WAN and LAN Plan and Design - Graduate Project Report | 0614 774, Study Guides, Projects, Research of Engineering

Download WAN and LAN Plan and Design - Graduate Project Report | 0614 774 and more Study Guides, Projects, Research Engineering in PDF only on Docsity!

WAN/LAN PLAN&DESIGN

GRADUATE PROJECT

Qian Li (Alison)

4/24/

WAN/LAN Planing&Design Professor Ronald G. Fulle 0614.774. Due April 24, 2008

Purpose

The purpose of this project is to design and develop a network, then to test different "what-if"

scenarios with varying topologies, traffic matrices, and configurations. Network I can use

reporting features to compare the results of different scenarios, and thereby determine the most effective and least costly network designs to meet future demands. To accomplish this

purpose I will create two baseline scenarios, both of them use the map of China. One is

transparent in OCH layer, and the other is opaque in OCH layer. The networks must be robust,

optimal design. The function of two scenarios is to implement communication through whole country area. They represent one worldwide big telecommunication company named QL

Communication Corporation, Ltd. All these requirements decide my topologies and network

performance. In the end of this project, I will compare two scenarios, and give the answer of which one is better.

Resources

This project was completed by using software SP Guru Transport Planner 14.0, Educational

Version by OPNET Technologies, Inc.

Approach

Baseline Scenarios

1. Create one new project with two scenarios

First step, I created two scenarios of my project, the topology is based on the Map of China, as our company provides nationwide communication of China.

Scenario 1—China (Transparent) Scenario 2— China (Opaque) Name: Nationwide Communication of China Name: Nationwide Communication of China

Network Topology: Network Topology:

2. Define the network topology

Fig 5- link design result

Fig 6- OTS layer link browser Fig 7- OMS layer link browser

Fig 8- OCH layer link browser

I didn’t specifically set DCL link. Later, when I dimension my project, DCL layer will be mesh topology.

Node Design:

Fig 9- OTS layer node browser Fig 10- OMS layer node browser

Fig 11- OCH layer node browser (before changing Fig 12- DCL layer node browser (before changing

traffic) traffic)

I use default equipment properties, which are shown below:

In DCL layer, traffics are 100% routed, in OCH layer are 59.5%. Transparent:

In DCL layer, traffics are 100% routed, in OCH layer are 63.6%. The result of my project:

Fig 13- DCL layer logical link

2. Dimension DCL/OCH Layer When you dimension a network at a network layer, SP Guru Transport Planner installs the required

node and link equipment to support a traffic matrix at that layer. SP Guru Transport Planner performs upgrade dimensioning: it upgrades the network equipment to meet the new traffic demands, but retains existing equipment and traffic. The reason why I used dimension can be seen from following test: (I tested the network after I changed network topology to get lower OCH layer utilization. ) Before dimension:

Layer Numberof Links

Link Utilization

Used Capacity

Total Capacity

Utilization (%) Unit

**Between

and 90.0 %**

Between 90.0 and 75.0 %

Between 75.0 and 0.0 %

OTS 26 26 0 0 1,300 1,300 100.00 Fiber Pairs

OMS 26 0 0 26 57 1,300 4.38 Fiber Pairs

OCH 26 3 1 22 1,117 2,280 48.99 Wavelengths

DCL^115 0 5 110 4,625^ 38,880^ 11.90^ STS-1 Units

After dimension: Dimension OC-

Traffic Matrix

Newly Routed Connections

Total Routed Connections

Demanded Connections

Percent Routed

Total Hops of Working Paths

Mean Hops of Working Paths OC-48 322 322 322 100.0 818 2. Utilization Variable Value OCH Used Link Capacity 818 wavelengths OCH Equipped Link Capacity 2,280 wavelengths OCH Link Utilization 35.88% OCH Node Utilization 100.0%

Dimension OC-

Traffic Matrix

Newly Routed Connections

Total Routed Connections

Demanded Connections

Percent Routed

Total Hops of Working Paths

Mean Hops of Working Paths OC- 192 122 122 122 100.0 297 2. Utilization Variable Value OCH Used Link Capacity 297 wavelengths OCH Equipped Link Capacity 2,280 wavelengths OCH Link Utilization 13.03%

The capacity in DCL layer is 100%, but in OCH layer OC-192 is 63.6%. In OCH layer,

system has higher link utilization.

Transparent:

The capacity in DCL layer is 100%, but in OCH layer OC-192 is 59.5%. In OCH layer, system has higher link utilization.

4. Traffic Variation Opaque:

OC-192 performances better than OC-48 here. Transparent:

OC-192 performances better than OC-48 here.

5. Network Cost

Opaque: Transparent:

Transparent:

In wavelength usage aspect, two scenarios are almost same.

8. Evaluate Availability

This report is too big; I will show it to Professor later by myself.

9. Failure Analysis

Opaque: Transparent:

The result shows there is no lost in my scenarios.

Growth of 20% Traffic Simulation

In this part, I will increase my traffic up to 20% of previous simulation.

1. Route DCL/OCH Traffic (from 1-8 are the results got before I made final changes, used

to compare with my changing)

Opaque:

In DCL layer, traffics are 100% routed, in OCH layer are 53.4%.

Transparent:

The capacity in DCL layer is 100%, but in OCH layer OC-192 is 68%. In OCH layer, system has higher link utilization.

3. Traffic Variation

Opaque:

OC-192 performances better than OC-48 here. Transparent:

OC-192 performances better than OC-48 here.

4. Network Cost

Opaque: Transparent:

5. Link Utilization per Layer Opaque:

Transparent:

In wavelength usage aspect, two scenarios are almost same.

7. Evaluate Availability This report is too big; I will show it to Professor later by myself. 8. Failure Analysis Opaque: Transparent:

The result shows there is no lost in my scenarios. If I fail some links in my network (use one scenario to test):

Layer Lost Recovered Unaffected Total Unit

OCH 45 0 399 444 Wavelengths

DCL 157 0 1,691 1,848 STS Connections

LOP 0 0 0 0 VT Connections

So, that means my network now is good, with all the links there is no lost.

Conclusion

In this project, I created two scenarios for my primary traffic, and then I increased my traffic

to 20% and duplicated two scenarios to see how the system handles increasing traffic. From above simulation results, we can see that transparent transport network is better than opaque.

The concept behind my conclusion is: In opaque transport networks, about 60 to 80 percent

of the network installation cost goes to the line cards and transponders that convert the electrical signals to optical signals. To reduce this cost, transparent networks regenerate

signals within nodes and move regeneration away from the WDM links—specifically the

long-reach transponders that terminate channels within these links. Instead, regeneration occurs within nodes selectively.

Scenario Comparison