Introduction to Network Design. e 2 Acknowledgement Edward Chow Robert Cahn.

Slides:



Advertisements
Similar presentations
There are many types of WAN technologies that can be used to solve the problems of users who need network access from remote locations. We will go through.
Advertisements

E&CE 418: Tutorial-4 Instructor: Prof. Xuemin (Sherman) Shen
Maciej Stasiak, Mariusz Głąbowski Arkadiusz Wiśniewski, Piotr Zwierzykowski Groups models Modeling and Dimensioning of Mobile Networks: from GSM to LTE.
11-1 Typical Capital Budgeting Decisions Plant expansion Equipment selection Equipment replacement Lease or buy Cost reduction.
QUEUING MODELS Based on slides for Hilier, Hiller, and Lieberman, Introduction to Management Science, Irwin McGraw-Hill.
Lecture 13 – Continuous-Time Markov Chains
CHAPTER 15 Funds-Transfer Pricing and the Management of ALM Risks What is in this Chapter? INTRODUCTION TRADITIONAL TRANSFE PRICING AND ITS PROBLEMS MATCHED-FUNDS-TRANSFER.
Model Antrian By : Render, ect. Outline  Characteristics of a Waiting-Line System.  Arrival characteristics.  Waiting-Line characteristics.  Service.
Waiting Lines and Queuing Theory Models
EEE440 Modern Communication Systems Cellular Systems.
Chapter 13 Queuing Theory
27 August EEE442 COMPUTER NETWORKS Test results & analysis.
Management of Waiting Lines
1 Internet Management and Security We will look at management and security of networks and systems. Systems: The end nodes of the Internet Network: The.
Trunking & Grade of Service
Lecture 14 – Queuing Systems
Local Number Portability Traffic Analysis Calculations ENGR 475 – Telecommunications October 19, 2006 Harding University Jonathan White.
BASIC TELECOMMUNICATIONS
Queuing Models and Capacity Planning
The Marriage Problem Finding an Optimal Stopping Procedure.
Traffic modeling and Prediction ----Linear Models
Lecture 11: Cellular Networks
CHAPTER TEN Dr. Rami Gharaibeh BUSINESS MODEL ANALYSIS.
Spreadsheet Modeling & Decision Analysis
Introduction to Management Science
Basic teletraffic concepts An intuitive approach
Capacity analysis of complex materials handling systems.
ISM 270 Service Engineering and Management Lecture 7: Forecasting and Managing Service Capacity.
1 Chapter 16 Applications of Queuing Theory Prepared by: Ashraf Soliman Abuhamad Supervisor by : Dr. Sana’a Wafa Al-Sayegh University of Palestine Faculty.
Designing a Voice over IP Network Chapter 9. Internet Telephony 2 Introduction The design of any network involves striking a balance between three requirements.
Management of Waiting Lines McGraw-Hill/Irwin Copyright © 2012 by The McGraw-Hill Companies, Inc. All rights reserved.
Florida Institute of technologies ECE 5221 Personal Communication Systems Prepared by: Dr. Ivica Kostanic Lecture 17: Traffic planning Spring 2011.
Capacity analysis of complex materials handling systems.
Computer Networks Performance Metrics. Performance Metrics Outline Generic Performance Metrics Network performance Measures Components of Hop and End-to-End.
1 Queuing Analysis Overview What is queuing analysis? - to study how people behave in waiting in line so that we could provide a solution with minimizing.
TCOM 540/11 TCOM 540 Network Optimization: Routing, Flow Management, Capacity Modeling Dr. John G. Leigh
EPL 476 Fundamental Concepts in Wireless Networks
1 Voice Traffic Engineering & Management. 2 PSTN and PBX networks are designed with 2 objectives: Maximize usage of their circuits Maximize usage of their.
M/M/1 queue λn = λ, (n >=0); μn = μ (n>=1) λ μ λ: arrival rate
COMT 2201 Managing Telecommunications Systems. COMT 2202 Managing Telecommunications Configuration Management Security Management Accounting Management.
1 Chapters 8 Overview of Queuing Analysis. Chapter 8 Overview of Queuing Analysis 2 Projected vs. Actual Response Time.
Network Design and Analysis-----Wang Wenjie Queuing Theory III: 1 © Graduate University, Chinese academy of Sciences. Network Design and Performance Analysis.
1 Dr. Ali Amiri TCOM 5143 Lecture 8 Capacity Assignment in Centralized Networks.
Waiting Lines and Queuing Theory Models
EE 489 Telecommunication Systems Engineering University of Alberta Dept. of Electrical and Computer Engineering Introduction to Traffic Theory Wayne Grover.
1 1 Slide © 2009 South-Western, a part of Cengage Learning Slides by John Loucks St. Edward’s University.
3-location Data Network Design 3 locations separated by 200 km among pairs. Give the new populations below 296 users, design the data network.
Maciej Stasiak, Mariusz Głąbowski Arkadiusz Wiśniewski, Piotr Zwierzykowski Model of the Nodes in the Packet Network Chapter 10.
1 Queuing Delay and Queuing Analysis. RECALL: Delays in Packet Switched (e.g. IP) Networks End-to-end delay (simplified) = End-to-end delay (simplified)
2/16/2016 Subject Name: Digital Switching Systems Subject Code:10EC82 Prepared By: Aparna.P, Farha Kowser Department: Electronics and Communication Date:
Small Office / Home Office (SOHO) IP Telephony Who says You Can’t Run with the Big Dogs?
Thought of the day Thought of the day Difference between science and spirituality is same as the difference between word and silence. Sameer Trapasiya.
Management of Waiting Lines Copyright © 2015 McGraw-Hill Education. All rights reserved. No reproduction or distribution without the prior written consent.
1 Lecture 06 EEE 441: Wireless And Mobile Communications BRAC University.
Tele traffic A telephone network is composed of a variety of common equipment, such as: Digital receivers Call processors Inter-stage switching links &
Fundamentals of Cellular Networks (Part IV)
Chapter 8 Switching Copyright © The McGraw-Hill Companies, Inc. Permission required for reproduction or display.
Management of Waiting Lines
Trunking Theory In general, there are many more users than channels--everyone can’t have a permanent personal channel! Fortunately, only a small fraction.
Introduction to Network Design
UNIT-V TRAFFIC ANALYSIS
Solutions Hwk Que3 1 The port of Miami has 3 docking berths for loading and unloading ships but is considering adding a 4th berth.
IV-2 Manufacturing Systems modeling
The M/G/1 Queue and others.
Lecture 13 – Queuing Systems
3-location Data Network Design
Network Planning & Capacity Management
Chapter-5 Traffic Engineering.
Network Planning & Capacity Management
Course Description Queuing Analysis This queuing course
Presentation transcript:

Introduction to Network Design

e 2 Acknowledgement Edward Chow Robert Cahn

e 3 Introduction to Network Design Network design is Create network structure (blue print) Decide how to allocate resource and spend money Two basic questions: How much it cost to build a usable network? How much improvement does $x buy? Answer: Depend on network services and components available We will concentrate the techniques and algorithms

e 4 Network Evaluation Every network has three characteristics: –Cost –Performance –Reliability First we need to find agree-upon quantitative numbers. Based on the quantitative numbers of these characteristics, we can evaluate different design alternative by ordering them and ruling out losers.

e 5 Example 1 Four designs for a network design problem

e 6 Cheap Network (by Intrepid)

e 7 Messy Network

e 8 Rank Designs by Attributes

e 9 Justify the Designs There can be factors that decides the final choice: Whether the company is expanding Maybe the proposed designs are not as expected. In outsourcing situation, you may ask for redesign You may not have to serve as designer but as an evaluator.

e 10 Compare Designs

e 11 What is more important, Performance or Cost? 150 cashiers  $13,725/month CEO  $152,500/month One user my justify the building a high performance network

e 12 Two-Location Problem It is called “Hello World” of Network Design. It is undaunting yet interersting problem! Design a network connecting two locations, 200km apart. Anagon city with 5 employees, Bregen with 10. Each employee –call other site 4 times/day, avg. 5 min. each. 4*5*15=300 min/day –call others in the same office 10 times/day about joint work, each last avg. 3 min. 10*3*15=450 min/day Note here we are not using C(10,2)+C(5,2) for the # of calls How can we best provide the communications between the two cities

e 13 Cost of Network Services and Components Network equipment purchase is typically amortized at 3% per month. The PBX Private Branch Exchange would cost $60/month

e 14 Public Switched Telephone Network Solution

e 15 Cost of PSTN Solution Assume 21 2/3 work days=65/3 days Local call: 450min/day*0.05$/min*65/3day=487.5$ Long distance call; 300min/day*0.4$/min*65/3day=2600$

e 16 Utilization Analysis 5 employees at Anagon place 4*5min*5=100 min calls/day to Bregen 10 employee at Bregen place 4*5min*10=200 min calls/day to Anagon 300 min long distance calls are shared among 5 employee at Anagon. That is 300min/5=1 hour/employee/day on long distance. For 8 hour day, each phone at Anagon is busy 25%=2/8? of the time. While phone at Bregen is busy 18.75%=1.5/8?  Low Utilization Resaon: Each employee at Anagon makes 10*3min/day=30min local call, but it will tie up other employee’s line. Assume no conference call. Therefore each line is 30min*2=1hour/day is busy on local call.

e 17 Design Principle 2.1 Good network designs tend to have many well- utilized components.

e 18 PBX Solution $487.5/month for local calls saved With$60*2=$120 amortized PBX cost, we actually save $367.5/month Reliability degraded?! Performance?

e 19 Reducing Trunks at Bregen There can be 5 intersite simultaneous calls. Reduce 10 outgoing trunks at Bregen to 5. $25/month*5=$125/month access fee saving. How can we reduce the cost further? Clue  Study the usage pattern

e 20 Famous 2 camel hump Telephone Traffic Daily Pattern Traffic measured in Erlangs.

e 21 Erlang: the Traffic Measure Unit Definition 2.1 If call arrive rate=  and departure rate= , Then the call intensity is E=  Erlangs. In honor of Danish Telephone engineer Erlang. Example 1. Calls arrive 2 per min., and hold for an average of 3 min, then  =2 and  =3, E=  =6 Erlangs. Note that hold time (H)= 1/departure rate; H=1/ . Apparently one line cannot handle this amount of traffic. When a call comes and all lines are busy, the call is blocked. How many lines can reduce the blocking probability to x%?

e 22 Erlang Calculation In our 2-location case, 15 places 4 long distance calls/day, each call last avg. 5min. Assume 8 hr day What is the call (traffic) intensity for the day? =15*4 calls/8 hr = 15/2 calls/hr H=5 min/call =1/12 hr/call  =1/H=12 calls/hr = 0.2 calls/min E=  =(15/2)/12=15/24=5/8 Erlangs. Assume 20% of the traffic in the busy hour. What is the call intensity in the busy hour? =60 calls/day * 0.2 = 12 calls/hr = 0.2 calls/min  =1/H=12 calls/hr = 0.2 calls/min E=  =12/12=1 Erlang.

e 23 Queueing Theory for System with Loss Assume a telephone system with multiple lines. –When a call comes and all lines are busy, the call is blocked. Unlike data network, calls are not buffered or queued if lines are not available. –Or, you can consider it is a finite queue, i.e., after queue full (line busy), further call are blocked. How many lines can reduce the blocking probability to x% for a given traffic density? Queueing theory can be used analyzed the telephone system’s performance, specifically the blocking probability.

e 24 M/M/2 Queue

e 25 Loss with m Lines (m servers, no queue) A: Arrive Rate; D: Departure Rate; E=A/D AP k-1 =kDP k  P k =E/k * P k-1

e 26 Erlang-B Function

e 27 Calculating the Blocking In 2-location case, with busy hour, A=0.2 calls/min, D=0.2calls/min. When 1 call in progress, departure rate is 0.2 calls/min When 2 calls in progress, departure rate is 0.4.

e 28 Design Intersite Link Given we can tolerate x% blocking, how many lines we actually need? Here E=1 (busy hour). Carried load=E(1-B(E,m)). qi=fraction of load on link i.

e 29 Simplified Traffic Profile Instead of 2 camel hump traffic pattern, simplify it to two levels: peak and off-peak. 60min*2*0.4*(21+2/3)= min*6*0.4*(21+2/30=1560

e 30 Reduce Cost by Using Leased Lines Instead of charged by # of calls, pay monthly cost. The leased line costs $275/month replacing two PSTN lines, one on each end, for a total of $50. The calls placed on the leased line save $0.4/min. Strategy: Place calls on leased lines first. Question: How many released lines should we use? Let us figure out the cost saving of each leased line. First focus on busy-hour analsys

e 31 Busy Hours Analysis for Leased Line Saving What is the value of busy-hours traffic carried by a single leased line? With 60min/hour usage, traffic is E=60/60=1 Erlang. From table 2.3, the fraction of calls on a single leased line is 0.5. The other half got blocked. Saving on dialup cost: 0.5*$1040=$520/month The net equipment cost of leased line=$275-$50 (replaced two PSTN lines)= $225/month. The net cost saving = $520-$225=$295/month It is justifiable.

e 32 2 nd Leased Line Saving For busy hours, the 2 nd leased line will carry 0.3 traffic. Dialup cost saving = 0.3*$1040=$312/month. $312 > $225  still justified. The third leased line only carries busy hour traffic. Dialup cost saving = *$1040=$143/month. We need to see if the cost saving of off-peak hour usage of the 3 rd leased line adds adds up to $225.

e 33 Off-Peak Analysis of 3 rd Leased Line Off-Peak hours traffic = 30/60= 0.5 Erlang. The fraction of calls carried by the first leased line =B(0.5,0)-B(0.5,1)=1-(0.5*B(0.5,0))/(0.5*B(0.5,0)+1)=2/3. Cost saving for the 1 st leased line=(2/3)*$1560=$1040/month. The fraction of calls carried by the 3rd leased line =B(0.5,2)-B(0.5,3)=(1/13)-(0.5*B(0.5,2))/(0.5*B(0.5,2)+1) =1/13-1/79= Cost saving for the 3 RD leased line during off-peak hours =(0.0643)*$1560=$100.25/month. $ $143=$243.25> $225  justified!

e 34 Final Design

e 35 Cost of Final Design 2PBX’ 3 leased lines 4 PSTN lines $ $ =$ /month saving