Energy-Aware Traffic Engineering For Wired IP Networks
UNIVERSITÉ DE MONTRÉALENERGY-AWARE TRAFFIC ENGINEERING FOR WIRED IP NETWORKSLUCA GIOVANNI GIANOLIDÉPARTEMENT DE GÉNIE ÉLECTRIQUEÉCOLE POLYTECHNIQUE DE MONTRÉALTHÈSE PRÉSENTÉE EN VUE DE L’OBTENTIONDU DIPLÔME DE PHILOSOPHIÆ DOCTOR(GÉNIE ÉLECTRIQUE)JUILLET 2014c Luca Giovanni Gianoli, 2014.
UNIVERSITÉ DE MONTRÉALÉCOLE POLYTECHNIQUE DE MONTRÉALCette thèse intitulée :ENERGY-AWARE TRAFFIC ENGINEERING FOR WIRED IP NETWORKSprésentée par : GIANOLI Luca Giovannien vue de l’obtention du diplôme de : Philosophiæ Doctora été dûment acceptée par le jury d’examen constitué de :M. GIRARD André, Ph.D., présidentMme SANSÓ Brunilde, Ph.D., membre et directrice de rechercheM. CAPONE Antonio, Ph.D., membre et codirecteur de rechercheMme CARELLO Giuliana, Ph.D., membreM. KLEIN Thierry, Ph.D., membre
iiiTo my family, to Hélène,and to all the friendswho supported me along these years. . .
ivACKNOWLEDGEMENTSThere is a long list of people who deserve my most sincere thanks for their precioussupport along these four years. First of all, a special thanks goes to my supervisors, Antonioand Brunilde, who, in 2010, gave me the opportunity to start the Ph.D, come to Montrealand live this amazing adventure. I wanna thank them for their help, for their suggestionsand for the attention that they always demonstrated in my regards. It has been a pleasureand an honour to work under their supervision.My deepest gratitude goes also to Edoardo, Giuliana, Bernardetta, Stefano C., Carmeloand Erick, who taught me so many valuable things and whose contribution has been fundamental along these four years.A special mention is also for Dr. A. Girard and Dr. T. Klein, who spent time an efforts toevaluate this thesis and gave me very precious suggestions to improve it. A particular thanksalso to Dr. J.C. Grégoire, who was member of the commission for the preliminary exam andwhose advice had been very useful to develop the research project.I wanna thank all the guys of the labs in both Montreal and Milano, i.e., Silvia, whowas the first to welcome me to Montreal and teach me everything about the city, Antimo,Alessandro, Arash, Stefano P., Federico, Ilario, Hadhami, Marnie, Michele and Carmelo(again!). My most sincere thanks for their precious companionship, for their support and forall the experiences that we shared along these years.A special appreciation is for all the personnel of École Polytechnique de Montréal andPolitecnico di Milano. In particular I would like to thank Nathalie Lévesque and NadiaPrada, who helped me out through all the bureaucratic requirements. Many thanks to thepersonnel of GERAD, too, and, in particular, to Marie Perrault, Francine Benoit, CarolDufour, Marilyne Lavoie, and Pierre Girard.Obviously, I cannot forget my parents, Andrea and Marina, who have never failed tosupport, and who I always felt very close to even when an entire ocean was among us (thanksSkype!): thanks for letting me go and find my own way in the world.A very special thanks is for Hélène, who left everything she had in Europe to accompanyme to Canada and live this adventure by my side: she is the only person who really knowshow hard I worked to get this Ph.D., she has been always there for me.I will never forget Roberto and all the guys from Tennis Canada, who have been alwaysso kind and who gave me the possibility to play tennis here in Motreal too. during the moststressful periods of a Ph.D., a tennis racket in your hands and a yellow ball to hit are thebest remedy!
vMy deepest thanks to my friends Nicolò, who taught me how to persist, Jacopo, whotaught me the true values of life and where to buy the best home-made cakes in the world,Filippo, who forced me to go out beyond a radius of four-hundred meters from my place,Alejandro, my personal tequila supplier and great football player, and Nicola, a very inspiringmaster of life.The final thanks are for all the other friends of both Milano and Montreal, with a specialmention which goes to Johnny, who crossed the ocean to visit me in Montreal without myknowing it, and Alessandro, whose terrific backhand kept me well-trained along my stays inMilano. Please, forgive me if I’ve forgotten anybody!
viRÉSUMÉMême si l’Internet est souvent considéré comme un moyen formidable pour réduire l’impactdes activités humaines sur l’environnement, sa consommation d’énergie est en train de devenir un problème en raison de la croissance exponentielle du trafic et de l’expansion rapidedes infrastructures de communication dans le monde entier. En 2007, il a été estimé que leséquipements de réseau (sans tenir compte de serveurs dans les centres de données) étaientresponsables d’une consommation d’énergie de 22 GW, alors qu’en 2010 la consommationannuelle des plus grands fournisseurs de services Internet (par exemple AT&T) a dépassé 10TWh par an.En raison de cette tendance alarmante, la réduction de la consommation d’énergie dans lesréseaux de télécommunication, et en particulier dans les réseaux IP, est récemment devenueune priorité. Une des stratégies les plus prometteuses pour rendre plus vert l’Internet est lesleep-based energy-aware network management (SEANM), selon lequel la configuration deréseau est adaptée aux niveaux de trafic afin d’endormir tous les éléments redondantes duréseau.Dans cette thèse nous développons plusieurs approches centralisées de SEANM, afind’optimiser la configuration de réseaux IP qui utilisent différents protocoles (open shortest path first (OSPF) or multi protocol Label Switching (MPLS)) ou transportent différentstypes de trafic (élastique or inélastique). Le choix d’adresser le problème d’une manière centralisée, avec une plate-forme de gestion unique qui est responsable de la configuration et dela surveillance de l’ensemble du réseau, est motivée par la nécessité d’opérateurs de mainteniren tout temps le contrôle complet sur le réseau.Visant à mettre en œuvre les approches proposées dans un environnement réaliste duréseau, nous présentons aussi un nouveau cadre de gestion de réseau entièrement configurableque nous avons appelé JNetMan. JNetMan a été exploité pour tester une version dynamiquede la procédure SEANM développée pour les réseaux utilisant OSPF.
viiABSTRACTEven if the Internet is commonly considered a formidable means to reduce the impact ofhuman activities on the environment, its energy consumption is rapidly becoming an issue dueto the exponential traffic growth and the rapid expansion of communication infrastructuresworldwide. Estimated consumption of the network equipment, excluding servers in data centers, in 2007 was 22 GW, while in 2010 the yearly consumption of the largest Internet ServiceProviders, e.g., AT&T, exceeded 10 TWh per year. The growing energy trend has motivatedthe development of new strategies to reduce the consumption of telecommunication networks,with particular focus on IP networks. In addition to the development of a new generation ofgreen network equipment, a second possible strategy to optimize the IP network consumption is represented by sleep-based energy-aware network management (SEANM), which aimsat adapting the whole network power consumption to the traffic levels by optimizing thenetwork configuration and putting to sleep the redundant network elements. Device sleepingrepresents the main potential source of saving because the consumption of current networkdevices is not proportional to the utilization level: so that, the overall network consumptionis constantly close to maximum. In current IP networks, quality of service (QoS) and networkresilience to failures are typically guaranteed by substantially over-dimensioning the wholenetwork infrastructure: therefore, also during peak hours, it could be possible to put to sleepa non-negligible subset of redundant network devices.Due to the heterogeneity of current network technologies, in this thesis, we focus ourefforts to develop centralized SEANM approaches for IP networks operated with differentconfigurations and protocols. More precisely, we consider networks operated with differentrouting schemes, namely shortest path (OSPF), flow-based (MPLS) and take into accountdifferent types of traffic, i.e., elastic or inelastic. The centralized approach, with a single management platform responsible for configuring and monitoring the whole network, is motivatedby the need of network operators to be constantly in control of the network dynamics. Tofully guarantee network stability, we investigate the impact of SEANM on network reliabilityto failures and robustness to traffic variations. Ad hoc modeling techniques are integratedwithin the proposed SEANM frameworks to explicitly consider resilience and robustness asnetwork constraints. Finally, to implement the proposed procedures in a realistic networkenvironment, we propose a novel, fully configurable network management framework, calledJNetMan. We use JNetMan to develop and test a dynamic version of the SEANM procedurefor IP networks operated with shortest path routing protocols.
viiiTABLE OF CONTENTSDEDICATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iiiACKNOWLEDGEMENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .ivRÉSUMÉ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .viABSTRACT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viiTABLE OF CONTENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viiiLIST OF TABLES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiiLIST OF FIGURES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xivLIST OF ANNEXES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xviLIST OF ACRONYMS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xviiLIST OF SYMBOLS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xixCHAPTER 1 Introduction . . . . . . . . . . . . . . .1.1 Definitions and Basic Concepts . . . . . . . .1.2 Elements of the Problem . . . . . . . . . . . .1.2.1 Routing Protocols . . . . . . . . . . . .1.2.2 QoS Constraints . . . . . . . . . . . .1.2.3 Optimization Frequency . . . . . . . .1.2.4 Decision Points . . . . . . . . . . . . .1.2.5 Network Survivability and Robustness1.3 Research Goals . . . . . . . . . . . . . . . . .1.3.1 Document Structure . . . . . . . . . .CHAPTER 2 Green Networking in IP Networks: an2.1 Power Models . . . . . . . . . . . . . . . . .2.2 Green Networking . . . . . . . . . . . . . . .2.3 Energy-Aware Network Optimization . . . .2.3.1 Problem Sets and Parameters . . . .1133456667Overview. . . . . . . . . . . . . . . . . . . . . 8. 8. 9. 11. 11
ix.1212141717192123252728293132323333CHAPTER 3 SEANM with Flow-Based Routing . . . . . . . . . . . . . . . .3.1 Our Approach for Energy-Aware Multi-Period Network Management3.2 Two Exact MILP Formulations . . . . . . . . . . . . . . . . . . . . .3.2.1 Power Aware Fixed Routing Problem (PAFRP) . . . . . . . .3.2.2 Power Aware Variable Routing Problem (PAVRP) . . . . . . .3.3 Heuristic Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . .3.3.1 A Quick Heuristic Algorithm . . . . . . . . . . . . . . . . . . .3.3.2 A Single-Period Heuristic . . . . . . . . . . . . . . . . . . . .3.4 Computational Results . . . . . . . . . . . . . . . . . . . . . . . . . .3.4.1 Test Instances . . . . . . . . . . . . . . . . . . . . . . . . . . .3.4.2 PAFRP and PAVRP Results . . . . . . . . . . . . . . . . . . .3.4.3 EA-LG . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3.4.4 Economic Evaluation . . . . . . . . . . . . . . . . . . . . . . .3.4.5 Single Period Heuristic . . . . . . . . . . . . . . . . . . . . . .3.4.6 Evaluation with Real Traces . . . . . . . . . . . . . . . . . . .3.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2 Problem Variables . . . . . . . . . . . . . . . . . . . . . . .Energy-Aware Network Management with Flow-Based Routing . . .2.4.1 State of the Art for SEANM-FB . . . . . . . . . . . . . . . .2.4.2 Our Contribution . . . . . . . . . . . . . . . . . . . . . . . .Energy-Aware Network Management with Shortest Path Routing .2.5.1 State of the Art for SEANM-SP . . . . . . . . . . . . . . . .2.5.2 Our Contribution . . . . . . . . . . . . . . . . . . . . . . . .Energy Minimization and Network Survivability . . . . . . . . . . .2.6.1 State of the Art on SEANM with Survivability Requirements2.6.2 Our Contribution . . . . . . . . . . . . . . . . . . . . . . . .Energy-Aware Network Management with Elastic Traffic . . . . . .2.7.1 State of the Art on the Management of Elastic Traffic . . . .2.7.2 Our Contribution . . . . . . . . . . . . . . . . . . . . . . . .Other Approaches . . . . . . . . . . . . . . . . . . . . . . . . . . . .2.8.1 Heuristic Algorithms . . . . . . . . . . . . . . . . . . . . . .2.8.2 Green Protocols . . . . . . . . . . . . . . . . . . . . . . . . .2.8.3 Optical Networks . . . . . . . . . . . . . . . . . . . . . . . .
xCHAPTER 4 On Robustness and Survivability in SEANM with Flow-based Routing .4.1 Our Approach to Network Survivability . . . . . . . . . . . . . . . . . . . . . .4.1.1 Network Resilience . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4.1.2 Network Robustness . . . . . . . . . . . . . . . . . . . . . . . . . . . .4.1.3 A Visual Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4.2 MILP Formulations for SEANM-FB with Network Resilience . . . . . . . . . .4.2.1 Variable Routing with Dedicated Protection . . . . . . . . . . . . . . .4.2.2 Variable Routing with Shared Protection . . . . . . . . . . . . . . . . .4.2.3 Variable Routing with Smart Protection . . . . . . . . . . . . . . . . .4.3 A MILP Formulation for Robust SEANM-FB . . . . . . . . . . . . . . . . . .4.4 Heuristic Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4.4.1 Warm Starting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4.5 Computational Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4.5.1 Test Instances . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4.5.2 Savings vs. Protection/Robustness . . . . . . . . . . . . . . . . . . . .4.5.3 Larger Networks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4.6 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .CHAPTER 5 SEANM with Shortest Path Routing . . . . . . . . .5.1 Our Approach for Energy-Aware Link Weight Optimization5.2 MILP Formulation for SEANM-SP . . . . . . . . . . . . . .5.3 Heuristic Methods for SEANM-SP . . . . . . . . . . . . . .5.3.1 Congestion-Aware Link Weight Optimization . . . . .5.3.2 Greedy Algorithm for Energy Savings (GA-ES) . . .5.3.3 GRASP for Energy Savings (GRA-ES) . . . . . . . .5.3.4 Two-Stages Algorithm for Energy Savings (TA-ES) .5.3.5 MILP-EWO Algorithm . . . . . . . . . . . . . . . . .5.4 Computational Results . . . . . . . . . . . . . . . . . . . . .5.4.1 Test Instances . . . . . . . . . . . . . . . . . . . . . .5.4.2 Results . . . . . . . . . . . . . . . . . . . . . . . . . .5.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . 07108109111113113114120CHAPTER 6 Dynamic SEANM with Shortest Path Routing . . . . . . . . .6.1 Dynamic Energy-Aware OSPF Optimization . . . . . . . . . . . . . .6.1.1 OSPF Switching Policy . . . . . . . . . . . . . . . . . . . . . .6.2 Java Framework for SNMP-based Network Management Applications6.3 Computational Results . . . . . . . . . . . . . . . . . . . . . . . . . .123124130130135
xi6.46.3.1 Test Instances . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1356.3.2 Experimentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140CHAPTER 7 SEANM with Elastic Traffic . . . . . . . . . . . .7.1 A Novel Bi-Level Network Management Problem . . . .7.1.1 A General MILP for SEANM with Elastic Traffic7.2 Max-Min-Fairness . . . . . . . . . . . . . . . . . . . . . .7.2.1 Exact MILP Formulation . . . . . . . . . . . . . .7.3 Proportional-Fair Allocation . . . . . . . . . . . . . . . .7.3.1 Heuristic MILP for PF . . . . . . . . . . . . . . .7.4 A Visual Example . . . . . . . . . . . . . . . . . . . . . .7.5 Restricted Path Heuristic . . . . . . . . . . . . . . . . . .7.6 Computational Results . . . . . . . . . . . . . . . . . . .7.6.1 Test Instances . . . . . . . . . . . . . . . . . . . .7.6.2 Numerical Results . . . . . . . . . . . . . . . . .7.7 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . .143145147149150152152156157158158160163CHAPTER 8 Conclusions . . . . . . . . . .8.1 Summary of Work . . . . . . . . . .8.1.1 Flow-Based Routing . . . . .8.1.2 Network Survivability . . . .8.1.3 Shortest Path Routing . . . .8.1.4 Practical Implementation . . .8.1.5 Elastic Traffic . . . . . . . . .8.2 Limitations of the Proposed Solutions8.3 Future Developments . . . . . . . . .165165165166166167167168168.REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169ANNEXES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 186
xiiLIST OF 75.85.96.16.26.37.17.2Main parameters and acronyms . . . . . . . . . . . . . . . . . . .Router chassis and cards . . . . . . . . . . . . . . . . . . . . . . .Card status for the 9-node network. . . . . . . . . . . . . . . . . .Chassis status for the 9-node network. . . . . . . . . . . . . . . .Routing for the 9-node network. . . . . . . . . . . . . . . . . . . .Power consumption and congestion values for 9-node network . .Computational results: card switching analysis . . . . . . . . . . .Computational results: MILP formulation with SNDLib networksComputational results: EA-LG with france and nobel-eu . . . . .Computational results for germany50. . . . . . . . . . . . . . . . .Computational results: online EA-STH . . . . . . . . . . . . . . .Test instances. . . . . . . . . . . . . . . . . . . . . . . . . . . . .MILP model with robustness to traffic variations . . . . . . . . .MILP model with classic robust dedicated protection . . . . . . .Computational results: MILP model with and without protectionComparison between MILP and EA-STH for protected problems .Comparison between classic and smart protection . . . . . . . . .Traffic levels supported by dedicated and shared protection . . . .CPLEX time limits . . . . . . . . . . . . . . . . . . . . . . . . . .Computational results for E-TESP formulation . . . . . . . . . .Sorting criteria for greedy algorithm for energy saving (GA-ES) .Test network topologies . . . . . . . . . . . . . . . . . . . . . . .Computational results: dual weight warm-start effectiveness . . .Traffic matrix notation . . . . . . . . . . . . . . . . . . . . . . . .Computational results: energy savings in backbone networks . . .Computational results: sleeping elements in backbone networks .Computational results: computing times in backbone networks . .Computational results: energy savings in access networks . . . . .Test network topologies . . . . . . . . . . . . . . . . . . . . . . .OSPF configurations for experimentations . . . . . . . . . . . . .Switching policies . . . . . . . . . . . . . . . . . . . . . . . . . . .Router chassis and line Cards . . . . . . . . . . . . . . . . . . . .Network data. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5115116116116117135137137159159
xiiiTable 7.3Table 7.4Performance results with small networks . . . . . . . . . . . . . . . . . 163Performance results with large networks . . . . . . . . . . . . . . . . . 163
xivLIST OF .15.25.35.45.5Network with 9 nodes . . . . . . . . . . . . . . . . . . . . . . . . . . .France topology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .nobel-eu topology . . . . . . . . . . . . . . . . . . . . . . . . . . . . .germany50 topology . . . . . . . . . . . . . . . . . . . . . . . . . . . .Traffic scenarios . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .Computational results: hourly power consumption . . . . . . . . . . . .Computational results: overall energy savings . . . . . . . . . . . . . .Computational results: congestion values for 9-node instances . . . . .Computational results: card switching analysis . . . . . . . . . . . . . .Computational results: france network energy consumption . . . . . .Computational results: nobel-eu network energy consumption . . . . .Computational results: sleeping vs. fully proportional scenario . . . . .Computational results: gap between MILP and EA-LG solutions . . . .Computational results: gap between MILP and EA-STH solutions . . .Computational results with real traces . . . . . . . . . . . . . . . . . .Energy consumption minimization vs resilience requirements. . . . . . .Warm-start of multi-period MILP with shared protection. . . . . . . .Warm-start of EA-STH with shared protection. . . . . . . . . . . . . .Trade-off between power consumption and survivability. . . . . . . . . .Energy-cost of robustness . . . . . . . . . . . . . . . . . . . . . . . . .Impact analysis of the backup utilization threshold . . . . . . . . . . .Energy-efficiency/survivability trade-off in polska . . . . . . . . . . . .Energy-efficiency/survivability trade-off in nobel-germany . . . . . . .Energy-efficiency/survivability trade-off in nobel-eu . . . . . . . . . .Energy-efficiency/robustness trade-off in nobel-germany with EA-STH .Energy-efficiency/robustness trade-off in nobel-eu with EA-STH . . . .Impact analysis on path restriction . . . . . . . . . . . . . . . . . . . .Energy-efficiency/survivability trade-off in germany50 . . . . . . . . . .Piecewise-linear link saturation cost function . . . . . . . . . . . . . . .Centralized off-line approach for SEANM-SP . . . . . . . . . . . . . . .Two-stages algorithm for energy saving (TA-ES). . . . . . . . . . . . .MILP-EWO flow chart. . . . . . . . . . . . . . . . . . . . . . . . . . . .Computational results: comparison with state-of-art-procedures . . . 494959699101109112117
06.116.126.136.146.156.167.17.27.37.47.57.6Visual representation of MILP-EWO solutions . . . . . . . . . . . .Computational results: energy savings in backbone networks . . . .Computational results: computing times in backbone networks . . .Computational results: normalized congestion in backbone networksGeneral approach for dynamic SEANM-SP . . . . . . . . . . . . . .Sampling the traffic matrices from a daily traffic profile . . . . . . .Flow chart of our network management framework. . . . . . . . . .EANI architecture . . . . . . . . . . . . . . . . . . . . . . . . . . .OSPF configuration graph . . . . . . . . . . . . . . . . . . . . . . .OSPF configuration chain . . . . . . . . . . . . . . . . . . . . . . .Switching policy . . . . . . . . . . . . . . . . . . . . . . . . . . . .Architecture of JNetMan. . . . . . . . . . . . . . . . . . . . . . . .Example of network with 2 nodes and 1 link. . . . . . . . . . . . . .Topology definition through TAM . . . . . . . . . . . . . . . . . . .Use case for JNetMan’s APIs . . . . . . . . . . . . . . . . . . . . .Traffic profile used in tests . . . . . . . . . . . . . . . . . . . . . . .OCs chains used in tests. . . . . . . . . . . . . . . . . . . . . . . .Simulation results: OSPF configuration distribution . . . . . . . . .Simulation results: device consumption . . . . . . . . . . . . . . . .Simulation results: utilization plots . . . . . . . . . . . . . . . . . .Why elastic demands may appear as inelastic. . . . . . . . . . . . .SEANM with elastic traffic. . . . . . . . . . . . . . . . . . . . . . .Visual example of bi-level traffic engineering . . . . . . . . . . . . .Piecewise linear approximation of Objective function (7.43). . . . .Computational results: MILP for MMF SEANM-ET . . . . . . . .Computational results: restricted-path MILP for MMF 4135136137139139142144145157159162164
xviLIST OF ANNEXESAnnexe ANetwork congestion measure . . . . . . . . . . . . . . . . . . . . . . . . 186
xviiLIST OF on program interfaceautonomous systemborder gateway protocolenergy-aware network intelligenceenergy-aware lexicographic GRASPenergy-aware single time-period heuristicenergy-aware single time-period heuristic with restricted pathsenergy-aware network managementenergy-aware traffic engineeringequal cost multi-pathenergy-aware traffic engineering with shortest path routinggreedy algorithm for energy savingglobal greenhouse gasgreedy randomized adaptive search proceduregreedy randomized algorithm for energy savinginformation and communication technologyInternet engineering task forceinterior gateway protocol weight optimizationInternet protocolInternet service providerjava-based network management platformleast-flowleast-linklinear programminglow power idleLagrangian relaxationmulti-commodity network designmanagement information baseMILP-based algorithm for energy-aware weight optimizationmixed line ratesmax-min-fairnessnext-generation optical accessnetwork management platform
Mobject identifieropen shortest path firstopen systems interconnectionquality of servicepower-aware fixed routing problempower-aware variable routing problemproportional fairnesspoint of presencerestricted candidate listrandom early detectionrobust optimizationround-trip-timesleep-based energy-aware network managementSEANM with flow-based routingSEANM with shortest path routingSEANM with elastic trafficstructure of management informationsimple network management protocolsum-of-weightstwo-stage algorithm for energy savingtopology abstraction managertransport control protocoltraffic engineeringuser datagram protocolwavelength division multiplexing
xixLIST OF SYMBOLSSetsGVVeVcADDePdSHUijσdVoNetwork topologySet of network nodesSet of edge nodesSet of core nodesSet of unidirectional network linksSet of traffic demandsSet of elastic traffic demandsSet of paths for demand d DCircular set of time periodsSet of pieces of a piece-wise linear functionSet of demands which are uncertain during time period σ SSet of nodes from which the source node od of demand d De hasbeen ̄σijCiC̄iσµijµbckijπijOrigin of demand d DDestination of demand d DBandwidth request of demand d DBandwidth request of demand d D during time period σ SAverage bandwidth request of demand d D during time period σ S(robust variant)Worst case bandwidth request deviation of demand d D during timeperiod σ S (robust variant)Peak traffic value or nominal value of demand d DNumber of line cards installed on link (i, j) ACapacity of a line card installed on link (i, j) ACapacity already used on link (i, j) A during time period σ SCapacity of node i VCapacity already used on node i V during time period σ SMaximum link utilization allowed on link (i, j) AMaximum link utilization allowed on link (i, j) A during failure periodsPower required to keep activated a single line card of link (i, j) A
xxπiΠijΠiδσijdpωmaxhσηonEcM IN ENCong(i,j)σCong(i,j)αhβhΓij σmdΩΞgi dBPower required to keep activated node i VEnergy profile function of a single line card of link (i, j) AEnergy profile function of node i VPower required to switch on a chassis from the sleeping state, as afraction of the hourly chassis consumption πiBinary parameter, equal to 1 if link (i, j) A belongs to path p P dMaximum link weight allowedDuration of time period σ SMaximum number of card switching on allowed over the entire set ofperiods SEnergy budget to
universite de montr eal energy-aware traffic engineering for wired ip networks luca giovanni gianoli departement de g enie electrique ecole polytechnique de montr eal
Bruksanvisning för bilstereo . Bruksanvisning for bilstereo . Instrukcja obsługi samochodowego odtwarzacza stereo . Operating Instructions for Car Stereo . 610-104 . SV . Bruksanvisning i original
10 tips och tricks för att lyckas med ert sap-projekt 20 SAPSANYTT 2/2015 De flesta projektledare känner säkert till Cobb’s paradox. Martin Cobb verkade som CIO för sekretariatet för Treasury Board of Canada 1995 då han ställde frågan
service i Norge och Finland drivs inom ramen för ett enskilt företag (NRK. 1 och Yleisradio), fin ns det i Sverige tre: Ett för tv (Sveriges Television , SVT ), ett för radio (Sveriges Radio , SR ) och ett för utbildnings program (Sveriges Utbildningsradio, UR, vilket till följd av sin begränsade storlek inte återfinns bland de 25 största
Hotell För hotell anges de tre klasserna A/B, C och D. Det betyder att den "normala" standarden C är acceptabel men att motiven för en högre standard är starka. Ljudklass C motsvarar de tidigare normkraven för hotell, ljudklass A/B motsvarar kraven för moderna hotell med hög standard och ljudklass D kan användas vid
LÄS NOGGRANT FÖLJANDE VILLKOR FÖR APPLE DEVELOPER PROGRAM LICENCE . Apple Developer Program License Agreement Syfte Du vill använda Apple-mjukvara (enligt definitionen nedan) för att utveckla en eller flera Applikationer (enligt definitionen nedan) för Apple-märkta produkter. . Applikationer som utvecklas för iOS-produkter, Apple .
SA Learner Driver Manual Road Traffic Signs Version: Draft Page 1 of 56 2. ROAD TRAFFIC SIGNS, SIGNALS AND MARKINGS The purpose of road traffic signs is to regulate traffic in such a way that traffic flow and road traffic safety are promoted. 1. SIGNS IN GENERAL Road traffic signs can be divided into the following six main groups:
och krav. Maskinerna skriver ut upp till fyra tum breda etiketter med direkt termoteknik och termotransferteknik och är lämpliga för en lång rad användningsområden på vertikala marknader. TD-seriens professionella etikettskrivare för . skrivbordet. Brothers nya avancerade 4-tums etikettskrivare för skrivbordet är effektiva och enkla att
Den kanadensiska språkvetaren Jim Cummins har visat i sin forskning från år 1979 att det kan ta 1 till 3 år för att lära sig ett vardagsspråk och mellan 5 till 7 år för att behärska ett akademiskt språk.4 Han införde två begrepp för att beskriva elevernas språkliga kompetens: BI
