Broadband and Wireless Networking Laboratory School of Electrical and Computer Engineering

Ian F. Akyildiz

C. Scoglio, J. de Oliveira, T. Anjali, L. Chen,C. Scoglio, J. de Oliveira, T. Anjali, L. Chen,J. A. Smith*, G. Uhl* and A. Sciuto*J. A. Smith*, G. Uhl* and A. Sciuto*

NEXT GENERATION INTERNET NEXT GENERATION INTERNET PHYSICAL TESTBED:PHYSICAL TESTBED:

RESEARCH AND MANAGEMENT ISSUESRESEARCH AND MANAGEMENT ISSUES

Broadband and Wireless Networking Broadband and Wireless Networking LaboratoryLaboratory

School of Electrical and Computer EngineeringSchool of Electrical and Computer Engineering

Georgia Institute of Technology, Atlanta, GA, Georgia Institute of Technology, Atlanta, GA, USAUSA

*NASA Goddard Space Flight *NASA Goddard Space Flight CenterCenter

Greenbelt, MD, 20771, USA Greenbelt, MD, 20771, USA

IFA'02IFA'02 NGI PHYSICAL TESTBED: RESEARCH AND MANAGEMENT ISSUESNGI PHYSICAL TESTBED: RESEARCH AND MANAGEMENT ISSUES 22

IP QoS NETWORK IP QoS NETWORK PROJECTPROJECT(NASA Goddard, Raytheon, Swayles)(NASA Goddard, Raytheon, Swayles)

ChallengesChallenges– DifferentiateDifferentiate

d Servicesd Services– End-to-End End-to-End

QoSQoS– Integrated Integrated

Services for Services for MultimediaMultimedia

– Network Network ManagementManagement

NASA Ames Research

Center

NASAGoddard SpaceFlight Center

BWNLaboratory

GATECH

Abilene


BWN-Lab Physical Testbed:BWN-Lab Physical Testbed:

Experiments and IssuesExperiments and Issues


BWN-Lab TESTBEDBWN-Lab TESTBED

ExternalLightstream

ATM155 Mbps

Gigabit Ethernet

Fas

t Eth

erne

t

7204 VXR

7204 VXR

Catalyst4000

Lightstream 1010

ATM 622 Mbps

7505Catalyst

6506


TESTBED HARDWARETESTBED HARDWARE

2 Cisco 7200 routers:2 Cisco 7200 routers:– FastEthernet/OC3/GigabitEthernet FastEthernet/OC3/GigabitEthernet

interfacesinterfaces Cisco 7500 router:Cisco 7500 router:

– GEIP+/OC12 interfacesGEIP+/OC12 interfaces Cisco Catalyst 6506 layer 3 switch:Cisco Catalyst 6506 layer 3 switch:

– GEIP+ interfaceGEIP+ interface Cisco Catalyst 4000 switch:Cisco Catalyst 4000 switch:

– FastEthernet portsFastEthernet ports Cisco LightStream 1010 switch:Cisco LightStream 1010 switch:

– OC3 interfacesOC3 interfaces


TESTBED SOFTWARETESTBED SOFTWARE

Cisco IOS Cisco IOS – Version 12.2(1)E1 on the routersVersion 12.2(1)E1 on the routers– Native IOS on switch 6500Native IOS on switch 6500

End-hostsEnd-hosts– MS Win ME and Linux RedHat 7.2MS Win ME and Linux RedHat 7.2– ALTQ 2.2:ALTQ 2.2: Scheduling/Queueing Software Scheduling/Queueing Software– Iperf:Iperf: traffic generation and traffic traffic generation and traffic

characteristic measurementcharacteristic measurement– MRTGMRTG and and MRTG++MRTG++ (modified MRTG for (modified MRTG for

10s sampling)10s sampling)


DiffServ TESTBED DiffServ TESTBED NETWORK TOPOLOGYNETWORK TOPOLOGY

DS domain

Abilene

DS domain

NASA

DS domain

marking

markingpolicingscheduling

shaping

BWN-Lab


DiffServ EXPERIMENTSDiffServ EXPERIMENTS EDGE COMPONENTSEDGE COMPONENTS

– Classification/Marking:Classification/Marking: Policy Based Routing Policy Based Routing (PBR)(PBR) allows allows classifications based on IP Precedence.classifications based on IP Precedence.

– Policing:Policing: Committed Access Rate Committed Access Rate (CAR)(CAR) enforces a enforces a specified traffic profile preventing non-conformant specified traffic profile preventing non-conformant traffic from entering the network. traffic from entering the network.

– Shaping:Shaping: Generic Traffic Shaping Generic Traffic Shaping (GTS)(GTS) follows the follows the token bucket algorithm.token bucket algorithm.

CORE COMPONENTSCORE COMPONENTS– Queueing:Queueing: Class-Based WFQ Class-Based WFQ (CBWFQ)(CBWFQ) regulates traffic regulates traffic

submitted to the network, which may delay packets to submitted to the network, which may delay packets to adjust traffic stream characteristics to a defined adjust traffic stream characteristics to a defined profile. profile.

– Congestion avoidanceCongestion avoidance: Weighted RED : Weighted RED (WRED)(WRED) allows allows definition of multiple drop probability profiles.definition of multiple drop probability profiles.


DiffServ EXPERIMENTSDiffServ EXPERIMENTSSCHEDULING WITH CBWFQSCHEDULING WITH CBWFQ

We validated CAR and CBWFQ as the policing and We validated CAR and CBWFQ as the policing and scheduling mechanisms for DiffServ implementationscheduling mechanisms for DiffServ implementation

0

10

20

30

40

50

60

1 3 5 7 9 11 13 15 17 19

Time (sec)

Ba

nd

wid

th (

Mb

ps

)

EF

AF4

AF1

BE

ClassClass Time of Time of StartStart

Requested BW (Mbps)Requested BW (Mbps) Minimum Guaranteed BW (Mbps)Minimum Guaranteed BW (Mbps)

EF (UDP)EF (UDP) 00 5050 3030

AF4 (UDP)AF4 (UDP) 99 4040 2525

AF1 (UDP)AF1 (UDP) 55 4040 2020

BE (UDP)BE (UDP) 1414 4040 NoneNone


MPLS TESTBED MPLS TESTBED NETWORK TOPOLOGYNETWORK TOPOLOGY

rtr3

rtr1 rtr2

Abilene

7505

7200 7200

FastEthernet

Gigabit

Gigabit

Gigabit

MPLS Tunnel 1

MPLS Tunnel 3MPLS Tunnel 2

LAN 1 LAN 2


MPLS TE MPLS TE EXPERIMENTSEXPERIMENTS

Goal: Goal: Evaluate the benefits of MPLS TEEvaluate the benefits of MPLS TE– Case Study 1: Case Study 1: Traditional IP Network (Min Hop Traditional IP Network (Min Hop rtr1-rtr2) rtr1-rtr2)

– Best-effort service onlyBest-effort service only– Two 40 Mbps UDP flows are sent from rtr1 to rtr2Two 40 Mbps UDP flows are sent from rtr1 to rtr2– Two 100 Mbps TCP flows are sent from rtr1 to rtr2Two 100 Mbps TCP flows are sent from rtr1 to rtr2– All flows take the min-hop path (FastEthernet) and are All flows take the min-hop path (FastEthernet) and are

limited to a total of 100 Mbps. UDP starves the TCP flows.limited to a total of 100 Mbps. UDP starves the TCP flows.

0

5

10

15

20

25

30

35

40

45

50

1 3 5 7 9

11 13 15 17 19

Time (sec)

Ban

dwid

th (M

bps) TCP1

TCP2

UDP1

UDP2



Goal: Goal: Evaluate the MPLS TE PropertiesEvaluate the MPLS TE Properties– Case Study 2: MPLS Network – Mixed FlowsCase Study 2: MPLS Network – Mixed Flows

– 3 MPLS tunnels were set up.3 MPLS tunnels were set up.– Two 40 Mbps UDP flows sent from rtr1 to rtr2Two 40 Mbps UDP flows sent from rtr1 to rtr2– Two 100 Mbps TCP flows sent from rtr1 to rtr2Two 100 Mbps TCP flows sent from rtr1 to rtr2– Tunnel1: UDP1 + TCP1; Tunnel2: TCP2; Tunnel3: UDP2Tunnel1: UDP1 + TCP1; Tunnel2: TCP2; Tunnel3: UDP2– TCP1 reduces rate when UDP1 arrives due to BW TCP1 reduces rate when UDP1 arrives due to BW

contentioncontention

0

10

20

30

40

50

60

70

80

90

100

1 3 5 7 9 11 13 15 17 19

UDP1

TCP1

TCP2

UDP2



Goal: Goal: Evaluate the MPLS TE propertiesEvaluate the MPLS TE properties– Case Study 3: MPLS Network – Separate FlowsCase Study 3: MPLS Network – Separate Flows

– 3 MPLS tunnels 3 MPLS tunnels – Two 40Mbps UDP flows sent from rtr1 to rtr2Two 40Mbps UDP flows sent from rtr1 to rtr2– Two 100 Mbps TCP flows sent from rtr1 to rtr2Two 100 Mbps TCP flows sent from rtr1 to rtr2– Tunnel1: TCP1; Tunnel2: TCP2; Tunnel3: UDP1 + UDP2Tunnel1: TCP1; Tunnel2: TCP2; Tunnel3: UDP1 + UDP2– No interference between TCP and UDPNo interference between TCP and UDP

0

10

20

30

40

50

60

70

80

90

100

1 3 5 7 9

11

13

15

17

19

TCP1

TCP2

UDP1

UDP2


EXPERIMENTAL EXPERIMENTAL CONCLUSIONSCONCLUSIONS

The MPLS TE provides better The MPLS TE provides better resource resource utilization and throughpututilization and throughput

– Cisco’s MPLS tunnels implementation Cisco’s MPLS tunnels implementation does not enforce the limit on the does not enforce the limit on the tunnel reserved bandwidth – needs tunnel reserved bandwidth – needs improvementimprovement

– CAR policing is not implementable on CAR policing is not implementable on Tunnel interfacesTunnel interfaces


DiffServ-AWARE TEDiffServ-AWARE TEPREEMPTION PREEMPTION EXPERIMENTSEXPERIMENTS

Goal: Goal: To evaluate Cisco’s preemption policyTo evaluate Cisco’s preemption policy– 3 MPLS tunnels were setup between routers, sharing 3 MPLS tunnels were setup between routers, sharing

a FastEthernet linka FastEthernet link– Tunnels 1, 2, and 3 together require the total link Tunnels 1, 2, and 3 together require the total link

bandwidthbandwidth– A new bandwidth request arrives for Tunnel 4, which A new bandwidth request arrives for Tunnel 4, which

has higher priority than the other 3 tunnelshas higher priority than the other 3 tunnels– One of the previously established tunnels must be One of the previously established tunnels must be

preempted. Which one?preempted. Which one?

New TunnelNew Tunnel


201Tunnel 4

206Tunnel 3

307Tunnel 2

X507Tunnel 1

PreemptedBandwidthPriorityTunnel

Oldest TunnelOldest Tunnel

201Tunnel 4

206Tunnel 3

507Tunnel 2

X307Tunnel 1

PreemptedBandwidthPriorityTunnel

Lowest PriorityLowest Priority

Cisco’s preemption policy:Cisco’s preemption policy:– Tunnel priority: Tunnel priority: lowest priority (numerically higher)lowest priority (numerically higher)– Tunnel age: Tunnel age: Tunnel created earliestTunnel created earliest

DiffServ-AWARE TEDiffServ-AWARE TEPREEMPTION PREEMPTION EXPERIMENTSEXPERIMENTS


EXPERIMENTAL EXPERIMENTAL CONCLUSIONSCONCLUSIONS

DiffServ-aware TE support in Cisco’s IOS DiffServ-aware TE support in Cisco’s IOS is not completely deployedis not completely deployed– Preemption is purely based on tunnel Preemption is purely based on tunnel

priority and age – waste of resourcespriority and age – waste of resources


TEAM/AA Architecture:TEAM/AA Architecture:Managing Multiple Domain Managing Multiple Domain

DiffServ MPLS NetworksDiffServ MPLS Networks

Research Contributions and IssuesResearch Contributions and Issues


RELATED WORKRELATED WORK RATES (Routing and Traffic Engineering Server)RATES (Routing and Traffic Engineering Server)

– Software by Bell Labs for MPLS Traffic Engineering (TE)Software by Bell Labs for MPLS Traffic Engineering (TE)– Uses Common Open Policy Service (COPS) and Minimum Uses Common Open Policy Service (COPS) and Minimum

Interference Routing Algorithm (MIRA) Interference Routing Algorithm (MIRA) – It achieves TE by routing of bandwidth guaranteed LSPs It achieves TE by routing of bandwidth guaranteed LSPs

TEQUILA (TE for QoS in the Internet at Large Scale)TEQUILA (TE for QoS in the Internet at Large Scale)– European research project for end-to-end QoS in European research project for end-to-end QoS in

DiffServ networkDiffServ network– Components for monitoring, TE, SLS management, and Components for monitoring, TE, SLS management, and

policy managementpolicy management– Algorithms and techniques are not concretely defined Algorithms and techniques are not concretely defined

yet and their quantitative evaluation has not been yet and their quantitative evaluation has not been carried outcarried out

MATE (Multipath Adaptive Traffic Engineering)MATE (Multipath Adaptive Traffic Engineering)– Software by Bell Labs for MPLS TESoftware by Bell Labs for MPLS TE– Assumes LSP layout using a long term traffic matrix. The Assumes LSP layout using a long term traffic matrix. The

focus is on load balancing short term traffic fluctuationsfocus is on load balancing short term traffic fluctuations– Not designed for bandwidth guaranteed servicesNot designed for bandwidth guaranteed services


To NeighboringTEAM

DiffServ/MPLSDomain

To NeighboringDomain

NetworkPlanning

TRAFFIC ENGINEERING TRAFFIC ENGINEERING AUTOMATED MANAGER (TEAM)AUTOMATED MANAGER (TEAM)“Design and Management Tools for an MPLS Domain QoS “Design and Management Tools for an MPLS Domain QoS Manager,”Manager,” to appear in Proceedings of SPIE ITCOM 2002, to appear in Proceedings of SPIE ITCOM 2002, Boston, August 2002.Boston, August 2002.

Rou

teR

esou

rce

Management Plane

Mob

ilit

yLSP Routing

Traffic Routing

LSP Preemption

LSP Setup/Dimensioning

LSP Capacity Allocation

LocationManagement

HandoffManagement

Intra-domain operationIntra-domain operation

TEAM


TEAM/AATEAM/AAARCHITECTUREARCHITECTURE

TEAM TEAMTEAM

AA

AA

AA

AA

AA

AA

Inter-domain operationInter-domain operation


TEAM/AA TEAM/AA ARCHITECTUREARCHITECTURE

Traffic Engineering Automated Traffic Engineering Automated Manager (TEAM) and Adaptive Manager (TEAM) and Adaptive Agent (AA):Agent (AA):– Manage heterogeneous networksManage heterogeneous networks

– Different services such as best-effort, real-Different services such as best-effort, real-time, etc.time, etc.

– Different network technologies such as Different network technologies such as wired and wireless mobile networkswired and wireless mobile networks

– Manage large networksManage large networks– Multiple domainsMultiple domains


TEAM COMPONENTSTEAM COMPONENTS Traffic Engineering Tool Traffic Engineering Tool [5][5]

– Resource ManagementResource Management

– Optimal Policy for LSP Setup [1, Optimal Policy for LSP Setup [1, 2]2]

– Adaptive preemption policy for LSPs [3]Adaptive preemption policy for LSPs [3]– Traffic estimation and resource allocation Traffic estimation and resource allocation

scheme [4,6]scheme [4,6]


RESOURCE MANAGEMENT – LSP RESOURCE MANAGEMENT – LSP SETUPSETUP“A New Threshold-Based Policy for Label Switched Path Setup in MPLS Networks,”“A New Threshold-Based Policy for Label Switched Path Setup in MPLS Networks,” in proceedings of ITC 2001, Salvador da Bahia, Brazil, pp. 1-11, December 2001. in proceedings of ITC 2001, Salvador da Bahia, Brazil, pp. 1-11, December 2001.“Optimal Policy for Label Switched Path Setup in MPLS Networks,”“Optimal Policy for Label Switched Path Setup in MPLS Networks,” accepted for publication in Computer Networks Journal, 2002.accepted for publication in Computer Networks Journal, 2002.

Based on Markov Decision Process theory.Based on Markov Decision Process theory. Objective Function: Objective Function:

– Minimize the expected infinite-horizon Minimize the expected infinite-horizon discounted total cost.discounted total cost.

To determine the optimal policy, To determine the optimal policy, the the transition probabilities and the optimality transition probabilities and the optimality equationsequations

The optimality equations are solved using The optimality equations are solved using the Value Iteration Algorithm.the Value Iteration Algorithm.

Determine an Determine an Adaptive Traffic DrivenAdaptive Traffic Driven PolicyPolicy for for LSP Setup and Dimensioning for each LSP Setup and Dimensioning for each MPLS Network.MPLS Network.


OPTIMIZATION OPTIMIZATION PROBLEMPROBLEM

Optimal policy Optimal policy * such that* such that

Optimality equationOptimality equation

wherewhere

),(),(),(),(

aSwaSwaSWaSr swb

sign


TEAM COMPONENTSTEAM COMPONENTS

Traffic Engineering Tool Traffic Engineering Tool [5][5]

– Resource ManagementResource Management– Optimal Policy for LSP Setup [1, 2]Optimal Policy for LSP Setup [1, 2]

– Adaptive Preemption Policy for Adaptive Preemption Policy for LSPs [3]LSPs [3]

– Traffic estimation and resource allocation Traffic estimation and resource allocation scheme [4,6]scheme [4,6]


RESOURCE MANAGEMENT – RESOURCE MANAGEMENT – LSP PREEMPTIONLSP PREEMPTION“A New Preemption Policy for DiffServ-Aware Traffic Engineering to Minimize “A New Preemption Policy for DiffServ-Aware Traffic Engineering to Minimize Rerouting,”Rerouting,” to appear in Proceedings of IEEE INFOCOM 2002, New York City, June to appear in Proceedings of IEEE INFOCOM 2002, New York City, June 2002.2002.

Non-real time applications may afford Non-real time applications may afford to have their transmission rate to have their transmission rate reduced.reduced.

By reducing the rate in a fair fashion:By reducing the rate in a fair fashion:– These LSPs would not be torn down,These LSPs would not be torn down,– There would be no service disruption, There would be no service disruption,

extra setup and tear down signalingextra setup and tear down signaling– THERE WOULD BE NO REROUTING THERE WOULD BE NO REROUTING

DECISIONS DECISIONS


ADAPTIVE PREEMPTION ADAPTIVE PREEMPTION POLICY (Contd.)POLICY (Contd.)

Combines the three main Combines the three main preemption criteria:preemption criteria:– Priority of preempted LSPsPriority of preempted LSPs– Number of preempted LSPsNumber of preempted LSPs– Bandwidth of preempted LSPsBandwidth of preempted LSPs

Optimization formulation and Optimization formulation and heuristicheuristic


ADAPTIVE PREEMPTION ADAPTIVE PREEMPTION POLICY – OPTIMIZATION POLICY – OPTIMIZATION FORMULATIONFORMULATION

Minimize:Minimize:

F = F = (priority cost) + (priority cost) + (number of LSPs) (number of LSPs) + + (preempted bandwidth) + BW module cost (preempted bandwidth) + BW module cost

Subject to:Subject to:– Number of preempted modules Number of preempted modules r r– Number of preempted modules in a preempted Number of preempted modules in a preempted

LSP is equal to total number of modules in the LSP is equal to total number of modules in the LSP.LSP.

– Number of preempted modules in a rate reduced Number of preempted modules in a rate reduced LSP is less than LSP is less than % of the total number of % of the total number of modules in the LSP.modules in the LSP.


PERFORMANCE COMPARISON:PERFORMANCE COMPARISON:COMMERCIAL VERSUS ADAPTIVE POLICYCOMMERCIAL VERSUS ADAPTIVE POLICY

Requested bw

Mbps

Commercial Policy

Adaptive Policy

15 1 0

25 1 0

30 2 0

40 2 0

55 2 0

65 2 0

90 2 1

100 2 2

155 6 2

185 7 3

240 7 4

280 8 5

325 10 6

Requested bw

Mbps

Commercial Policy

Adaptive Policy

15 10 0

25 0 0

30 70 0

40 60 0

55 45 0

65 35 0

90 10 0

100 0 0

155 16 0

185 71 0

240 16 0

280 1 0

325 66 0

Number of Preempted LSPs Bandwidth Wastage


TEAM COMPONENTSTEAM COMPONENTS

Traffic Engineering Tool Traffic Engineering Tool [5][5]

– Resource ManagementResource Management– Optimal Policy for LSP Setup [1, 2]Optimal Policy for LSP Setup [1, 2]

– Adaptive Preemption Policy for LSPs [3]Adaptive Preemption Policy for LSPs [3]

– Traffic Estimation and Resource Traffic Estimation and Resource Allocation Scheme [4,6]Allocation Scheme [4,6]


RESOURCE MANAGEMENT –RESOURCE MANAGEMENT –LSP BANDWIDTH LSP BANDWIDTH ALLOCATIONALLOCATION

Simple method is over-provisioning or Simple method is over-provisioning or cushioncushion

New Method New Method based on Kalman filter for based on Kalman filter for optimal estimation of the traffic and optimal estimation of the traffic and capacity prediction by determining capacity prediction by determining transition probabilitiestransition probabilities

A method to determine A method to determine Bandwidth AllocationBandwidth Allocation for for LSPsLSPs with with less bandwidth wastageless bandwidth wastage and and less less re-dimensioningre-dimensioning in an in an MPLS Network.MPLS Network.


To NeighboringTEAM

DiffServ/MPLSDomain

To NeighboringDomain

NetworkPlanning

TRAFFIC ENGINEERING TRAFFIC ENGINEERING AUTOMATED MANAGER (TEAM)AUTOMATED MANAGER (TEAM)“Design and Management Tools for an MPLS Domain QoS “Design and Management Tools for an MPLS Domain QoS Manager,”Manager,” to appear in Proceedings of SPIE ITCOM 2002, to appear in Proceedings of SPIE ITCOM 2002, Boston, August 2002.Boston, August 2002.

Rou

teR

esou

rce

Management Plane

Mob

ilit

yLSP Routing

Traffic Routing

LSP Preemption

LSP Setup/Dimensioning

LSP Capacity Allocation

LocationManagement

HandoffManagement

Intra-domain operationIntra-domain operation

TEAM


TEAM/AATEAM/AAARCHITECTUREARCHITECTURE

TEAM TEAMTEAM

AA

AA

AA

AA

AA

AA

Inter-domain operationInter-domain operation


PUBLICATIONSPUBLICATIONS[1] C. Scoglio, T. Anjali, J. de Oliveira, I. Akyildiz, and G. Uhl, [1] C. Scoglio, T. Anjali, J. de Oliveira, I. Akyildiz, and G. Uhl,

“A New Threshold-Based Policy for Label Switched Path Setup in “A New Threshold-Based Policy for Label Switched Path Setup in MPLS Networks,”MPLS Networks,” in proceedings of ITC 2001, Salvador da Bahia, in proceedings of ITC 2001, Salvador da Bahia, Brazil, pp. 1-11, December 2001.Brazil, pp. 1-11, December 2001.

[2] T. Anjali, C. Scoglio, J. de Oliveira, I. Akyildiz, and G. Uhl, [2] T. Anjali, C. Scoglio, J. de Oliveira, I. Akyildiz, and G. Uhl, “Optimal “Optimal Policy for Label Switched Path Setup in MPLS Networks,”Policy for Label Switched Path Setup in MPLS Networks,” accepted accepted for publication in Computer Networks Journal, 2002.for publication in Computer Networks Journal, 2002.

[3] J. de Oliveira, C. Scoglio, I. Akyildiz, and G. Uhl, [3] J. de Oliveira, C. Scoglio, I. Akyildiz, and G. Uhl, “A New Preemption “A New Preemption Policy for DiffServ-Aware Traffic Engineering to Minimize Policy for DiffServ-Aware Traffic Engineering to Minimize Rerouting,”Rerouting,” to appear in proceedings of IEEE INFOCOM 2002, New to appear in proceedings of IEEE INFOCOM 2002, New York City, June 2002.York City, June 2002.

[4] C. Bruni, C. Scoglio, and S. Vergari, [4] C. Bruni, C. Scoglio, and S. Vergari, “Optimal Capacity Provisioning “Optimal Capacity Provisioning for Label Switched Paths in MPLS Networks,” for Label Switched Paths in MPLS Networks,” to appear in to appear in proceedings of IFIP-TC6 Networking 2002, Pisa, Italy, May 2002.proceedings of IFIP-TC6 Networking 2002, Pisa, Italy, May 2002.

[5] J. de Oliveira, C. Scoglio, T. Anjali, L. Chen, I. Akyildiz, and G. Uhl, [5] J. de Oliveira, C. Scoglio, T. Anjali, L. Chen, I. Akyildiz, and G. Uhl, “Design and Management Tools for an MPLS Domain QoS Manager,”“Design and Management Tools for an MPLS Domain QoS Manager,” to appear in Proceedings of SPIE ITCOM 2002, Boston, August 2002.to appear in Proceedings of SPIE ITCOM 2002, Boston, August 2002.


PUBLICATIONS PUBLICATIONS (Contd.)(Contd.)

[6] T. Anjali, C. Scoglio, I. Akyildiz, and G. Uhl, [6] T. Anjali, C. Scoglio, I. Akyildiz, and G. Uhl, “A New “A New Scheme for Traffic Estimation and Resource Allocation Scheme for Traffic Estimation and Resource Allocation for Bandwidth Brokers,”for Bandwidth Brokers,” submitted for publication, submitted for publication, 2002.2002.

[7] T. Anjali, C. Scoglio, L. Chen, I. Akyildiz, and G. Uhl, [7] T. Anjali, C. Scoglio, L. Chen, I. Akyildiz, and G. Uhl, “ABEst: An Available Bandwidth Estimator within an “ABEst: An Available Bandwidth Estimator within an Autonomous System,”Autonomous System,” submitted for publication, 2002. submitted for publication, 2002.

[8] J. de Oliveira, F. Martinelli, and C. Scoglio, [8] J. de Oliveira, F. Martinelli, and C. Scoglio, “SPeCRA: A “SPeCRA: A Stochastic Performance Comparison Routing Algorithm Stochastic Performance Comparison Routing Algorithm for LSP Setup in MPLS Networks,” for LSP Setup in MPLS Networks,” submitted for submitted for publication, 2002.publication, 2002.

[9] J. L. Marzo, E. Calle, C. Scoglio, and T. Anjali, [9] J. L. Marzo, E. Calle, C. Scoglio, and T. Anjali, “Adding “Adding QoS Protection in Order to Enhance MPLS QoS QoS Protection in Order to Enhance MPLS QoS Routing,”Routing,” submitted for publication, 2002. submitted for publication, 2002.

Documents

Broadband and Wireless Networking Laboratory School of Electrical and Computer Engineering