Chapter 11:Approaches to Networking

Business Data Communications, 4e

LANs, WANs, and MANs Ownership

WANs can be either public or private LANs are usually privately owned

Capacity LANs are usually higher capacity, to carry greater internal

communications load Coverage

LANs are typically limited to a single location WANs interconnect locations MANs occupy a middle ground

Comparison of Networking Options

Types of WANs

Circuit-switched Packet-switched

Circuit-Switching

Definition: Communication in which a dedicated communications path is established between two devices through one or more intermediate switching nodes

Dominant in both voice and data communications today e.g. PSTN is a circuit-switched network

Relatively inefficient (100% dedication even without 100% utilization)

Circuit-Switching Stages

Circuit establishment Transfer of information

point-to-point from endpoints to node internal switching/multiplexing among nodes

Circuit disconnect

Circuit Establishment Station requests connection from node Node determines best route, sends message to next

link Each subsequent node continues the establishment

of a path Once nodes have established connection, test

message is sent to determine if receiver is ready/able to accept message

Information Transfer

Point-to-point transfer from source to node Internal switching and multiplexed transfer

from node to node Point-to-point transfer from node to receiver Usually a full-duplex connection throughout

Circuit Disconnect

When transfer is complete, one station initiates termination

Signals must be propagated to all nodes used in transit in order to free up resources

Public Switched Telephone Network (PSTN)

Subscribers Local loop

Connects subscriber to local telco exchange

Exchanges Telco switching centers Also known as end office >19,000 in US

Trunks Connections between

exchanges Carry multiple voice

circuits using FDM or synchronous TDM

Managed by IXCs (inter-exchange carriers)

Digital Circuit-Switching Node

Circuit Switching Node:Digital Switch

Provides transparent signal path between any pair of attached devices

Typically full-duplex

Circuit-Switching Node:Network Interface

Provides hardware and functions to connect digital devices to switch

Analog devices can be connected if interface includes CODEC functions

Typically full-duplex

Circuit-Switching Node:Control Unit

Establishes on-demand connections Maintains connection while needed Breaks down connection on completion

Blocking/Nonblocking Networks

Blocking: network is unable to connect two stations because all possible paths are already in use

Nonblocking: permits all possible connection requests because any two stations can be connected

Switching Techniques

Space-Division Switching Developed for analog

environment, but has been carried over into digital communication

Requires separate physical paths for each signal connection

Uses metallic or semiconductor “gates”

Time-Division Switching Used in digital transmission Utilizes multiplexing to place

all signals onto a common transmission path

Bus must have higher data rate than individual I/O lines

Routing in Circuit-Switched Networks

Requires balancing efficiency and resiliency Traditional circuit-switched model is

hierarchical, sometimes supplemented with peer-to-peer trunks

Newer circuit-switched networks are dynamically routed: all nodes are peer-to-peer, making routing more complex

Alternate Routing

Possible routes between two end offices are predefined

Originating switch selects the best route for each call

Routing paths can be fixed (1 route) or dynamic (multiple routes, selected based on current and historical traffic)

Control Signaling Manage the establishment, maintenance, and

termination of signal paths Includes signaling from subscriber to network, and

signals within network In-channel signaling uses the same channel for

control signals and calls Common-channel signaling uses independent

channels for controls (SS7)

ISDN 1st generation: narrowband ISDN

Basic Rate Interface (BRI) two 64Kbps bearer channels + 16Kbps data channel (2B+D)

= 144 Kbps circuit-switched

2nd generation: broadband ISDN (B-ISDN) Primary Rate Interface (PRI) twenty-three 64Kbps bearer channels + 64 data channel

(23B+D) = 1.536 Mbps packet-switched network development effort led to ATM/cell relay

Past Criticism of ISDN

“Innovations Subscribers Don’t Need” , “It Still Doesn’t Network” , “It Still Does Nothing”

Why so much criticism? overhyping of services before delivery high price of equipment delay in implementing infrastructure incompatibility between providers' equipment.

Didn’t live up to early promises

ISDN Principles Support of voice and nonvoice using limited set of

standard facilities Support for switched and nonswitched applications Reliance on 64kbps connections Intelligence in the networks Layered protocol architecture (can be mapped onto OSI

model) Variety of configurations

ISDN User Interface

“Pipe” to user’s premises has fixed capacity Standard physical interface can be used for

voice, data, etc Use of the pipe can be a variable mix of voice

and data, up to the capacity User can be charged based on use rather than

time

ISDN Network Architecture

Physical path from user to office subscriber loop, aka local loop full-duplex primarily twisted pair, but fiber use growing

Central office connecting subscriber loops B channels: 64kbps D channels: 16 or 64kbps H channels: 384, 1536, or 1920 kbps

ISDN B Channel

Basic user channel (aka “bearer channel”) Can carry digital voice, data, or mixture

Mixed data must have same destination Four kinds of connections possible

Circuit-switched Packet-switched Frame mode Semipermanent

ISDN D Channel

Carries signaling information using common-channel signaling call management billing data

Allows B channels to be used more efficiently Can be used for packet switching

ISDN H Channel

Only available over primary interface High speed rates Used in ATM

ISDN Basic Access

Basic Rate Interface (BRI) Two full-duplex 64kbps B channels One full-duplex 16kbps D channel Framing, synchronization, and overhead bring total

data rate to 192kbps Can be supported by existing twisted pair local

loops 2B+D most common, but 1B+D available

ISDN Primary Access

Primary Rate Interface (PRI) Used when greater capacity required No international agreement on rates

US, Canada, Japan: 1.544mbps (= to T1) Europe: 2.048mbps

Typically 23 64kbps B + 1 64kbps D Fractional use of nB+D possible Can be used to support H channels

Packet-Switching Networks

Includes X.25, ISDN, ATM and frame-relay technologies

Data is broken into packets, each of which can be routed separately

Advantages: better line efficiency, signals can always be routed, prioritization option

Disadvantages: transmission delay in nodes, variable delays can cause jitter, extra overhead for packet addresses

Packet-Switching Techniques

Datagram each packet treated independently and referred to as a

datagram packets may take different routes, arrive out of sequence

Virtual Circuit preplanned route established for all packets similar to circuit switching, but the circuit is not dedicated

Packet-Switched Routing Adaptive routing changes based on network

conditions Factors influencing routing are failure and

congestion Nodes must exchange information on network status Tradeoff between quality and amount of overhead

Packet-Switched Congestion Control

When line utilization is >80%, queue length grows too quickly

Congestion control limits queue length to avoid througput problems

Status information exchanged among nodes Control signals regulate data flow using

interface protocols (usually X.25)

X.25 Interface Standard

ITU-T standard for interface between host and packet-switched network

Physical level handles physical connection between host and link to the node Technically X.21, but other standards can be substituted,

including RS-232 Link level provides for reliable data transfer

Uses LAPB, which is a subset of HDLC Packet level provides virtual circuits between

subscribers

Virtual-Circuit Service

External virtual circuit: logical connection between two stations on the network

Internal virtual circuit: specific preplanned route through the network

X.25 usually has a 1:1 relationship between external and internal circuits

In some cases, X.25 can be implemented as a packet-switched network

WANs for Voice

Requires very small and nonvariable delays for natural conversation--difficult to provide this with packet-switching

As a result, the preferred method for voice transmission is circuit-switching

Most businesses use public telephone networks, but a few organizations have implemented private voice networks

WANs for Data

Public packet-switched networks (X.25) Private packet-switched networks Leased lines between sites (non-switched) Public circuit-switched networks Private circuit-switched networks (interconnected

digital PBXs) ISDN (integrated X.25 and traditional circuit-

switching)

WAN Considerations

Nature of traffic stream generally works best with dedicated circuits bursty better suited to packet-switching

Strategic and growth control--limited with public networks

Reliability--greater with packet-switching Security--greater with private networks