CS 622 Distributed Networks

Similar documents
There are two ways to establish VCs:

Intranets and Virtual Private Networks (VPNs)


Routing. Information Networks p.1/35

Wide Area Networks (WANs) Slide Set 6

Frame Relay Topologies and Designs

CSS 343 Data Structures, Algorithms, and Discrete Math II. Graphs II. Yusuf Pisan

CSC 421: Algorithm Design & Analysis. Spring 2015

Precept 4: Traveling Salesman Problem, Hierarchical Clustering. Qian Zhu 2/23/2011

The Shortest Path Problem. The Shortest Path Problem. Mathematical Model. Integer Programming Formulation

WAN Technologies CCNA 4

Lecture 19. Broadcast routing

Internetworking Part 1

looking ahead to see the optimum

CSC Design and Analysis of Algorithms. Lecture 4 Brute Force, Exhaustive Search, Graph Traversal Algorithms. Brute-Force Approach

COMP 182: Algorithmic Thinking Prim and Dijkstra: Efficiency and Correctness

Advanced Algorithms Class Notes for Monday, October 23, 2012 Min Ye, Mingfu Shao, and Bernard Moret

Greedy Algorithms. At each step in the algorithm, one of several choices can be made.

Routing Outline. EECS 122, Lecture 15

UNIT 5 GRAPH. Application of Graph Structure in real world:- Graph Terminologies:

CAD Algorithms. Categorizing Algorithms

CS/COE

Greedy Approach: Intro

DESIGN AND ANALYSIS OF ALGORITHMS GREEDY METHOD

Dijkstra s algorithm for shortest paths when no edges have negative weight.

Dijkstra s Algorithm. Dijkstra s algorithm is a natural, greedy approach towards

UNIT 3. Greedy Method. Design and Analysis of Algorithms GENERAL METHOD

Week 11: Minimum Spanning trees

Lecture Summary CSC 263H. August 5, 2016

Network Topologies & LAN,MAN and WAN. By: Mr. Binesh Kr. Singh. What is Topology

Greedy Algorithms. Previous Examples: Huffman coding, Minimum Spanning Tree Algorithms

Chapter 9. Greedy Technique. Copyright 2007 Pearson Addison-Wesley. All rights reserved.

CSE 100: GRAPH ALGORITHMS

Shortest Paths. Nishant Mehta Lectures 10 and 11

MST & Shortest Path -Prim s -Djikstra s

22 Elementary Graph Algorithms. There are two standard ways to represent a

CS200: Graphs. Rosen Ch , 9.6, Walls and Mirrors Ch. 14

Communication Networks I December 4, 2001 Agenda Graph theory notation Trees Shortest path algorithms Distributed, asynchronous algorithms Page 1

Presentation for use with the textbook, Algorithm Design and Applications, by M. T. Goodrich and R. Tamassia, Wiley, Approximation Algorithms

Module 6 NP-Complete Problems and Heuristics

tree follows. Game Trees

Shortest Paths. CSE 373 Data Structures Lecture 21

Algorithms for Minimum Spanning Trees

Chapter 9 Graph Algorithms

1.264 Lecture 23. Telecom Enterprise networks MANs, WANs

Module 6 NP-Complete Problems and Heuristics

Undirected Graphs. Hwansoo Han

On Routing Performance of MENTOR Algorithm

Math 3012 Combinatorial Optimization Worksheet

Lecture 4: Graph Algorithms

Integrated t Services Digital it Network (ISDN) Digital Subscriber Line (DSL) Cable modems Hybrid Fiber Coax (HFC)

Competitive Public Switched Telephone Network (PSTN) Wide- Area Network (WAN) Access Using Signaling System 7 (SS7)

DHANALAKSHMI COLLEGE OF ENGINEERING, CHENNAI. Department of Computer Science and Engineering CS6301 PROGRAMMING DATA STRUCTURES II

Lecture (08, 09) Routing in Switched Networks

UNIT Name the different ways of representing a graph? a.adjacencymatrix b. Adjacency list

10. Network dimensioning

Computers Are Your Future

Lecture 11: Analysis of Algorithms (CS ) 1

Graph Algorithms. A Brief Introduction. 高晓沨 (Xiaofeng Gao) Department of Computer Science Shanghai Jiao Tong Univ.

What is Multicasting? Multicasting Fundamentals. Unicast Transmission. Agenda. L70 - Multicasting Fundamentals. L70 - Multicasting Fundamentals

Computer Networks. ENGG st Semester, 2010 Hayden Kwok-Hay So

ITU Regional Seminar. Belgrade, Serbia and Montenegro, June Session 5.2. Service and applications matrix forecasting

Notes for Lecture 24

Trees, Trees and More Trees

managing an evolving set of connected components implementing a Union-Find data structure implementing Kruskal s algorithm

Optimal tour along pubs in the UK

Routing. 4. Mar INF-3190: Switching and Routing

Routing & Congestion Control

The Shortest Path Problem

Types of Computer Networks. ICS 614: Computer Networks Concepts and Principles 11

(the bubble footer is automatically inserted into this space)

CMPSC 250 Analysis of Algorithms Spring 2018 Dr. Aravind Mohan Shortest Paths April 16, 2018

Module 6 NP-Complete Problems and Heuristics

Computer Networks and Internet

Ad hoc and Sensor Networks Topology control

Fairness Example: high priority for nearby stations Optimality Efficiency overhead

Minimum-Spanning-Tree problem. Minimum Spanning Trees (Forests) Minimum-Spanning-Tree problem


Lecture 6 Basic Graph Algorithms

Frame Relay Network Performance

Module 6 P, NP, NP-Complete Problems and Approximation Algorithms

Overview. Problem: Find lowest cost path between two nodes Factors static: topology dynamic: load

AT&T OPERATING COMPANIES TARIFF NO. 1 Original Page 4-1 ADVANCED SERVICES

Chapter 7 Slicing and Dicing

Graph Algorithms: Part 2. Dr. Baldassano Yu s Elite Education

Mathematical Thinking

Chapter 4. Greedy Algorithms. Slides by Kevin Wayne. Copyright 2005 Pearson-Addison Wesley. All rights reserved.

BLM6196 COMPUTER NETWORKS AND COMMUNICATION PROTOCOLS

Dr./ Ahmed Mohamed Rabie Sayed

is the Capacitated Minimum Spanning Tree

Very Large Scale Neighborhood Search. Collaborators include: Ravi Ahuja, Ozlem Ergun, Abraham Punnen, Dushyant Sharma

CSE 100 Minimum Spanning Trees Prim s and Kruskal

1 Minimum Spanning Trees (MST) b 2 3 a. 10 e h. j m

Chapter 9 Graph Algorithms

We will discuss about three different static routing algorithms 1. Shortest Path Routing 2. Flooding 3. Flow Based Routing

CSE 202: Design and Analysis of Algorithms Lecture 3

Basic Switch Organization

Search and Optimization

ALGORITHM DESIGN GREEDY ALGORITHMS. University of Waterloo

Design and Analysis of Algorithms

Transcription:

CS 622 Distributed Networks Access Network Design Dr. Xiaobo Zhou Department of Computer Science CS622 AccessNetwork.1 Review: Traffic Normalization Traffic Generators Uniform traffic Random traffic Realistic traffic (in relation with population and distance) Traffic Normalization Total normalization Row normalization Row & Column normalization Different applications have difference characteristics CS622 AccessNetwork.2

What is Access Network Backbone network Access network Backbone link Access link A Backbone network connects major sites Access networks connect small sites to the backbone network CS622 AccessNetwork.3 Access Network Design How to decide which sites should be in the backbone network? Traffic volume Close to multiple small sites Access network collect traffic from small sites into the high speed backbone network. Sharing high speed links, enjoy economic of scale benefit. Examples of local access networks Local subscriber loop connects users of a central office. Lottery network ATM network ISP s local access network. CS622 AccessNetwork.4

A Simple Access Design Example Consider a problem with 6 access locations and 1 backbone site N1 is the backbone site, the traffic is all from and to N1 Symmetric traffic CS622 AccessNetwork.5 The Cost Matrix for 56 Kbps links Linear Cost model: fixed cost = $400/month, and $3.00/km/month for the first 300km, and then a cost of $1.75/km/month after that Cost if distance-dependent. CS622 AccessNetwork.6

A Star Design Cost=$9650; Max. Utilization=23.2% CS622 AccessNetwork.7 A Cheaper Local-Access Design N2 serves as a concentrator for N6 and N7. Local link can use shorter less expensive link Cost $8666 CS622 AccessNetwork.8

A Even Cheaper Local-Access Design Two concentrators: N2 for N6 and N7; N4 for N3 Cost: $8158 CS622 AccessNetwork.9 Move to MST the cost is optimal Choose N7 as concentrator instead of N2 The design becomes an MST T2-1+T6-1+T7-1 = 26000 26000/56000=46.4% CS622 AccessNetwork.10

MSTs not always Optimal Access Designs The MST was the optimal, because the traffic volumes were moderate When the traffic grows 50%, the MST costs $10,616 and the links to concentrators N4 and N7, i.e, N1 - N4 and N1 - N7, must have two links to keep utilization below 50% CS622 AccessNetwork.11 An Optimal Design with Increased Traffic A Constraint MST Problem N3 connects directly to N1 since through N4 will violate the utilization constraint CS622 AccessNetwork.12

Frame Relay Design Frame Relay is a relatively new service Frame relay is accomplished by connecting each site into a frame relay cloud provided by the service carrier, just like plain old phone service is accomplished by attaching the phone into PSTN cloud. Packets exceed Committed Information Rate (CIR) will be discardable by having discard eligibility (DE) bit set Three classes of charges: access link costs, provider port costs (cost to frame relay), and CIR costs. It is volume dependent and not distance-bw-nodes dependent Port Charges CIR charges per PVC (permanent virtual circuits) CS622 AccessNetwork.13 Frame Relay Design Cost Model Each node is 20 km from the provider point of presence (POP) Access link cost is same for all 7 nodes: 7 * (400 + 3 * 20) N3 PVC port POP POP PVC port: 56/64 kbps N2 Frame Relay Cloud PVC port N4 POP POP PVC port: 128 kbps Σi T i ->1 = 59 bps; 50% N1 CS622 AccessNetwork.14

Frame Relay vs. Lease-Line Cost Let x be the distance from the site to the center Fixed cost=$400/month; $3.00/km/month; Leased-line Cost Model 6 * 400 + 6 * 3.00 * x Frame Relay Cost Model: Assume frame relay provider has point of presence (pop) at each site (20km away). Each site connects to the frame relay network N1 uses a 128 kbps link port (59 bps total traffic on way, 50%) Other nodes use 56kbps links - Port charges: 6 * 250 + 500 = 2000 Access charges: 7 * (400 + 3.0 * 20) = 7*460 = 3220 CIR charges: 4 * 30 + 2 * 25 = 170, if 4 PVCs with 16kbps CIR and 2PVCs with 8kbps Frame relay cost: 2000 + 3220 + 170 = $5390/mo Solve 2400 + 18 * x = 5390 x = 166.11 km, Break even point Most WAN are larger than this Frame Relay is a good candidate CS622 AccessNetwork.15 Choosing Backbone Nodes If the division between large sites and small sites is distinct, there is usually no problem in deciding backbone nodes Definition 5.1: Given a set of sites Ni and traffic matrix T(i,j), weight(ni) = Σ j(t(i,j)+t(j,i)) Sometimes, the weights of nodes indicate the choices of backbone nodes or traffic centers. Design Principle 5.3 It is acceptable for small nodes to route their traffic via big nodes, but generally we do not want to route the traffic between big nodes via the small nodes In airline networks, there are four kind of nodes Hubs Cities with enough traffic to have nonstop flights to cities other than hubs Cities with enough traffic to have nonstop flights to their hubs Little cities CS622 AccessNetwork.16

3 Types of Local Access Problems 1. Access node s traffic are considerably smaller than the smallest link. But occasionally, they may need to download huge file Use frame relay or access tree capacitated spanning tree building problem 2. Access node s traffic is comparable to the capacity of the smallest link Choice 1: connect them directly to the hub/center Choice 2: put a concentrator between hub and those nodes Concentrator placement problem, local access tree problem 3. Access node s traffic can fill several low-speed access lines Choice 1: multiple links to multiple backbone nodes Choice: a high speed link to a backbone node - Multiple low-speed links give better reliability - The high-speed link gives better performance and lower cost CS622 AccessNetwork.17 One-speed One-Center Design and Capacitated Trees Example: 19 nodes to be connected to a hub, N14 (total 20 nodes) Requirement: 4 sites can share a line Symmetric traffic: to and from each node Ni to N14 is 1200 bps Capacity of the one-speed link: 9600 bps Link utilization limitation: < 50% The problem becomes a capacitated access tree building problem Solutions: SPT by Dijkstra s algorithm MST by Prim s algorithm Intermediate trees by Prim-Dijkstra with 0<α<1 Exhaustive search for an optimal tree Other new algorithms? CS622 AccessNetwork.18

SPT(Star) High cost: $26358 Max_uitlization is 12.5% (12000 / 56000) How about aggregate traffic of nodes and connect to the center? CS622 AccessNetwork.19 MST Cost: $18,730 15 sites from N12 4 parallel links CS622 AccessNetwork.20

Prim-Dijkstra with α=0.3 Lower cost: $15930 Potential improvements: N11 can go through N4; Two clusters with N18 and N9 as concentrators ( a double link -> 2 single links). A double D96 link CS622 AccessNetwork.21 How Many Trees to Search for Optimal Designs? Exhaustive search for an optimal design is usually infeasible. Cayley s Theorem: Given n nodes, there are n n-2 different spanning tree For 20 nodes, there are 20 18 =2.621*10 23 trees. CS622 AccessNetwork.22

Constraint Minimum Spanning Tree Problem CMST problem: Given a central node N0 and a set of other nodes (N1,, Nn), a set of weights(w1,,wn) for each node, the capacity of a link, W, and a cost matrix Cost(i, j), find a set of trees T1,, Tk such that each Ni belongs to exactly one Tj and each Tj contains N0, and i T j w i, i> 0 min < W Trees l Links Cost ( end1, end2 ) l l CS622 AccessNetwork.23 A Greedy CMST Algorithm Sort the edges/links according to the cost take the lowest cost edge from sorted list; add it to the solution subtrees if the addition does not violated the constraint w < W; go to s1. Assume W=3, each node has wi = 1, and the following topology: i i T j, i> 0 T1 T2 CS622 AccessNetwork.24

Esau-William Algorithm Initially, each node starts off in a tree with itself Compute the tradeoff function: Tradeoff(Ni) = min j Cost(i, j) - Cost(Comp(N i ), Center) // attractive if 0 saving by linking Ni to Nj, rather than linking directly to Center (STAR) Cost(Comp(i),Center) is the cost of connecting the component with Node Ni to the Center. It is equivalent to the cost of the shortest path from the Center to any node in the component. Cost(i,j) is the link cost from Node Ni to Node Nj min j Cost(N i,n j ) suggests pick the closest neighboring Node Nj Maintain a sorted list of links based on the Tradeoff() value Tradeoff() is attractive if negative, smaller -> more attractive Actually, in each iteration, we only consider the shortest link out of a node to a neighbor that does not belong the component of the node L1: adds the top link (min Tradeoff) in the list to the solution if the weight constraint of the component is satisfied; otherwise, reject it update the tradeoffs in other links due to the newly added link and resort the list go to L1 CS622 AccessNetwork.25 Apply Esau-William Algorithm Assume W=3, each node has wi = 1, and the following topology: Tradeoff(1) = min j Cost(1, j) - Cost(Comp(1),Center) = Cost(1,3) - 7 //comp(1) contains Node 1 = 5 7 = -2 // pick closest neighbor, Node 3 Tradeoff(2,0) = 6-8 = -2 Tradeoff(3,1) = 5 11 = -6 Tradeoff(4,2) = 7 14 = -7 Tradeoff(5,3) = 8 17 = -9 Tradeoff(5) is lowest one Accept link(5,3) to the solution since weight constraint on component with nodes 5 and 3 are not violated. Σwi = w 5 + w 3 = 2 <= W=3 Effectively this picks the faraway node with short link to its neighbor and group them as component. CS622 AccessNetwork.26

Apply Esau-William Algorithm(2) Update the tradeoffs in other links due to the newly added link and resort the Tradeoff() list Tradeoff(5,4) = 9 11 = -2 // Next shortest link out of 5 is (5,4) // (Comp(5)=11,node 5 goes through node 3 to center) // 3 is already in the component, not considered Tradeoff(3,1) = 5 11 = -6 // not changed Tradeoff(1,3) = 5 7 = -2 Tradeoff(2,4) = 6 8 = -2 Tradeoff(4,2) = 7 14 = -7 Tradeoff(5,4) = 9 11 = -2 Pick Tradeoff(4,2) since it is smallest Accept link(4,2) since weight constraint on component with nodes 4 and 2 are not violated: Σwi = w 4 + w 2 = 2 <= W=3 CS622 AccessNetwork.27 Apply Esau-William Algorithm(3) Tradeoff(4,3) = 8 8 = 0 // Tradeoff(2,1)= -2 not changed Tradeoff(3,1) = 5 11 = -6, not changed Tradeoff(5,4) = 9 11 = -2 Tradeoff(1,3) = 5 7 = -2 Tradeoff(2,1) = 6-8 = -2 Tradeoff(4,3) = 8 8 = 0 Pick Tradeoff(3,1) Accept link (3,1) since weight constraint on component with nodes 1, 3 and 5 are not violated Σwi = w 1 + w 3 + w 5 = 3 <= W=3 CS622 AccessNetwork.28

Apply Esau-William Algorithm (4) and (5) Since nodes 5 and 3 now go through node 1 to the Center 0, Tradeoff(5,4) = 9 7 = 2 Tradeoff(3,4) = 8 7 = 1 Tradeoff(1,2) = 6 7 = -1 Tradeoff(2,1) = 6-8 = -2 Tradeoff(4,3) = 8-8 = 0 Tradeoff(2,1) is lowest but adding link(2,1) result a component with 5 nodes violate Σwi <= 3 Reject(2,1) recompute Tradeoff(2,0) = 8 8 = 0 Reject(1,2) due to the same reason Recompute Tradeoff(1,0) = 7 7 =0 Pick link(1,0) Pick link(2,0) completes the access network CS622 AccessNetwork.29 Creditability of Esau-Williams Algorithm The Esau-Williams Algorithm is heuristic, though works quite well Example: an Esau-Williams design with up to 7 sites per line CS622 AccessNetwork.30

Creditability of Esau-Williams Algorithm (Cont.) Given a set of sites N and a capacitated tree T, 1-exchange test: no cheaper link can be substituted for an existing link without violating the capacity constraints Example: for homogeneous traffic, up to 4 sites on a line For homogeneous traffic, the Esau-Williams does well on 1-change test CS622 AccessNetwork.31 Esau-Williams and InHomogenous Traffic One-speed link, but different sites have different traffic Link capacity is 9600 bps, 50% of sites have a requirement of 2400 bps, and the other 50% have a requirement of 4800 bps CS622 AccessNetwork.32

Line Crossing and Esau-Williams Are crossings an indication of a lack of creditability in capacitated trees? In tours, yes! CS622 AccessNetwork.33 Line Crossing in Access Designs: Sharma s Algorithm 1. Compute the angle θ s from each site S to the central site C. If S and C have the same coordinate, set θ s =0 2. Sort the angles θ s 3. Beginning at a site S first, create a set of nodes clockwise (or counterclockwise) from S first A set is complete when adding the next node would put Σ set w(site) > W. The next set starts with that node. 4. The design is completed by building a MST on each set with the addition of the central node C. Theorem 5.2 Sharma s algorithm builds CMSTs withour line crossings CS622 AccessNetwork.34

Sharma s Algorithm Result A better solution CS622 AccessNetwork.35 Creditability of Sharma Algorithm Much higher failure rate than Esau-Williams by 1-change test CS622 AccessNetwork.36

Sharma vs. Esau-Williams EW_Ratio = SharmaCost/EWCost; S_Ratio = 1/EW_Ratio CS622 AccessNetwork.37