Running head: APA WRITING STYLE AND MECHANICS - Transtutors

blackstartNetworking and Communications

Oct 26, 2013 (3 years and 7 months ago)

68 views










WAN Paper

(Stephen Craig
)

(IT230
)


(
9

June

2008
)











Introduction
:



Wan is a geographically dispersed telecommunications network such as state, province or
country. It distinguishes a broader telecommunication structure from a local area netw
ork (LAN)
and connects multiple smaller networks such as LAN or a MAN (metro area network). A wide
area network may be privately owned or rented, but the term usually connotes the inclusion of
public (shared user) networks. With a WAN in place, employees a
t the LANs in any of your
offices, regardless of how geographically distant they may be still they can freely exchange files
and data with each other as if they were physically connected to one network.



Hardware:


DSL
(Digital Subscriber Line) is a techn
ology that

delivers reliable, high
-
speed office
-
to
-
office
connectivity over traditional copper wires and is available in most regions from ISPs, local phone
companies or alternative exchange carriers. DSL is more expensive than 56 Kbps or ISDN
access; it i
s affordable to most small businesses.

DSL is an excellent connectivity choice for WANs as it delivers outstanding performance over
existing telephone lines. The most common and cost
-
effective way for small businesses to build
DSL
-
powered WANs is over pub
lic networks, such as the Internet or the public telephone system
and it is always on, or open, connection. To use DSL you require installing a DSL router at each
of your offices and establishing an enterprise
-
wide account with an ISP. Employees will have
instant and easy access to the WAN.



Access Routers:

Access routers are useful to connect two or more LANs together. Each LAN
requires a device called an access router that connects to a switch or hub on the local network
and serves as a gateway to the WAN
. The access routers establish LAN
-
to
-
LAN connections and
forward network traffic between users at the remote sites.


Wan Connection Types


A wide area network (WAN) connection type compromises of different components such as
cables, switches, routers etc.

All these components help the internet service provider to forward
data from one subscriber’s location to another. There are generally three types of WAN
connections namely:




Leased Lines



Packet Switching Connections



Circuit Switched Connections.



Leased

Lines:

A
Leased line is a dedicated connection which is typically used when
organization needs connection that is always available. It is not shared with other subscribers.
Several encapsulation methods used on leased lines are PPP, HDLC and SLIP. Leased

lines
connection is also considered as Point
-
to
-
point connection.



Packet Switched Connections:
Packet switched connections provides end
-
to
-
end connection
that allows multiple subscribers. It lets multiple subscribers to use same wire at any given time. I
t
performs slightly different role than leased line as it maintains a virtual point
-
to
-
point or point
-
to
-
multipoint connection by using different virtual circuits for each subscribers.


Circuit Switched Connections:
Circuit switched connections are used o
n a per connection basis
and are normally implemented in a dial
-
up modem or ISDN environment. In this type of
connection, service is only made available when data needs to be transmitted. Encapsulation used
is normally PPP, SLIP or HDLC.


Wan Design Goal:


The goal of designing a wan is to interconnect two offices which provide high level security and
continuous availability of network connection between offices. The design should be such that
any changes to be made in future should be easily implemented.


Wan Design Strategies:


Frame Relay: Frame relay is a layer 2 protocol used in Wide Area Networking (WAN). It uses
telecommunication provider’s packet switching infrastructure to transfer data. Frame relay can
provide speed from 56kbps to 43 Mbps dependin
g on the service provider’s network capabilities.



Connection to frame relay network is done by attaching a point
-
to
-
point link from customer’s
DTE to providers DCE. This connection lets the customer to connect to the provider’s frame
cloud. After the fra
me cloud is in place at two or more customer’s site PVC’s can be set up to
allow communication between the sites.


Advantages of Frame Relay


The most important advantage of Frame Relay over point
-
to
-
point leased line is cost. Frame relay
can provide per
formance similar to leased line but cuts down significant amount of cost if the
distance between the computers on network is long. The customer has to make a point
-
to
-
point
connection to provider’s nearest frame switch to get on the network.


In the scena
rio where we have to design a WAN for office, Frame Relay is a good choice since
the office can be located in any part of country and still have a WAN connection at low cost.


Disadvantages of Frame Relay


The main disadvantages of Frame Relay are:


1)

Slow
down due to network congestion

2)

Quality of Service


As all of provider’s frame relay customer use common network, there is always a high

probability of high data transmission which exceeds the network’s capacity. The other
disadvantage is ensuring quality o
f service as frame relay uses variable length packet it’s always
difficult to ensure quality of service compared to fixed length packets.


To overcome potential congestion, it should be ensured that the provider’s Committed
Information Rate (CIR) is suffi
cient to meet the needs of their network’s requirement. CIR is
minimum level that the provider guarantees to deliver at least this amount of data transmission
even during heavy network load.


Before making any final decision we should first take a look at

the alternative WAN protocol.


X.25
is older technology similar to Frame Relay but it’s not as efficient as Frame relay is. X.25
was designed when most telecommunications lines were analog and therefore X.25 loses a big
percentage of throughputs to error
checking overhead compared to Frame Relay. X.25 has no
advantages over Frame relay and should not be used unless there are no other alternatives
available.


ATM:
Difference between ATM and Frame Relay is that ATM uses fixed
-
length packet whereas
Frame Rela
y uses variable
-
length packets. ATM assures quality of service as it uses fixed
-
length
packets compared to Frame Relay. ATM is a very good option to choose only if the office can
afford higher price.


After considering all the WAN protocols, it’s recommend
ed to use Frame Relay for the offices in

consideration for designing a WAN. As we have examined all other alternatives and decided that
Frame Relay offers the best combination of performance and price for our offices’ WAN needs
and so the next step should
be to design the network to efficiently connect the office sites.


Topology:


We would be connecting two offices, and in future there could be a requirement to add more
offices and thus here, we choose mesh topology to provide some redundancy in communicat
ion
links. This arrangement can be either full mesh where each site has a connection to every other
site on the network or a partial mesh where sites have connection to one or two other sites but not
all the sites on network. It seems an easy option to cho
ose a full
-
mesh as a better option to
connect as it’s most redundant. But, full redundancy means more virtual circuits and more virtual
circuit connections means more time spent on setting it up and monitoring the whole network.


Thus, we leave it to the o
ffice’s management team to decide how much redundancy they require.
If they want most redundant network then, full
-
mesh should be implemented else partial
-
mesh.


All the data are stored in office headquarter, so the simpler solution would be to keep the
he
adquarter office network in center and each office have a virtual circuit connected to the office
headquarter. This simple design solves the basic connectivity issue and all offices can
communicate with main office and all other offices can communicate to
each other by routing
layer
-
3 data through headquarter.



Considering Flaw in Design


The above solution seems simple and practical, but it has a fatal flaw as the whole network relies
completely on the main office. If there is an emergency like fire, flood

etc, it will break the whole
network and will make it unusable. It is a good practice to have all data route through the main
office, but it is possible to continue business in less centralized fashion which can provide
redundant links between key offices

and also ensure the business continuity in an unplanned
event like fire, flood etc.


Bandwidth Calculation


Total Bandwidth Utilization formula determines the percentage of available bandwidth that is
required to support the usage profile (browsing/viewin
g sites and uploading/download files)
associated with your regional site (number of user and their percentage of activity). This result
represents the average continuous impact that the group of regional users will likely have on the
WAN connection between

the regional and central site.


Following is the symbolic representation of the formula to find percentage of available
bandwidth:



A x B









---------

= Percentage of available bandwidth.



C


A =
Network cost (KB) of the baseline user operati
on. In simpler language it is total number of
kilobytes transferred across the network when a common user operation is performed




B =
Number of Normalized User Operations per second. The total number of Normalized

User Operations per second perfor
med on the portal site by the group of users.


C =
Total bandwidth available. The total bandwidth

(Mbps) available on the network connection
or the connection that links group of users to the portal site. Consider only those connections that
link the user
directly to the portal site.


Final Notes to be considered:


If there is a rapid growth in the office sites then the mesh topology which we have already
considered could become a real ‘mess’ as adding one site can result in couple (or more) of new
PVC conn
ections. So to control the number of PVC connections in the network it’s would be
better to have a dual
-
hub design. With the dual
-
hub design, any one office chooses one of its sites
to serve as a second hub and makes PVC connections to each of the other si
tes from there.



References:



http://www.ciscopress.com/articles/article.asp?p=25259


http:/
/www.cisco.com/warp/public/779/largeent/design/frame_relay_wan.html


http://office.microsoft.com/en
-
us/sharepointportaladmin/HA100906131033.aspx


http://www.networkcomputing.com/netdesign/frame1.html