
* Collisions are often in a hub environment, because
each port on a hub is in the same collision domain.
* By contrast, each port on a bridge, a switch or a
router is in a separate collision domain.

6 Collision
Domain

Idawaty Ahmad, PhD


27 18 March 2017
Network Protocols
Broadcast Domain
* All ports on a hub or a switch are by default in the
same broadcast domain. All ports on a router are in
the different broadcast domains and routers don’t
forward broadcasts from one broadcast domain to
another.

3 Broadcast
Domain

Idawaty Ahmad, PhD


28 18 March 2017
Network Protocols
Ethernet Overview

* Originally developed at Xerox


PARC
* The transmission rate has
evolved from 10Mbps Ethernet
Standard Ethernet, 100Mbps
Fast Ethernet, 1Gbps Gigabit
Ethernet to 10Gbps Ethernet
* Today, Ethernet is the
mainstream LAN technology

Idawaty Ahmad, PhD


29 18 March 2017
Network Protocols
Token Ring Overview

* It was first designed and developed Token


Ring
by IBM, and later accepted as an Network
IEEE standard: IEEE802.5
* Physical connections: star topology.
Logically, the devices are organized in a ring topology.
Token passing is used for multiple access
* At present, the token ring network usage is declining,
and its technology development is at a standstill
Idawaty Ahmad, PhD
30 18 March 2017
Network Protocols
FDDI Overview

l FDDI (Fiber Distributed


FDDI Ring
Data Interface) is another Network

ring-based LAN technology,


standardized by ANSI
l FDDI LAN provided better reliability and bandwidth than Ethernet
and Token Ring in early to mid 1990s
l However, the advantages of FDDI have disappeared with the
continuous development of Ethernet
Idawaty Ahmad, PhD
31 18 March 2017
Network Protocols
WLAN Overview
Internet

WLAN

AP

AP: Access Point

* Wireless LAN (WLAN) uses wireless transmission


technology to replace the need for wires
* It has become more and more popular due to ease of
installation (no cables) and the convenience of mobility
* IEEE 802.11 (Wi-Fi) is the predominant WLAN standard
Idawaty Ahmad, PhD
32 18 March 2017
Network Protocols
Contents

1. Datalink Layer Protocols


2. LAN Overview
3. Ethernet
4. Ethernet Device Principle
5. Ethernet Port
6. Ethernet Configuration

Idawaty Ahmad, PhD


33 18 March 2017
Network Protocols
History of Ethernet

* In 1973, Doctor Robert Metcalfe together with David


Boggs invented Ethernet, a technology for local area
networks with an initial data transmission rate of
2.94Mbps.
* After thirty years of development, Ethernet has
developed into an IEEE standard supporting
transmission rates in excess of 10000Mbps.

Idawaty Ahmad, PhD


34 18 March 2017
Network Protocols
Brief Development History of
Ethernet
IEEE802.3 Ethernet standard 10G Ethernet
IEEE802.3u 100BASE-T Fast Ethernet
Gigabit Ethernet
IEEE802.3z/ab 1000Mb/s Gigabit Ethernet
IEEE802.3ae 10GE Ethernet 100M Fast Ethernet

Shared turn to LAN switch

10M Ethernet

Ethernet

1970s 1980s 1990s 1992 1996 2002

Idawaty Ahmad, PhD


35 18 March 2017
Network Protocols
Shared Ethernet Transmission Media

10Base5: A thick coaxial cable. The ‘5’ refers to a


maximum transmission distance of 500 meters.

10Base2: A thin coaxial cable . The ‘2’ refers to a


maximum transmission distance of 185 meters.

Idawaty Ahmad, PhD


36 18 March 2017
Network Protocols
Shared Ethernet Medium
10Base-T: twisted-pair

Twisted wire pair RJ45

scarfskin

STP Twisted wire UTP Twisted wire pair


pair
Idawaty Ahmad, PhD
37 18 March 2017
Network Protocols
Pin Order of Category 5 Twisted-pair
Side 1 Side 1 Side 2
12345678
Straight Cable 1=white/orange 1=white/orange
2=orange 2=orange
3=white/green 3=white/green
4=blue 4=blue
5=white/blue 5=white/blue
6=green 6=green
12345678 7=white/brown 7=white/brown
Side 2 8=brown 8=brown
Side 1 Side 1 Side 2
Crossover Cable 12345678 1=white/orange 1=white/green
2=orange 2=green
3=white/green 3=white/orange
4=blue 4=blue
5=white/blue 5=white/blue
6=green 6=orange
12345678 7=white/brown 7=white/brown
Side 2 8=brown 8=brown
Idawaty Ahmad, PhD
38 18 March 2017
Network Protocols
General View of 802.3 Cable

Standard Cable type Transmission distance

10BASE-5 Thick coaxial-cable 500m

10BASE-2 Thin coaxial-cable 200m

10BASE-T Twisted-pair 100m

Standard Ethernet(10Mbit/s)defined by IEEE 802.3

Standard Ethernet (10Mbit/s)usually orients at access


layer
Idawaty Ahmad, PhD
39 18 March 2017
Network Protocols
Fast Ethernet

Fast Ethernet is a LAN based technology with a capability for


data transmission rates of 100Mbps. It can provide higher
bandwidth for terminal users, servers and server groups.

IEEE802.3u is defined by IEEE as the standard for Fast


Ethernet supporting two medium types.
* Twisted-pair

* Fiber

Idawaty Ahmad, PhD


40 18 March 2017
Network Protocols
General View of Fast Ethernet

Transmission
standard Cable type
distance
EIA/TIA category 5 (UTP) Unshielded Twisted
100BaseTX 100m
Paired 2pair

EIA/TIA category 3, 4, 5 (UTP) Unshielded


100BaseT4 100m
Twisted Pair (4pair)

Multiple mode fiber (MMF) 550m-2km


100BaseFX
Single mode fiber (SMF) >2km

18 March 2017
Gigabit Ethernet

Gigabit Ethernet is the extension of the IEEE802.3


standard, for which transmission speeds of 1Gbps are
achieved.
Two standards are defined:
* IEEE802.3z (fiber and copper cable)
* IEEE802.3ab(twisted-pair)

Idawaty Ahmad, PhD


42 18 March 2017
Network Protocols
General View of Gigabit
Ethernet
Transmission
standard Cable type
distance
Copper EIA/TIA type 5(UTP)Unshielded
1000BaseT 100m
Twisted Paired 4 pair

1000BaseCX Copper Shielded Twisted Pair 25m

Multiple mode fiber, 50/62.5um fiber, uses


1000BaseSX 550m/275m
laser which wavelength is 850nm
Single mode fiber, 9um fiber, uses laser
1000BaseLX 2km-15km
which wavelength is 1300nm

18 March 2017
General View of 10G Ethernet Cable
Transmission
standard Cable type
distance
Multimode fiber, 62.5/125um fiber or 50/125um
10GBASE-SR/SW 2m-300m
fiber

Multimode fiber, 62.5/125um fiber or 50/125um


10GBASE-LR/LW 2m-10km
fiber; single mode fiber, 10/125um fiber

10GBASE-ER/EW 10/125um fiber 2m-40km


Multiple mode
Multimode fiber, 62.5/125um fiber or 50/125um
10GBase-LX4 300m, single
fiber; single mode fiber, 10/125um fiber
mode 10km
10GBase-CX4 coaxial copper cable 15m
10GBase-T Twisted-pair copper cable 100m
Standards for 10G include IEEE 802.3ae (fiber), IEEE 802.3ak
(copper cable) and IEEE 802.3an(twisted-pair) 18 March 2017
Mainly applied between core nodes over backbone networks
Contents

1. Datalink Layer Protocols


2. LAN Overview
3. Ethernet
4. Ethernet Device Principle
5. Ethernet Port
6. Ethernet Configuration

Idawaty Ahmad, PhD


45 18 March 2017
Network Protocols
Hub and Switch

* Ethernet relies on the operation of devices such as


the HUB and Switch to provide lower layer operation.
* The principle characteristics of such devices is crucial
to fully understanding the behavior surrounding
frame transmission

Idawaty Ahmad, PhD


46 18 March 2017
Network Protocols
Shared Ethernet Principles: CSMA/CD
CS Carrier Sense
* The LAN device listens to the Ethernet network to sense
the carrier signal on the network. This ensures that line
is idle to reduce the likelihood of collisions occurring.
MA Multiple Access
* Traffic can be received by multiple devices.
CD Collision Detection
* Detects the collision as transmission occurs. If there has
been a collision, the LAN device wait for a random
amount of time to retransmit.

Idawaty Ahmad, PhD


47 18 March 2017
Network Protocols
HUB Operation

Application Application
layer layer
Presentation Presentation
layer layer
Session Session
layer layer
Transport Transport
layer layer
Network Network
layer layer
HUB
Data Link Data Link
layer layer
Physical Physical Physical Physical
layer layer layer layer

Idawaty Ahmad, PhD


Note:A HUB is just a connection
48 18 March 2017
Network Protocols device at physical layer
HUB Operation

1 2 3 4 5

OU OU OU OU
IN T T T T

All the HUB devices are half-duplex

Idawaty Ahmad, PhD


49 18 March 2017
Network Protocols
Weaknesses of HUB based
Ethernet Connections

A network connected to a HUB is considered as shared


Ethernet, and has all the same weaknesses of shared
Ethernet:
* Collisions;
* Broadcast flooding;
* No guarantee of security

Idawaty Ahmad, PhD


50 18 March 2017
Network Protocols
Ethernet Switch

* The Ethernet switch operates at the data-link layer


and has two basic functions:
* Learning MAC addresses
* Switching or filtering data
* Preventing Loop

Idawaty Ahmad, PhD


51 18 March 2017
Network Protocols
Bridge/L2 Ethernet Switch
Operation
Application Application
layer layer
representation representation
layer layer
Session Session
layer layer
Transport Transport
layer L2 Switch layer
Network Network
layer layer
Data Link Data Link Data Link Data Link
layer layer layer layer
Physical Physical Physical Physical
layer layer layer layer

Idawaty Ahmad, PhD


52 18 March 2017
Network Protocols
Switching…

Idawaty Ahmad, PhD


53 18 March 2017
Network Protocols
Ethernet II Frame

0800 IP datagram

0806 ARP request/response

Idawaty Ahmad, PhD 8035 RARP request/response


54 18 March 2017
Network Protocols
802.3 Frame Structure
DMAC SMAC Length DATA/PAD FCS

ORG
DSAP SSAP CTRL CODE TYPE DATA
1 1 1 3 2 38~1492

0800 IP datagram

0806 ARP request/response

8035 RARP request/response


Idawaty Ahmad, PhD
55 18 March 2017
Network Protocols
How do switches work?

* After taking a switch out the box, plugging it in, and


connecting devices to it, the switch goes through the
following processes:
* Learning Process
* Filtering
* Forwarding
* Flooding

Idawaty Ahmad, PhD


56 18 March 2017
Network Protocols
Learning Process

* As a PC sends a frame to another device through the


switch, the switch captures the source MAC address
of the frame and the interface that received it.
* The switch confirms or adds the MAC address and the
port to the lookup table.
* A switch also keeps a timer for each of the MAC
address entries in its lookup table. (Aging timer)
* By default, many vendors set this time to hold an
address entry to 300 seconds (5 minutes)
Idawaty Ahmad, PhD
57 18 March 2017
Network Protocols
Learning MAC Address

Idawaty Ahmad, PhD


58 18 March 2017
Network Protocols
Flooding

A switch will flood the single frame out all of its other ports
when it cannot find the destination MAC address in the
Idawaty Ahmad, PhD
Network Protocols
switch’s lookup
59 table 18 March 2017
Forwarding and Filtering

* When a switch has learned the locations of the


devices connected to it, the switch is ready to either
forward or filter frames based on the destination
MAC address of the frame and the contents of the
switch lookup table.
* The switch filters out (or does not send the frame to)
other ports on the switch since they do not have the
target MAC address in the lookup table. That way, no
one else can look at the contents of the frame.

Idawaty Ahmad, PhD


60 18 March 2017
Network Protocols
Forwarding and Filtering

Idawaty Ahmad, PhD


61 18 March 2017
Network Protocols
Flooding

* A switch also floods all of its other ports under two other
circumstances:-
* If a broadcast frame is received with a destination MAC address
ff ff ff ff ff ff, the switch forwards the broadcast frame out all of
the other ports. VLAN exceptions!!
* If a multicast frame is received, the switch forwards the multicast
frame
* The flooding process for broadcast is necessary for protocols
such as ARP which maps an IP address to a MAC address via
broadcasting. One significant problem with this flooding process
occurs when switches in a network are redundantly connected
into physical loops. Flooding broadcast frames can cause a
“broadcast storm” on a network, causing it to fail.
Idawaty Ahmad, PhD
62 18 March 2017
Network Protocols
Flooding

* The flooding process for broadcast is necessary for


protocols such as ARP which maps an IP address to a
MAC address via broadcasting.
* One significant problem with this flooding process
occurs when switches in a network are redundantly
connected into physical loops.
* Flooding broadcast frames can cause a “broadcast
storm” on a network, causing it to fail.
* Refer to STP notes
Idawaty Ahmad, PhD
63 18 March 2017
Network Protocols
L2 Switch Principles

Frame Input

Destination
unicast broadcast
MAC address
Lookup Frame flooding
address table (except source port)
Not matched

matched

Forwarding frame
(except source port)
Idawaty Ahmad, PhD
64 18 March 2017
Network Protocols
Three Switch Modes
Cut-Through
* Forward as soon as it receives destination MAC address.
* Low delay
* No error check
Store-and-Forward
* Forwards only after it receiving full frame
* Delay lies on frame length
* Switch check error, the error packet will be discarded
Fragment-free
* Switch receives the first 64 bytes of the packet, then performs
a lookup up of the address table by header and forwards
* Switch check the first 64 bytes, if it finds error, the packet will
be discarded.

Idawaty Ahmad, PhD


65 18 March 2017
Network Protocols
L2 Switch Limitations

The L2 switch has brought about important


improvements in Ethernet technology, solving problems
with collisions, enhancing Ethernet capabilities
remarkably, and the security is also enhanced. There still
remain some Ethernet shortcomings:
* Broadcast flooding
* No security assurance
The main shortcoming of these is broadcast flooding.

Idawaty Ahmad, PhD


66 18 March 2017
Network Protocols
L3 Switch Features

1. From a logical point of view, L3 switching is the same


as routing, the process of L3 switching is just the
process of routing.
2. Layer 3 switch divides a Layer 2 network into
multiple VLANs. They implement Layer 2
switching within the VLANs and Layer 3 IP
connectivity between VLANs.

Idawaty Ahmad, PhD


67 18 March 2017
Network Protocols
Select L2 Switch or L3 Switch
Frame input

VLAN
attribution

Whether it is
route interface’s
MAC address

L3 forwarding L2 forwarding
Inter-VLAN Inside VLAN

Idawaty Ahmad, PhD


68 18 March 2017
Network Protocols
Router vs. L3 Switch

* Routers use CPUs (software) to complete Layer 3


forwarding, whereas Layer 3 switch uses hardware to
complete Layer 3 forwarding.
* Hardware forwarding has a much higher performance than
software forwarding (CPU based forwarding).
* Layer 3 Switches cannot replace routers in all scenarios
because routers provide rich interface types, good service
class control, and powerful routing capabilities that Layer 3
switch cannot provide.

Idawaty Ahmad, PhD


69 18 March 2017
Network Protocols
Routers

* Traditionally used a Bus Architecture


* Each packet has to pass over the bus twice
* From the incoming interface to the processor
* From the processor to the outgoing interface
* This ties up the router for two bus cycles
Processor
Bus

Interface Interface
Idawaty Ahmad, PhD
70 18 March 2017
Network Protocols
Layer 3 Switches

* Interfaces are Connected by a Switching Matrix


* Can handle multiple frame forwardings simultaneously

Interface Interface

IP Packets Switching Matrix

Interface Interface
Idawaty Ahmad, PhD
71 18 March 2017
Network Protocols
Routers vs Layer 3 Switches

* Both forward on the basis of IP addresses


* But Layer 3 switches are faster and cheaper
* However, Layer 3 switches are more limited
* Usually do not support WAN connections and only Ethernet for
LANs
* Usually support only TCP/IP
* There are other differences as well

Idawaty Ahmad, PhD


72 18 March 2017
Network Protocols
Routers

* Forward each packet independently


* Ignore previous IP packets to the same destination
* Make a full router forwarding decision (Ch. 3) for each
arriving IP packet
* This takes a great deal of processing time
* This makes routers slow and inexpensive
* However, this allows routers to adjust rapidly to failures
Idawaty Ahmad, PhD
73 18 March 2017
Network Protocols
Layer 3 Switches

* Do NOT forward each packet independently


* For instance decision caching; Remember previous
decisions based on IP address
* No need to make a full router forwarding decision for
each arriving IP packet
* This saves a great deal of processing time
* This makes Layer 3 switches fast and inexpensive
* However, this may make failure recovery slow
Idawaty Ahmad, PhD
74 18 March 2017
Network Protocols
Contents

1. Datalink Layer Protocols


2. LAN Overview
3. Ethernet
4. Ethernet Device Principle
5. Ethernet Port
6. Ethernet Configuration

Idawaty Ahmad, PhD


75 18 March 2017
Network Protocols
Ethernet Port

1. Auto-Negotiation Technology
2. Port Aggregation
3. Port Mirroring
4. Flow Control

Idawaty Ahmad, PhD


76 18 March 2017
Network Protocols
Auto Negotiation

* The auto-negotiation function allows interfaces on


both ends of a link to select the same operating
parameters by exchanging capability information
* The interfaces negotiate the duplex mode, speed, and
flow control parameters. After a successful
negotiation, the interfaces use the same duplex
mode, speed, and flow control parameters.

Idawaty Ahmad, PhD


77 18 March 2017
Network Protocols
Auto Negotiation is enabled on the
Interface by default

Idawaty Ahmad, PhD


78 18 March 2017
Network Protocols
Auto-negotiation

100Mb/s auto-negotiation 100Mb/s full-duplex

Port 2 auto-negotiation 10Mb/s auto-negotiation

Port 3 auto-negotiation
Port 1 auto-negotiation

Port 4 auto-negotiation
Port 5 auto-negotiation

10Mb/s full-duplex

Idawaty Ahmad, PhD 10Mb/s half-duplex 79 18 March 2017


Network Protocols
Auto-Negotiation

* Auto-negotiation: determine the highest shared


operational mode (i.e., transmission speed, duplexing
mode) of Ethernet devices
* Implemented in physical layer chips. Thus it is
independent of the upper layer protocols
* Basic mechanism:
* Negotiation information is encapsulated into a series of linking conformity
testing pulses
* Clock information in the pulses allows the devices to learn about the
operational mode of link partner and set the operational mode
accordingly

Idawaty Ahmad, PhD


80 18 March 2017
Network Protocols
Precedence in Auto-Negotiation

* Basic principle: Precedence Duplexing Mode


* 100Mbps mode has 1 100BASE-TX full-duplex
higher precedence than
10Mbps mode 2 100BASE-T4
* Full duplex is better than 3 100BASE-TX
half duplex
4 10BASE-T full-duplex
* 100 BASE-T4 is listed
before 100 BASE-TX as 5 10BASE-T
100 BASE-T4 supports
more cable types
Idawaty Ahmad, PhD
81 18 March 2017
Network Protocols
Ethernet Port Duplex Mode

* Ethernet twisted pair interface works in the following


three duplex modes:
* Full-duplex: port sends and receives data packet at the same
time
* Half-duplex: port sends or receives data packet at one time
* Auto-negotiation: port duplex mode is negotiated by the port
and its peer’s port
* Ethernet optical interface can only work in full-duplex
mode

Idawaty Ahmad, PhD


82 18 March 2017
Network Protocols
Setting the Duplex Mode

Idawaty Ahmad, PhD


83 18 March 2017
Network Protocols
Ethernet Port Speed Settings

When port speed is set as auto-negotiation, the speed


will be determined by negotiation of the port and its
peer’s port.

Idawaty Ahmad, PhD


84 18 March 2017
Network Protocols
Setting the Rate of Interface (Speed)

Idawaty Ahmad, PhD


85 18 March 2017
Network Protocols
Port Aggregation

Logical
Link

* Also called link aggregation


* Link aggregation refers to a method of bundling a group of
physical interfaces into a logical interface to increase
bandwidth. It is also called multi-interface load sharing group or
link aggregation group. For details, refer to IEEE802.3ad.
* Port aggregation is realized at data link layer
* Also known as link aggregation
Idawaty Ahmad, PhD
86 18 March 2017
Network Protocols
Advantages of Port Aggregation

* Improve bandwidth capacity between devices


* Traffic load-balancing capability
* Improve reliability: each aggregated port provides
redundancy should a link fail

Data traffic

18 March 2017
Limitations of Port Aggregation

The physical parameters of aggregation ports of two


peers must be same
* Number of the aggregation ports
* Speed of the aggregation ports
* Duplex mode of the aggregation ports
The logical parameters of aggregation ports of two
peers must be same
* The basic configuration of ports in the same aggregation
group must be same, the basic configuration include STP,
QoS, VLAN, Port and so on.
Idawaty Ahmad, PhD
88 18 March 2017
Network Protocols
Port Aggregation Features

* Manual Load Balancing Mode


* Static LACP Mode
* Active and Inactive Interface
* Actor and Partner

Idawaty Ahmad, PhD


89 18 March 2017
Network Protocols
Manual Load Balancing Mode

* Manually add member interfaces to the link


aggregation group. All the interfaces configured with
load balancing are in forwarding state.
* Must set up the Eth-Trunk and add an interface to the
Eth-Trunk manually.
* The Link Aggregation Control Protocol (LACP) is not
used.
* The manual load balancing mode is usually used when
the peer device does not support LACP.
Idawaty Ahmad, PhD
90 18 March 2017
Network Protocols
Manual Load Balancing Mode
Configuration (Basic)

1. Create an Ethernet Trunk Interface (Eth-Trunk)


2. Adding Member Interfaces to an Eth-Trunk
2.1 Configuration in the Eth-Trunk interface view
2.2 Configuration in the member interface view

Idawaty Ahmad, PhD


91 18 March 2017
Network Protocols
1. Create an Ethernet Trunk interface
(Eth-Trunk)

Idawaty Ahmad, PhD


92 18 March 2017
Network Protocols
2.1 Configuration in the Eth-Trunk
interface view

Idawaty Ahmad, PhD


93 18 March 2017
Network Protocols
2.2 Configuration in the member
interface view

Idawaty Ahmad, PhD


94 18 March 2017
Network Protocols
Idawaty Ahmad, PhD
95 18 March 2017
Network Protocols
Static LACP Mode

* Link Aggregation Control Protocol (LACP)


* Static LACP mode is a link aggregation mode in which
the two parties negotiate aggregation parameters by
exchanging LACP packets.
* After the negotiation, the two parties determine the
active interface and the inactive interface.
* In static LACP mode, you need to create an Eth-Trunk
manually and add members to the Eth-Trunk. The
active interfaces and inactive interfaces are
determined by LACP negotiation.
Idawaty Ahmad, PhD
96 18 March 2017
Network Protocols
Static LACP

* The static LACP mode is also called the M:N mode.


* In this mode, links can implement load balancing and
redundancy at the same time.
* In a link aggregation group, M links are active and they
forward data in load balancing mode.
* N links are inactive and they function as backup links. The
backup links do not forward data. When an active link fails,
the backup link with the highest priority replaces the failed
link to forward data and its status changes to active.

Idawaty Ahmad, PhD


97 18 March 2017
Network Protocols
Static LACP vs. Manual Load
Balancing Mode

* In static LACP mode, some links function as backup links.


When a fault occurs on some links, the backup links replace
the faulty ones to keep data transmission uninterrupted.
The active links have the load balancing capability.

* In manual load balancing mode, all member interfaces


work in forwarding state to share the traffic. Only load
balancing. This is the main difference between the two
modes.

Idawaty Ahmad, PhD


98 18 March 2017
Network Protocols
Dynamic LACP Mode

* Link aggregation can also be implemented in dynamic


LACP mode.
* In dynamic LACP mode, LACP creates the Eth-Trunk
and adds member interfaces automatically without
human intervention.
* This mode is easy for users, but is too flexible and
hard for management; therefore, the S3700 does not
support dynamic LACP mode.

Idawaty Ahmad, PhD


99 18 March 2017
Network Protocols
Active and Inactive Interface

* Active interfaces refer to the interfaces that are in


active state and are responsible for forwarding data.
* The interfaces that do not forward data and are in
inactive state are called inactive interfaces.

Idawaty Ahmad, PhD


100 18 March 2017
Network Protocols
Active and Inactive Interface

* According to the operation modes, active and inactive


interfaces are classified as follows:
* Manual load balancing mode:
* Generally, all member interfaces are active interfaces unless a
fault occurs on these interfaces.
* Static LACP mode:
* The interfaces connected to M links are active interfaces that
are responsible for forwarding data; the interfaces connected
to N links are inactive interfaces that are used for redundancy
backup.

Idawaty Ahmad, PhD


101 18 March 2017
Network Protocols
Actor and Partner

* In static LACP mode, the device in the link


aggregation group with a higher LACP priority is the
Actor and the device with a lower LACP priority is the
Partner.
* If the two devices have the same LACP priority, the
Actor is selected based on the MAC addresses of the
devices. The device with a smaller MAC address
becomes the Actor.
* Differentiating the Actor and the Partner is to keep
the active interfaces of devices at both ends
Idawaty Ahmad,consistent.
PhD
102 18 March 2017
Network Protocols
Actor and Partner

Idawaty Ahmad, PhD


103 18 March 2017
Network Protocols
Static LACP Mode Configuration
(Basic)

1. Create an Ethernet Trunk Interface (Eth-Trunk)


2. Adding Member Interfaces to an Eth-Trunk
2.1 Configuration in the Eth-Trunk interface view
2.2 Configuration in the member interface view

Idawaty Ahmad, PhD


104 18 March 2017
Network Protocols
1. Create an Ethernet Trunk Interface
(Eth-Trunk)

Idawaty Ahmad, PhD


105 18 March 2017
Network Protocols
2.1 Configuration in the Eth-Trunk
interface view

Idawaty Ahmad, PhD


106 18 March 2017
Network Protocols
2.2 Configuration in the member
interface view

Idawaty Ahmad, PhD


107 18 March 2017
Network Protocols
Set the LACP priority of system

Idawaty Ahmad, PhD


108 18 March 2017
Network Protocols
Setting the LACP Priority of an
Interface

Idawaty Ahmad, PhD


109 18 March 2017
Network Protocols
Contents

1. Datalink Layer Protocols


2. LAN Overview
3. Ethernet
4. Ethernet Device Principle
5. Ethernet Port
6. Ethernet Configuration

Idawaty Ahmad, PhD


110 18 March 2017
Network Protocols
1. Perform basic Configuration on the
Ethernet switches
G0/0/9 G0/0/9
Trunk

S1 G/0/0/10 G0/0/10 S2

Idawaty Ahmad, PhD


111 18 March 2017
Network Protocols
1. Perform basic Configuration on the
Ethernet switches
G0/0/9 G0/0/9
Trunk

S1 G/0/0/10 G0/0/10 S2

Idawaty Ahmad, PhD


112 18 March 2017
Network Protocols
1. Perform basic Configuration on the
Ethernet switches (Verify)
G0/0/9 G0/0/9
Trunk

S1 G/0/0/10 G0/0/10 S2

Idawaty Ahmad, PhD


113 18 March 2017
Network Protocols
2.Manual Link Aggregation
Configuration

G0/0/9 G0/0/9
Trunk

S1 G/0/0/10 G0/0/10 S2

Idawaty Ahmad, PhD


114 18 March 2017
Network Protocols
2. Manual Link Aggregation
Configuration (Verify)
G0/0/9 G0/0/9
Trunk

S1 G/0/0/10 G0/0/10 S2

Idawaty Ahmad, PhD


115 18 March 2017
Network Protocols
3. Link Aggregation Configuration in
Static LACP Mode (If necessary)
G0/0/9 G0/0/9
Trunk

S1 G/0/0/10 G0/0/10 S2

Idawaty Ahmad, PhD


116 18 March 2017
Network Protocols
3. Link Aggregation Configuration in
Static LACP Mode

G0/0/9 G0/0/9
Trunk

S1 G/0/0/10 G0/0/10 S2

Idawaty Ahmad, PhD


117 18 March 2017
Network Protocols
3. Link Aggregation Configuration in
Static LACP Mode (Verify)
G0/0/9 G0/0/9
Trunk

S1 G/0/0/10 G0/0/10 S2

Idawaty Ahmad, PhD


118 18 March 2017
Network Protocols
3.Link Aggregation Configuration in
Static LACP Mode
G0/0/9 G0/0/9
Trunk

S1 G/0/0/10 G0/0/10 S2

Idawaty Ahmad, PhD


119 18 March 2017
Network Protocols
3.Link Aggregation Configuration in
Static LACP Mode (Verify)
G0/0/9 G0/0/9
Trunk

S1 G/0/0/10 G0/0/10 S2

Idawaty Ahmad, PhD


120 18 March 2017
Network Protocols
3. Link Aggregation Configuration in
Static LACP Mode (Verify)
G0/0/9 G0/0/9
Trunk

S1 G/0/0/10 G0/0/10 S2

Idawaty Ahmad, PhD


121 18 March 2017
Network Protocols
Thank you J

Idawaty Ahmad, PhD


122 18 March 2017
Network Protocols
Port Mirroring

* Port mirroring is a technique uses by a switch to make a


copy of data on one switch port and send the copy to
another port
* Two types of port mirroring
* Port based
* Flow based
* Usage of port mirroring
* Troubleshooting
* Traffic monitoring and observation, e.g., for intrusion
detection

Idawaty Ahmad, PhD


123 18 March 2017
Network Protocols
Port Based Mirroring

1. Port based mirroring allows a copy of sent and


received traffic on a port to be duplicated and
forwarded (mirrored) to an alternative port.
2. This can be used for traffic observation or
troubleshooting .
3. An Ethernet switch supports many to one mapping,
meaning the traffic of multiple ports can be
mirrored to a single monitor port.

Idawaty Ahmad, PhD


124 18 March 2017
Network Protocols
Port Mirroring (Cont.) Switch duplicates the
original data and
forward the duplicates
Internet to the desired port

SW1 E0/1

E0/2 Mirrored
Original
E0/3 data Monitoring
device
data

SW3 SW2

PC1
PC2

Idawaty Ahmad, PhD


125 18 March 2017
Network Protocols
Page125
Port-Based Mirroring
* Data from one port is copied and forwarded to
another port
* Usage: traffic observation and troubleshooting
Incoming data from
port E0/1 is mirrored at
Original port E0/2
data

E0/1
E0/2

Mirrored
data
Monitoring
device
Idawaty Ahmad, PhD
126 18 March 2017
Network Protocols
Flow-Based Mirroring

* Data traffic is mirrored based on some traffic


classification rules
* Usage: traffic monitoring and analysis
Incoming data that matched a
Data specified traffic classification
rule will be mirrored at port
E0/2
E0/1
E0/2

Mirrored
data
Monitoring
device
Idawaty Ahmad, PhD
127 18 March 2017
Network Protocols
Flow Based Mirroring
Flow based mirroring classifies mirrored traffic
according to classification rules, based which data will
be mirrored to the monitor port, again which is used
to analyze and monitor classified traffic.
A switch can only configure a single monitor port.
Service data of data traffic 1
Service data of data traffic 1

E0/1
E0/2
mirrored data of data
traffic 1

Idawaty Ahmad, PhD


128 18 March 2017
Network Protocols
Configuration of Port Mirroring

E0/1 E0/24

PC1 SW PC2
Configuration of SW:
[SW]observe-port 1 interface Ethernet 0/1
[SW]interface ethernet 0/24
[SW-Ethernet0/24] port-mirroring to observe-port 1
inbound

Idawaty Ahmad, PhD


129 18 March 2017
Network Protocols
Flow Control

Standard Ethernet
Congestion
Fast Ethernet
Gigabit Ethernet

Half-duplex Full-duplex

Backpressure PAUSE frame

Idawaty Ahmad, PhD


130 18 March 2017
Network Protocols
Flow Control in Half-Duplex Mode

* Half-duplex Ethernet uses a method called backpressure to regulate


the speed between devices of different speeds
* Basic concept:
* To force the sender to slow down, the switch “pretends” it has encountered
collision
* Due to CSMA/CD, the sender will defer transmission (backpressure)
* Backpressure is removed once the switch is recovered from congestion

Idawaty Ahmad, PhD


131 18 March 2017
Network Protocols
Flow Control in Full-Duplex Mode

* Backpressure doesn’t work in full-duplex mode. Why?


* A new method for flow control in full-duplex mode is
defined in IEEE 802.3x
* Switch creates a PAUSE frame destined for a reserved
multicast address: 01-80-C2-00-00-01
* The PAUSE frame is sent to the traffic source
* On receiving the PAUSE frame, the source will stop sending
* By using a reserved multicast address, the PAUSE will not be
further forwarded by switches

Idawaty Ahmad, PhD


132 18 March 2017
Network Protocols
Flow Control in Full-Duplex Mode

* Application scenarios of PAUSE frame:


* Between a pair of terminals, i.e., simple point-to-point
connection
* Between a switch and a terminal
* Between two switches

Idawaty Ahmad, PhD


133 18 March 2017
Network Protocols
Page133
Set Flow Control of Ethernet Port

To avoid message dropping, it need to enable flow control


on own port and peer

Implement the following configuration in Ethernet port view


* Enable flow control of Ethernet port
* flow-control
* Shutdown flow control of Ethernet port(default closed)
* undo flow-control

Idawaty Ahmad, PhD


134 18 March 2017
Network Protocols

Vous aimerez peut-être aussi