Transcript Chapter 1. Introduction to Data Communications

Chapter 5
Hardware
Layers:
Backbone
Networks
Networking
in the
Internet Age
by Alan Dennis
1
Copyright © 2002 John Wiley & Sons, Inc.
Copyright John Wiley & Sons, Inc. All rights reserved.
Reproduction or translation of this work beyond that named in
Section 117 of the United States Copyright Act without the
express written consent of the copyright owner is unlawful.
Requests for further information should be addressed to the
Permissions Department, John Wiley & Sons, Inc. Adopters of
the textbook are granted permission to make back-up copies for
their own use only, to make copies for distribution to students of
the course the textbook is used in, and to modify this material to
best suit their instructional needs. Under no circumstances can
copies be made for resale. The Publisher assumes no
responsibility for errors, omissions, or damages, caused by the
use of these programs or from the use of the information
contained herein.
2
Chapter 5. Learning Objectives
• Understand hierarchical backbones and the devices
they use
• Understand flat backbones and the devices they use
• Understand collapsed backbones and the devices
they use
• Understand VLANs and the devices they use
• Be familiar with FDDI
• Be familiar with ATM
• Understand the best practice recommendations for
backbone design
3
Chapter 5. Outline
• Introduction
• Backbone Architectures
– Hierarchical backbones, Flat backbones, Collapsed
backbones, Virtual LANs
• Fiber Distributed Data Interface
– Topology, Medium Access Control, Error Control, Message
Delineation, Data Transmission in the Physical Layer
• Asynchronous Transfer Mode
– Topology, Medium Access Control, Error Control, Message
Delineation, Data Transmission in the Physical Layer, ATMs
and LANs
• The Best Practice Backbone Design
– Architectures, Effective Data Rates, Conversion Between
Protocols, Recommendations
4
Introduction
5
Backbone Networks
• Backbone networks are high speed networks that
link an organization’s LANs and also provide
connections to other backbones, MANs, WANs
and the Internet.
• A backbone that connects backbones in several
buildings is also often called a campus network.
• A backbone is also sometimes called an enterprise
network if it connects all the networks within a
company, especially if this includes large WAN
segments.
6
Backbone Architecture Layers (Figure 5-1)
• Network designers view networks as made of
three technology layers:
– The access layer which is the technology used in
LANs.
– The distribution layer which is the part of the backbone
that connects the LANs together.
– The core layer connects different backbone networks
together, often between buildings.
• Some organizations are not large enough to have a
core layer. In such cases their backbone only spans
the distribution layer.
7
Figure 5-1 Backbone network design layers
8
Backbone Architectures
9
Backbone Network Types
• There are four basic types of backbone
networks:
• Hierarchical Backbones
• Flat Backbones
• Collapsed Backbones
• Virtual LANs
10
Hierarchical Backbones
• Figure 5-2 shows an example of a distribution
layer hierarchical backbone.
• Each LAN is a separate and isolated network,
connected by a TCP/IP gateway (usually a router)
to a shared media backbone network.
• Within the LANs messages are sent based on the
data link layer addresses.
• To move between LANs, message traffic needs to
be sent specifically to the router, which forwards
the message based on its network layer address.
11
Figure 5-2 Hierarchical backbone architecture 12
Flat Backbones
• Figure 5-3 gives an example of a distribution
layer flat backbone with a bus topology.
• With a flat backbone, LANs are connected using
bridges or layer-2 switches.
• Packets are forwarded based on their data link
layer addresses, making the entire flat backbone a
single subnet.
• Flat backbones using bridges were developed in
the mid-1980s to reduce costs, because at the time
routers were very expensive.
• Bridges have now become obsolete and are
typically replaced by layer-2 switches, which
have continued to fall in price.
13
Figure 5-3 Flat backbone architecture
14
Collapsed Backbones (Figure 5-4)
• Collapsed backbones use a star topology, usually
with a high speed switch at the center.
• Collapsed backbones can use either layer-2
switches or layer-3 routing switches.
• The two main advantages are:
– 1) each connection to the switch becomes a separate
point-to-point circuit also giving much higher
performance (from 200-600% higher)
– 2) the network has far fewer devices and so is much
simpler to manage.
• Two minor disadvantages are: 1) use more cable
and the cable runs for longer distances, 2) if the
central switch fails, the network goes down.
15
Figure 5-4 Collapsed backbone architecture
16
Rack-Mounted Collapsed Backbones
• Rack-mounted backbones collapse the
backbone into a single room, called a main
distribution facility (MDF) where
networking equipment is connected and
mounted on equipment racks (Figure 5-5).
• Devices are connected using short patch
cables.
• Moving computers between LANs is
relatively simple since equipment is all in
the same location.
17
Fig. 5-5 Rack-mounted collapsed backbone architecture
18
Chassis-based Collapsed Backbones
• Chassis switch designs include a number of
open slots and have an internal capacity
capable of supporting all active modules.
• A variety of modules (i.e., card-mounted
networking devices) can be inserted into
these slots providing a high level flexibility
in network configuration.
• By turning the backbone into an internal
bus, chassis switches also can provide very
high performance speeds capable of
aggregate data rates in the Gbps range.
19
Figure 5-7
Central
Parking’s
collapsed
backbone
20
Virtual LANs
• VLANs are a new type of LAN/BN architecture
using intelligent, high-speed switches.
• Unlike other LAN types, which physically connect
computers to LAN segments, VLANs assign
computers to LAN segments by software.
• VLANs have been standardized as IEEE802.1q
and IEEE802.1p.
• The two basic designs are:
– Single-switch VLANs
– Multiswitch VLANs
21
Single Switch VLANs (Figure 5-8)
• With single switch VLANs, computers are assigned
to VLANs using special software, but physically
connected together using a large physical switch.
• Computers can be assigned to VLANs in four ways:
– Port-based VLANs assign computers according to the
VLAN switch port to which they are attached
– MAC-based VLANs assign computers according to each
computer’s data link layer address
– IP-based VLANs assign computers using their IP-address
– Application-based VLANs assign computers depending
on the application that the computer typically uses. This
has the advantage of allowing precise allocation of
network capacity.
22
Figure 5-8 Single-switch VLAN architecture
23
Multiswitch VLANs (Figure 5-9)
• Multiswitch VLANs send packets between multiple
switches, making VLANs with segments in separate
locations possible.
• When a frame is sent between switches it is modified
and to include a tag field carrying VLAN information
field. When the frame reaches the final switch, the tag
field is removed prior to the frame being sent to its
destination computer.
• Multiswitch VLANs can also prioritize traffic using
the IEEE802.1p standard in the hardware layers and
the RSVP standard in the internetwork layers.
• IEEE802.1p works with the IEEE802.11ac frame
definition which includes a special priority field.
24
Figure 5-9 Multiswitch VLAN architecture
25
Figure 5-10 IONA VLAN
26
Fiber Distributed Data Interface
27
Fiber Distributed Data Interface (FDDI)
• FDDI (standardized as ANSI X3T9.5) backbone
protocol was developed in the 1980s and popular
during the 80s and 90s.
• FDDI operates at 100 Mbps over a fiber optic
cable.
• Copper Distributed Data Interface (CDDI) is a
related protocol using cat 5 twisted wire pairs.
• FDDI’s future looks limited, as it is now losing
market share to Gigabit Ethernet and ATM.
28
FDDI Topology (Figure 5-11)
• FDDI uses both a physical and logical ring
topology capable of attaching a maximum
of 1000 stations over a maximum path of
200 km. A repeater is need every 2 km.
• FDDI uses dual counter-rotating rings
(called the primary and secondary). Data
normally travels on the primary ring.
• Stations can be attached to the primary ring
as single attachment stations (SAS) or both
rings as dual attachment stations (DAS).
29
Figure 5-11 Optical cable topology for an FDDI
local area network.
30
FDDI’s Self Healing Rings
• An important feature of FDDI is its ability to
handle a breaks in the network by forming a single
temporary ring out of the pieces of the primary
and secondary rings.
• Figure 5-12 shows an example of a cable break
between two dual attachment stations.
• Once the stations detect the break, traffic is
rerouted through a new ring formed out of the
parts of the primary and secondary rings not
affected by the break.
• The network then operates over this temporary
ring until the break can be repaired.
31
Figure 5-12 Managing a broken circuit
32
FDDI Media Access Control
• FDDI uses a token passing system. Computers wanting to
send packets wait to receive a token before transmitting.
• Multiple packets can be attached to the token as it moves
around the network.
• When a station receives the token, it looks for attached
packets addressed to it and removes them from the incoming
packet.
• If the station wants to send a packet it attaches it to the token
and sends the token with its attached packets to the next
station.
• This controlled access technique provides a higher
performance level at high traffic levels compared to a
contention-based technique like Ethernet.
33
FDDI Message Delineation (Fig. 5-13)
• The FDDI frame can be broken into three parts:
• Frame Start: like Ethernet, the frame begins with a
preamble (8-bytes in this case) and a 1-byte start
delimiter.
• Frame Body: the main body of the frame includes
the following fields:
– 1-byte frame control field (used for the token)
– 2 or 6 byte fields for the destination and source addresses
(6 bytes is more common)
– the data field contains 0-4500 bytes of data
– the frame check sequence (FCS) used in error control.
• Frame End: the frame ends with a 1-byte end
delimiter and a 2-byte frame status field.
34
Figure 5-13 FDDI frame layout
35
Asynchronous Transfer Mode
36
Asynchronous Transfer Mode (ATM)
• Asynchronous Transfer Mode (ATM) (also called
cell relay) was originally designed to carry both
voice and data traffic over WANs. It is also used in
backbone networks.
• In the WAN, ATM almost always uses SONET as
its hardware layer. In backbones, ATM is often
implemented as a standalone protocol.
• On order to interconnect with the TCP/IP world,
an ATM gateway is used that converts TCP/IP and
Ethernet frames into ATM cells and then converts
them back once they have reached their
destination network.
37
ATM Topology
• ATM uses a mesh topology (see Figure 5-14)
• This mesh is made up of point-to-point, full
duplex circuits that interconnect ATM switches.
• ATM circuits typically operate at 155 Mbps in
each direction, although higher speeds, esp. 622
Mbps (1.24 Gbps total) is also possible.
• Although originally designed to run on optical
fiber, some versions of ATM can run on cat-5e
twisted pair cables.
38
Figure 5-14 ATM mesh architecture
39
ATM Media Access Control
• ATM uses full-duplex circuits, so media access
control is less of an issue.
• To handle circuit congestion, ATM prioritizes
transmissions based on Quality of Service
(QoS). Priorities are based on 5 ATM service
class definitions.
• Real time applications, such as voice, get a
high priority, since it can not allow delays.
• E-mail gets a lower priority, since small delays
don’t matter very much.
40
ATM Addressing (Figure 5-15)
• ATM addressing uses virtual channels (VCs).
• Each cell’s VC has two parts: a virtual path
identifier and a virtual channel identifier.
• Virtual channels are also assigned a service class
when they are created.
• When a cell reaches an ATM switch, the switch
looks up the VC number in its VC table to
determine where to send it next (similar to how a
routing table works).
41
Figure 5-15 Addressing and forwarding
with ATM virtual circuits
42
ATM virtual circuits
• ATM is connection-oriented: all packets
travel in order in the same virtual channel.
• VCs can be set up in one of two ways:
– Permanent Virtual Circuits (PVCs) – permanent
virtual circuits set up for long periods.
– Switched Virtual Circuits (SVCs) - temporary
virtual circuits set up for one transmission and
deleted when the transmission is completed.
43
ATM Error Control
• ATM’s error control technique is called
throw-it-on-the-floor.
• Error checking is only done on the ATM
header.
• If an error is detected, the cell is discarded.
• Full error control, including requests for
retransmission are handled at the source and
destination computers (on a LAN this is
typically done using TCP).
44
ATM Message Delineation (Fig. 5-16)
• ATM has a 53-byte frame called a cell.
• The ATM header includes these fields:
• Generic Flow Control: controls the flow of data
across the circuit
• Virtual Path Identifier: identifies the group of
channels the data is moving with.
• Virtual Circuit Identifier: identifies the specific
channel.
• Payload Type: indicates type of data in data field
• Cell Loss Priority: whether or not the cell is
discarded if the circuit gets busy.
• Header Error Control: uses CRC-8 for error control
but only on the header portion of the field.
45
Figure 5-16 ATM cell layout
46
ATM and LANs
• Ethernet and TCP/IP use large variable length
frames/packets with fixed addresses while ATM
uses small fixed length cells addressed using
virtual channels.
• For that reason, Ethernet and TCP/IP must first be
translated before being sent over ATM networks.
• Two approaches for this are:
– LAN Encapsulation (LANE), which splits frames into
48 byte pieces, reassembling them when they reach
their destination LAN.
– Multiprotocol Over ATM (MPOA) is an extension of
LANE that uses both IP and Ethernet addresses.
47
LAN Encapsulation (LANE)
• LANE works by breaking Ethernet frames into 48byte chunks. This occurs in the LAN’s gateway
ATM edge switch (see Figure 5-17).
• The edge switch also creates a virtual channel
identifier for the cells to use.
• The cells are then sent over the ATM backbone
using this virtual channel identifier.
• When they reach the destination edge switch, the
frame is reassembled.
• LANE’s high overhead creates significant delays,
lowering network performance as a consequence.
48
Figure 5-17 ATM in the backbone
49
The Best Practice Backbone
Design
50
Current Backbone Technology Trends
• The following trends in backbone technologies have
been taking place in recent years:
• Organizations are moving to collapsed backbones or
VLANs.
• Gigabit Ethernet use is growing.
• FDDI seems to be on its way out.
• ATM, while still popular in WANs, is losing ground
to Gigabit Ethernet as a backbone technology.
• Taken together, it appears that Ethernet use will
dominate both the LAN and backbone environments.
51
Technology
Effective Data Rate
Full Duplex 1 GbE
1.8 Gbps
Full Duplex 10 GbE
18 Gbps
FDDI
7-70 Mbps depending on
traffic
ATM (155 Mbps, Full
Duplex)
160 Mbps
ATM (622 Mbps, Full
Duplex)
760 Mbps
Assumes: collapsed backbone connecting Ethernet LANs transmitting mostly large frames
Figure 5-19 Effective data rates
for backbone technologies
52
Backbone Recommendations (Fig. 5-20)
• The best practices are recommended for backbones:
– 1. Architecture: collapsed backbone or VLAN.
– 2. Technology: gigabit Ethernet. ATM and FDDI use has
started to fall off over the past year.
– 3. The ideal network design combines use of layer-2 and
layer-3 Ethernet switches.
– 4. The access layer (LANs) uses 10/100 layer-2 switches
using cat 5e or cat 6 twisted pair cables (cat 6 is needed
for 1000BaseT).
– 5. The distribution layer uses layer-3 Ethernet switches
that use 1000BaseT or fiber, Cat 6 or Cat 7 TP.
– 6. The core layer uses layer-3 Ethernet switches running
10GbE or 40GbE over fiber.
– 7. Network reliability is increased using redundant
switches and cabling.
53
Figure 5-20 The best practice network design
54
End of Chapter 5
55