. Study questions.
What are the three layers of the hierarchical network design model
Access, distribution and core
What are the advantages of designing LANs using the hierarchical model?
The hierarchical model
separates out different functions of a network,
modular design. This should make the network easier to manage and
troubleshoot. It should be easier to expand the network (scalability), and it
should be easier to maximise performance.
What is the
purpose of the access layer?
It allows end devices to connect to the network
and controls which devices
What sort of devices are found at the access layer?
End devices, such as PCs, printers, and IP phones. Network devices such as
switches, bridges, hubs, and wireless access points.
What is the purpose of the distribution layer?
It controls the flow of network traffic. In particular, it controls traffic between
different broadcast domains (subnetworks, VLANs)
. It aggregates traffi
the access layer that needs to be passed to the core layer for longer distance
What sort of devices are found at the distribution layer?
performance switches. They should have high availability and
redundancy to ensure reliabili
What is the purpose of the core layer?
It is the high speed backbone of the network. Its main task is to forward large
amounts of data quickly.
What is a collapsed core model and where might it be used?
The distribution layer and the core layer are
combined. It is used for smaller
What is a wiring closet?
A room designed to hold network devices such as switches and routers. It is a
central point where network cabling comes together.
Why is it not easy to see the logical hierarchical desi
gn of a network when
looking at the network layout in a building?
The devices operating at different layers may not be physically separated. For
example, switches operating at different layers may be kept in the same
Why is redundancy important
in a network?
It gives better availability and reliability. If one switch or link goes down then
an alternative path can be used.
Which layers normally have redundancy built in?
Core and distribution layers.
How can the hierarchical design help to give
Traffic (and particularly long distance traffic) is
forwarded through distribution
layer and core layer switches that are designed to work at high speeds, and
not through a series of lower
speed access layer switches.
How can switches
at the different layers contribute to network security?
Access layer switches can restrict the devices that are permitted to connect to
Distribution layer devices can be configured with access control
policies that restrict traffic according
to IP addresses or application layer
protocols. (The core layer is optimised for speed. Security measures are not
appropriate at this layer because they slow things down.)
How does a hierarchical design help to make a network manageable?
Switches at a gi
have similar functions and therefore are likely
have similar configurations
. This makes it easier to check configurations and
Should the same type of switch be used at each layer of the hierarchical
distribution and core layers need very fast switches. The access layer
can use less expensive switches.
” in hierarchical network design
Network diameter is the number of devices that a packet has to cross before it
What is network device latency?
Network device latency is the time spent by a device as it processes a packet
What sort of processing does a switch have to do on each packet?
The switch has to look in the frame header and find t
he destination MAC
address. It then looks this address up in its MAC address table and finds the
matching exit port. It then forwards the frame out of the port.
How can bandwidth aggregation be implemented?
Switches can be linked by more than one physica
l link. These physical links
can be combined into one logical link that makes use of the combined
Why is redundancy not normally provided at the access layer?
It would be too expensive. The advantage of improved reliability would not
he extra cost.
When designing a new network, at which layer would you start?
At the access layer. Make sure that all end devices will be able to connect to
What is a converged network?
A network that carries voice, video and data.
ctors have slowed the move towards converged networks?
that was originally
affordable only by big companies
need Quality of Service management to give voice and
video traffic priority
expertise to set
up and manage
Where firms have existing separa
te networks that work well, the firms
reluctant to make a change.
What are the benefits of a converged network?
In a new buil
d, there is only one set of cabling to install instead of two or three.
There is only one network to manage instead of two or three. This may mean
having one set of people to run the network instead of several sets of people
so the company may employ fewe
When changes are needed, there is only one network to change.
New communications options are possible.
A PC can be used to make phone
calls or for videoconferencing by adding appropriate software and peripherals.
There is no need for a separate IP
phone or videoconferencing equipment.
Why was videoconferencing equipment originally kept on a separate network
from data or phones, and what made it possible to combine videoconferencing
with other networks?
Videoconferencing requires large amounts of
and is sensitive to
delays. It was kept separate to avoid conflict with data traffic. Developments in
network design, using the hierarchical model and quality of service policies,
have allowed videoconferencing to share a network with voice and
without significant reduction of quality.
What is traffic flow analysis, and why might you want to do it?
Traffic flow analysis is the measurement of bandwidth usage.
It can show
what demands different sources of traffic are making on the netwo
might do it to get the information you need to improve network performance,
plan network changes
, decide if you need to upgrade,
and decide what type of
new equipment is needed.
How can measurements be made in order to carry out a traffic flow ana
It can be done manually by monitoring switch ports, but this is time consuming
and not feasible for large networks. There are software analysis tools that take
the measurements and perform the analysis, giving a visual display
How can a
help with network design?
Users are normally grouped by job function because they have similar
requirements for bandwidth, applications, access to servers etc. The User
Communities Analysis shows what demands the different groups of users
make on the network. Some design implications are:
The number of switch ports needed for each group of users, including
The location of servers that the users access.
The bandwidth requirements of the users.
Which two types of traffic n
eed to be considered when planning data storage
server traffic and server
How can server
server traffic be optimised?
Servers and data stores should be kept physically close to each other and
should be connected by hig
h performance switches. Businesses often keep
their servers and data stores in a secure data centre.
How might you deal with a bottleneck where there is insufficient bandwidth?
Use link aggregation (multiple links between switches) or install higher
Why is it important to produce a topology diagram as part of the network
It is difficult to produce a topology diagram later by examining the network
itself, and the diagram is necessary for the efficient management of t
network. It shows all the switches and how they are connected, including
aggregated links and redundant links, and the ports used. It shows how
servers, storage devices and end user devices such as workstations are
How is the thickness of a
switch (top to bottom) measured?
In rack units (U).
What is the difference between a fixed configuration switch and a modular
A fixed configuration switch has a fixed number of ports and these cannot be
changed. A modular switch has a chassis th
at can take line cards. The line
cards contain the ports. A modular switch can therefore have different
numbers or types of ports, depending on the choice of line cards.
What are stackable switches?
Stackable switches are designed to be connected togethe
r and act as a single
The connection between the switches is through a special high speed
port and not through the normal switch ports.
What is the port density of a switch?
The number of available access ports.
If 80 outlets need to be connecte
d to a switch, why is a single large modular
switch likely to be a better choice than four fixed configuration switches with
24 ports each?
The four switches each need their own power supply, and they use ports to
connect to each other. They may need seve
ral ports for each link if
aggregation is needed. The single large switch needs a single power supply
No ports are needed for internal links, and the switch’s internal backplane
provides a higher speed connection than links between switches.
What is wire
speed, and what is forwarding rate?
Wire speed is the data rate that a switch port
A port may be
designed to operate at 100 Mb/s Fast Ethernet or 1000 Mb/s Gigabit Ethernet.
The forwarding rate is the amount of data that the switch can proces
second. If the forwarding rate is less than the total wire speed of all the ports
then it is not possible for all the ports to operate simultaneously at their wire
What is Etherchannel?
A system that uses link aggregation to provide high ban
switches. Up to 8 ports
can be bound together so that they act as a single link.
What is PoE and what are its advantage and disadvantage?
Power over Ethernet is a feature that lets a switch provide a power supply to a
device such as an IP
phone or a wireless access point. The power is supplied
over the Ethernet cable. The advantage is that the devices do not need a
separate power supply, so it may be possible to position them more flexibly.
The disadvantage is that the feature makes the swi
tch much more expensive.
At which OSI layer to typical traditional switches operate?
What is a multilayer switch?
A switch that operates at layer 3 as well as at layer 2. For example, it can
forward packets based on IP addressing, not just on
What features are required for access layer switches?
Port security, VLANs, Fast Ethernet/Gigabit Ethernet, PoE, and link
Quality of Service is needed for converged networks.
What features are required for distribution l
High forwarding rate to allow all ports to operate to their full bandwidth
High bandwidth links
and link aggregation
Redundancy to ensure reliability and availability
swappable power supplies
Quality of servi
ce to maintain priority for video and voice traffic
VLAN routing (layer 3 function)
Access control lists for security (layer 3 function)
What features are required for core layer switches?
Very high forwarding rates as this is the high speed backbo
ne of the network.
High bandwidth ports, preferably 10Gbps. Support for link aggregation.
Redundancy to give reliability and availability.
Redundant power supplies.
swappable hardware components to avoid downtime for maintenance.
port for quality of service. (Particularly important where there are links to
lower bandwidth WANs.)
Layer 3 features.
Which Cisco Catalyst switch does not provide a command line interface?
Catalyst Express 500
Which Catalyst switch is suitable for acce
ss level use in small organisations,
provides up to 48 ports, but does not provide PoE?
What does the Catalyst 3560E switch provide that the 3560 switch does not?
10 Gbps connectivity.
Which catalyst switch has forwarding rates up to 720
Which type of
catalyst switch can be stacked s
o that up to 9 switches can
operate as a single logical switch?
What is special about the
Catalyst 4900 series switches
They are designed for use in data centres
ing servers and data stores