Business Data Communications

Chapter Six

Backbone and Metropolitan Area

Network Fundamentals

Primary Learning Objectives

Define backbone and metropolitan area networks Differentiate between horizontal and vertical

network segments Recognize the function of a backbone network

protocol Understand the advantages and disadvantages of

distributed and collapsed backbones Explain the concept of backbone fault tolerance List examples of backbone design considerations Identify common backbone problems Describe Switched Multimegabit Data Services

Backbone and Metropolitan Area Networks

A backbone network (BN): Connects other networks of an organization

Networks connected are typically LANs Generally spans a building or campus Has its own address

A metropolitan area network (MAN): Is often used to connects BNs Generally spans a city Is sometimes viewed as a citywide backbone Has its own address

The distinction between BNs and MANs is blurring

Horizontal and Vertical Networks

Each individual LAN is called a network segment

A network segment may be horizontal or vertical

Horizontal networks are configured on a single floor

Vertical networks are configured on multiple floors

Whether horizontal or vertical, network segments are usually connected to each other by means of a backbone network

A Horizontal Network

A Vertical Network

Connecting to a backbone

Backbone NetworkProtocols

BNs most often support high traffic demand for connected LANs

Therefore a BN generally uses a protocol that provides higher throughput than the protocol used by the connected LANs

Gigabit Ethernet: Is a common protocol choice for BNs In its initial form was labeled 802.3z by the IEEE Supports 1 Gbps (billion bits per second) Does not change the underlying Ethernet format Is both half- and full-duplex capable

ATM and Frame Relay are also possible BN protocols

Distributed and Collapsed Backbones

In addition to having a protocol, BNs have an architecture

Two common types of BN architecture are: Distributed Collapsed

Factors that influence the BN architecture choice include:

Business need Condition of the facility or physical plant The way that users need to communicate Budget Placement of networked devices

Distributed Backbones

“Distributed” implies “in more than one location” A Distributed Backbone:

Runs throughout the entire facility Uses a central cable Requires its own protocol Is its own network Is usually connected to network segments, LANs,

by switches and/or routers Can have directly connected devices that are part

of the BN

Distributed Backbones

Each router will have two network addresses

Distributed Backbones

When configured with multiple routers, may need to pass traffic through several routers for that traffic to reach its final destination

Going through several routers can result in traffic delay

Internetwork traffic can be expected to increase when more routers are used

Require that each network segment have its own cabling and connecting device to the distributed backbone, adding to expense

Distributed Backbones

Are generally more complex than collapsed backbones, resulting in more complicated:

Security Maintenance Monitoring

Allow the placing of commonly needed networked devices or resources directly onto the backbone

May be the only viable solution for a business, depending on the facility and layout of network segments

Distributed Backbone with a Directly Connected DB

Server

Collapsed Backbones

Use a single central device, namely a router or switch, to which network segments are connected

This central router or switch is in essence the backbone

Connect network segments to the collapsed backbone, using other hubs, switches, or routers

Generally reduce cabling needs, however: Connected devices must be able to support cable

segment lengths that span the distance to the collapsed backbone

Legacy networks using lower grade UTP may not be collapsed-backbone compatible

Collapsed Backbones

Do not require a protocol different from that of connected network segments

Having one protocol can make network administration easier

However, depending on how the backbone needs to be used relative to traffic demands, having one protocol may not be an advantage

Utilize a “backplane”--a high-speed communications bus--in the switch or router

Collapsed Backbones

Use fiber optic cabling to connect network segments to a collapsed backbone’s backplane

Fiber allows network segments to be widely scattered across a building or campus

Might not allow legacy Ethernet networks to utilize the collapsed backbone architecture

Pass internetwork traffic through only one device Centralize security, monitoring, and

maintenance Can achieve significant cost savings

Collapsed Backbones

Internetwork traffic passes through only one connecting device, in this example a switch, to its ultimate destination

Distributed versus Collapsed

Backbone Fault Tolerance

“Fault tolerance” is the capability of a technology to recover in the event of an error, failure, or some other unexpected event

Backbones, because they connect and provide communication to various segments of an enterprise, must be fault tolerant

Resource redundancy is a common means of providing fault tolerance

A Simplified Redundant Backbone

Backbone Fault Tolerance

Redundant backbones also allow for traffic load balancing

By permitting placement of half of all network segments onto one or the other backbone, duplicate backbones allow internetworking traffic to be shared or balanced

Duplicating the backbone can be done in whole or in part, based on cost and need

Documentation is also part of recovery procedures

Backbone Design Considerations

Internal versus external wiring Identification and labeling of all backbone

networks, devices, wiring Knowledge of collision domain boundaries

An enterprise with a mix of Ethernet networks (such as Standard, Fast, Gigabit) might have a mix of collision domains

Wiring closets Data centers

Backbone Design Considerations

Backbone Design Considerations

Wiring Closet Design

A few Good Tools

Common Gigabit Backbone Problems

Packet errors Early collision Late collision Runts Giants and jabbering Broadcast storms

Common Gigabit Backbone Problems

Cable errors: Near-end cross talk Attenuation Impedance Attenuation to cross talk Capacitance Cable length

Common Gigabit Backbone Problems

Network Interface Card errors: Improper configuration – diagnostic software Physical failure

Connectivity testing with ping: Based on ICMP Various vendor implementations Echo_Request Echo_Reply Time-to-live

Common Gigabit Backbone Problems

Switched Multimegabit Data Services - SMDS

Were designed specifically for MANs Support exchanging data between:

LANs in different parts of a city Network segments over a large campus

Provide packet-switched datagram delivery Are associated with a common carrier’s SMDS

network Require subscribers to pay only when they use

the common carrier’s network

In Summary

The distinction between BNs and MANs is blurring

BNs are critical in connecting the various network segments of an organization

BNs have both a protocol and architecture: Gigabit Ethernet is a popular BN protocol Two BN architectures are distributed or collapsed

Fault tolerance is important for the backbone Common Gigabit backbone problems include

packet, cable, and NIC errors SMDS is associated particularly with MANs