Transcript from Networks.

Internetworking
Hussain Ali, MS
[email protected]
Department of Computer Engineering
King Fahd University of Petroleum and Minerals
Dhahran, Saudi Arabia
What is Internetworking ?
Internetworking stands for
connectivity and communication
between two or more networks.
dropping the “s” from Networks.
How is Internetworking
Achieved ?
Cables and physical interfaces (physical
connectivity)
Protocols, management and applications
needed to support user (Internetworking)
Motivation for Internetworking
Overcome distance limitations and protocol
differences for more effective sharing of data
and resources
Productive communication between people
across a single network or multiple networks
Email, newsgroups, mailing lists, live conferencing
Components of an Internetwork
Campus Network
Locally connected users in a building or group of
buildings
Wide Area Networks (WANs)
Distant campuses connected together usually
through connection providers such as the phone
company
Remote Connections
Linking branch offices and mobile users to a
corporate campus
Campus Networks
A campus is a building or group of buildings
all connected into one enterprise network that
consists of many local area networks.
The distinct characteristic of a campus is that
the company network owns the physical wires.
Campus network topology is primarily LAN
technology connecting all the end systems
together.
Campus networks generally use LAN
technologies such as Ethernet, Token Ring,
FDDI, Fast Ethernet, and ATM.
Wide Area Networks
WAN communication occurs between
geographically separate areas.
In enterprise internetworks, WANs connect
campuses together.
When a local end station wants to
communicate with a remote end station,
information must be sent over one or more
WAN links.
WAN services are provided through the
following 3 primary switching technologies:
Circuit Switching
Packet Switching
Cell Switching
Remote Connection
Remote connections link single mobile users
and branch offices to a local campus.
Typically a branch office is a small site that
has few users and needs a low bandwidth
WAN connection.
These small sites or single users, seldom
need to remain connected 24 hours a day.
Remote connections are generally dial-up
links or low bandwidth dedicated WAN links.
Trends in LAN/WAN Integration
Today, most of the computing power resides
on the desktop, and this power is growing.
Distributed applications are increasingly
bandwidth hungry.
Voice communications have increased
significantly.
All of this is driving towards an integration of
LANs and WANS under one roof.
In the LAN, bandwidth is free and connectivity
is limited only by hardware.
In the WAN, bandwidth is an excessive cost.
The existence and development of bandwidth
sensitive traffic such as voice and real-time
video has forced a requirement of better and
more predictable LAN and WAN performance.
Interaction of Different Networks
1. LAN-to-LAN 2. LAN-to-WAN 3. WAN-to-WAN
4. LAN-to-WAN-to-LAN
Host Host
Host
Host
Host
802.5
LAN
MR
802.3 LAN
Host Host
B
802.4 LAN
SNA WAN
MR
802.3 LAN
MR
X.25 WAN
MR
Host Host
B: Bridge
MR: Multi-protocol router
Relays
Devices that interconnect LANs are
known as relays and operate at one
layer of OSI model
There are four common types of relays
Repeater: at physical layer (bits)
Bridge: at data-link layer (frames)
Router: at network layer (packets)
Gateways: at transport and higher
layers (protocols)
Repeater (Hub)
Overcomes restrictions caused by single
segment usage such as number of users,
cable length.
Amplifies or regenerates weak signals .
Extends cable length
Can connect LANs of a similar type but which
use different media.
Provides simple connection between adjacent
LANs at the expense of increased network
congestion
Use of Repeaters for a Multisegment LAN
Station
Station
Printer
Segment A
Repeater
Segment B
Stations
File
Server
Bridge
Interconnects two or more LANs (either similar
or dissimilar) at the MAC level.
Capable of deciding whether or not to forward
frame.
Creates an extended network and keeps local
traffic off.
Can make minor changes to frame header.
Does not inspect or modify the network layer
packets inside frames.
Characteristics of Bridges
 Routing Tables
 Filtering
 Forwarding
 Learning Algorithm
Routing table
Contains one entry per station of network to
which bridge is connected.
Is used to determine the network of
destination station of a received packet.
Filtering
Is used by bridge to allow only those packets
destined to the remote network.
Packets are filtered with respect to their
destination and multicast addresses.
Forwarding: the process of passing a packet
from one network to another.
 Learning: the process by which the bridge
learns how to reach stations on the
internetwork.

Operation of a LAN bridge from
802.3 to 802.4
Host A
Network
Packet
LLC
Packet
MAC
802.3 Packet
Physical
802.3 Packet
802.3
CSMA/CD
Host B
Bridge
Packet
Network
Packet
Packet
LLC
802.3
802.3
802.4
802.4
802.4
802.4 Packet
MAC
802.4 Packet
Physical
Token bus
Transparent Bridges
The first IEEE 802 bridge is a transparent
bridge or spanning tree bridge.
People wanted to have complete transparency:
when a site with multiple LANs buys bridges
designed to the IEEE standard, just plug
connectors into bridges. So,
no need for hardware/software changes,
no setting of address switches,
no downloading of routing tables or
parameters.
A transparent bridge accepts every frame
transmitted on all the LANs to which it is
attached.
LAN 4
A
B
LAN 1
Bridge
F
C
Bridge
LAN 2
G
H
D
LAN 3
E
Topology can change dynamically.
 There must be only one path of bridges and
LANs between any two segments in the bridged
LAN
 Bridges must support Spanning Tree Protocol if
network contains loops.
 Have the advantage of being easy to install
 Use only a subset of topology.
 Are chosen by the CSMA/CD and token bus.

Source Routing Bridges
Token ring people chose the source routing
bridge.
Transmitter, or source, of frame in source
routing specifies which route the frame is to
follow.
Every machine in the network knows, or can
find, the best path to every other machine;
discovery frame is used.
Sender knows whether or not the destination is
on its own LAN.
Comparison of Bridges
Issue
Transparent
Bridge
Connectionless
Source Routing Bridge
Not transparent
Configuration
Transparent to
hosts
Automatic
Routing
Suboptimal
Optimal
Locating
Destinations
Failures
Backward learning Discovery frames
Orientation
Transparency
Complexity
Handled by
bridges
In the bridges
Connection-oriented
Manual
Handled by hosts
In the hosts
Router
Provides a more intelligent service
makes a decision as to the best way to
deliver a packet from source to destination
may fragment packets to meet packet size
requirements of LANs
are slower than bridges
Permits translation between different address
domains such as addresses of IEEE 802 LAN
and X.25
Connects dissimilar networks, provided that
end-systems use a common network layer
protocol, such as IP.
Unlike bridge, router receive only those packets
addressed to it by either a user machine or
another router.
Select the best route.
The question of who owns, operates, and
maintains a router arises especially when two
networks belong to independent organizations.
Full Router and
Two Half-Routers
Buffer
Full Router:
Network 1
Net 1 to
internet
internet
to Net 1
Two-Half
Routers:
Net 1 to
internet
Machine owned jointly by both
networks
Net 2 to
internet
Network 2
internet
to Net 2
Net 2 to
internet
Network 2
Network 1
internet
to Net 1
internet
to Net 2
Disadvantages of Routers
Routers
are protocol-dependent devices that must
understand the protocol they are forwarding.
can require a considerable amount of initial
configuration.
are relatively complex devices, and generally are
more expensive than bridges.
Advantages of Routers
Routers

provide sophisticated routing, flow
control, and traffic isolation

are configurable, which allows network
manager to make policy based on routing
decisions

allow active loops so that redundant paths
are available
Gateway
Connects end-systems whose host protocols
have varying degrees of difference
Transport gateways make a connection
between two networks at the transport layer.
Application gateways connect two parts of an
application in the application layer, e.g.,
sending email between two machines using
different mail formats

Connect two networks above the network
layer of OSI model.

Are capable of converting data frames and
network protocols into the format needed by
another network.

Provide for translation services between
different computer protocols.
Routers versus Bridges
Addressing
Routers are explicitly addressed.
Bridges are not addressed.
Availability
Routers can handle failures in links, stations,
and other routers.
Bridges use only source and destination
MAC address, which does not guarantee
delivery of frames.
Message Size
» Routers can perform fragmentation on
packets and thus handle different packet
sizes.
» Bridges cannot do fragmentation and should
not forward a frame which is too big for the
next LAN.
 Forwarding
» Routers forward a message to a specific
destination.
» Bridges forward a message to an outgoing
network.

Priority
» Routers can treat packets according to
priorities
» Bridges treat all packets equally.
 Error Rate
» Network layers have error-checking
algorithms that examines each received
packet.
» The MAC layer provides a very low
undetected bit error rate.


Security
» Both bridges and routers provide the
ability to put “security walls” around
specific stations.
» Routers generally provide greater
security than bridges because
–they can be addressed directly and
–they use additional data for
implementing security.
Brouters: Bridging Routers
Combine features of bridges and routers.
 Capable of establishing a bridge between two
networks as well as routing some messages
from the bridge networks to other networks.
 Are sometimes called (Layer 2/3) switches and
are a combination of bridge/router hardware
and software.

Network Connectivity Devices
 Entry-level Hubs
Interconnect PCs in a single network segment
Simple stand-alone device that provides a starting point
cost-effective connectivity for many organizations.
Network Connectivity Devices
(contd.)
 Stackable Hubs
Let you start small and grow your network at your own
pace.
Are connected by flexible expansion cables, and once
stacked together, function as one hub.
Manageable as one logical unit.
Network Connectivity Devices
(contd.)
 Chassis Hub
Big iron box that can contain a variety of network
modules.
It has a power supply, a high speed backplane, and
expansion slots for plug-in Hub modules.
Network Connectivity Devices
(contd.)
 Workgroup switches
Low-end network devices that aggregate multiple
shared segments
Use switching technology
Typically deployed at the desktop level
Ethernet, Token-Ring, or ATM
Network Connectivity Devices
(contd.)
 Workgroup switches
Low-end network devices that aggregate multiple
shared segments
Use switching technology.
Typically deployed at the desktop level.
Network Connectivity Devices
(contd.)
 Backbone switches
High-end network devices deployed at the core of the
network.
Use switching technology.
Aggregate data from Hubs and Workgroup switches.
Typically accept various networking options.
Network Connectivity Devices
(contd.)
 Routers
Perform routing of packets among LANs.
Provide most effective way of segmenting the network.
Move data by finding the best path from the sender to
the receiver.
Suitable for organizations with many large LANs.