Distributed Queuing Dual Bus (DQDB) is a media access control (MAC) technique, which is being considered by the IEEE 802.6 for the metropolitan area network (MAN). The DQDB medium access technique has many promising advantages over other access methods. However it has one drawback, which is its unfairness in terms of node-dependent medium access delay. In this paper a mathematic model is formulated to describe this detrimental behavior of DQDB. The access control method is first modeled as a M/G/1 queueing system with a single priority level, then it is remodeled as a non-preemptive priority system with three priority levels. By employing these models, the approximate medium access delay analysis of a DQDB network is investigated for a metropolitan area network containing 50 stations with a channel bandwidth of 150 Mbps. Numerical results are then presented to illustrate the network unfairness performance under various traffic intensities and under different priority levels. The results have been obtained for non-isochronous (asynchronous) traffic. / Master of Science

No description available.

Approved for public release, distribution is unlimited / The performance of network management tools for SONET/SDH networks subject to the load conditions is studied and discussed in this thesis. Specifically, a SONET network which consists of four CISCO ONS 15454s, managed by a CISCO Transport Manager, is set up in the Advanced Network Laboratory of the Naval Postgraduate School. To simulate a realistic data transfer environment for the analysis, Smartbits Avalanche software is deployed to simulate multiple client-server scenarios in the SONET network. Traffic from the management channel is then captured using a packet sniffer. Queuing analysis on the captured data is performed with particular emphasis on properties of self-similarity. In particular, the Hurst parameter which determines the captured traffic's degree of self-similarity is estimated using the Variance-Index plot technique. Link utilization is also derived from the computation of first-order statistics of the captured traffic distribution. The study shows that less management data was exchanged when the SONET network was fully loaded. In addition, it is recommended that CTM 4.6 be used to manage not more than 1552 NEs for safe operation. The results presented in this thesis will aid network planners to optimize the management of their SONET/SDH networks. / Civilian, Ministry of Defense, Singapore

published_or_final_version / Business Administration / Master / Master of Business Administration

Previous research on ECN and RED usually considered only a limited traffic domain, focusing on networks with a small number of homogeneous flows. The behavior of RED and ECN congestion control mechanisms in TCP network with many competing heterogeneous flows in the bottleneck link, hasn't been sufficiently explored. This thesis first investigates the behavior and performance of RED with ECN congestion control mechanisms with many heterogeneous TCP Reno flows using the network simulation tool, ns-2. By comparing the simulated performance of RED and ECN routers, this study finds that ECN does provide better goodput and fairness than RED for heterogeneous flows. However, when the demand is held constant, the number of flows generating the demand has a negative effect on performance. Meanwhile, the simulations with many flows demonstrate that the bottleneck router's marking probability must be aggressively increased to provide good ECN performance. Based on these simulation results, an Adaptive ECN algorithm (AECN) was studied to further improve the goodput and fairness of ECN. AECN divides all flows competing for a bottleneck into three flow groups, and deploys a different max for each flow group. Meanwhile, AECN also adjusts min for the robust flow group and max to get higher performance when the number of flows grows large. Furthermore, AECN uses mark-front strategy, instead of mark-tail strategy in standard ECN. A series of AECN simulations were run in ns-2. The simulations show clearly that AECN treats each flow fairer than ECN with the two fairness measurements: Jain's fairness index and visual max-min fairness. AECN has fewer packet drops and alleviates the lockout phenomenon and yields higher goodput than ECN.

Parallèlement au prix, le délai de livraison est un facteur clé de compétitivité pour les entreprises. De plus les entreprises sont plus que jamais obligées de respecter ce délai promis. La combinaison du choix du prix et du délai promis implique de nouveaux compromis et offre de nombreuses perspectives. Un délai plus court peut entraîner une augmentation de la demande, mais augmente également le risque de livraison tardive et donc décourager les clients. A contrario un délai plus long ou un prix plus élevé entraîne généralement une baisse de la demande. Or malgré le rôle stratégique conjoint du prix et des délais et leurs impacts sur la demande, dans la littérature en gestion des opérations on suppose très généralement une demande exogène (fixée a priori) même si la conception de la chaîne impacte fortement les délais (localisation des sites, positionnement des stocks,..) et donc la demande. Nous nous sommes donc intéressés à ces choix de fixation des délais promis et du prix dans un contexte de demande endogène.La littérature traitant du choix du délai et du prix sous demande endogène a principalement considéré un contexte de fabrication à la commande (Make to Order). Un papier fondateur de Palaka et al en 1998 a présenté cette problématique avec une modélisation de l’entreprise par une file d’attente M/M/1 et nos travaux se placent dans la suite de ce travail. Notre revue de la littérature a permis d'identifier de nouvelles perspectives et nous proposons trois extensions dans cette thèse.Dans notre première contribution, en utilisant le cadre de Palaka et al, nous considérons que le coût de production est une fonction décroissante du délai. Dans tous les articles publiés dans ce contexte, le coût de production unitaire a été supposé constant. Pourtant en pratique, le coût de production unitaire dépend du délai promis, l'entreprise pouvant mieux gérer le processus de production et réduire les coûts de production en proposant des délais plus longs aux clients.Dans la deuxième contribution, nous considérons toujours le cadre de Palaka et al, mais modélisons l'entreprise comme une file d'attente M/M/1/K, pour laquelle la demande est donc rejetée s'il y a déjà K clients dans le système. Dans la littérature issue du travail de Palaka seule la file d'attente M/M/1 a été utilisée, ce qui signifie que tous les clients sont acceptés, ce qui peut entraîner de longues durées de séjour dans le système. Notre idée est basée sur le fait que rejeter certains clients, même si cela peut apparaitre dans un premier temps comme une perte de demande, pourrait aider à proposer un délai plus court pour les clients acceptés, et finalement conduire à une demande et donc un profit plus élevé.Dans la troisième contribution nous étudions un nouveau cadre pour le problème du délai et du prix en fonction de la demande endogène, en modélisant une chaîne logistique composée de deux étapes de production, modélisée par un réseau de files d’attente tandem (M/M/1-M/M/1). Dans la littérature avec ce cadre multi-entreprise, tous les articles ont considéré qu'un seul acteur avait des opérations de production, l'autre acteur ayant un délai nul. Nous avons étudié les scénarios centralisés et décentralisés.Pour chacun des nouveaux problèmes nous avons proposé des formulations maximisant le profit composé du revenu diminué des coûts de production, de stockage et pénalité de retard, et fourni des résolutions optimales, analytiques ou numériques. Ces résolutions nous ont amenés à démontrer de nouveaux résultats (retard moyen dans une M/M/1/K ; condition pour que des contraintes de service locales permettent d’assurer une contrainte de service globale dans un système en tandem). Nous avons mené des expériences numériques pour voir l’influence des différents paramètres. / Along with the price, the delivery lead time has become a key factor of competitiveness for companies and an important purchase criterion for many customers. Nowadays, firms are more than ever obliged to meet their quoted lead time, which is the delivery lead time announced to the customers. The combination of pricing and lead time quotation implies new trade-offs and offers opportunities for many insights. For instance, on the one hand, a shorter quoted lead time can lead to an increase in the demand but also increases the risk of late delivery and thus may affect the firm’s reputation and deter future customers. On the other hand, a longer quoted lead time or a higher price generally yields a lower demand. Despite the strategic role of joint pricing and lead time quotation decisions and their impacts on demand, in the operations management literature an exogenous demand (a priory a known demand) is generally used in supply chain models, even if the design of the supply chain has a strong impact on lead times (i.e., sites location, inventory position, etc.) and thus affects the demand. Therefore, we are interested in the lead time quotation and pricing decisions in a context of endogenous demand (i.e., demand sensitive to price and quoted lead time).The literature dealing with pricing and lead time quotation under an endogenous demand mainly considered a make to order (MTO) context. A pioneer paper, Palaka et al. (1998), investigated this issue by modeling the company as an M/M/1 queue, and our work follows their footsteps. Our review of the literature allowed to identify new perspectives for this problem, which led to three main contributions in this thesis.In our first contribution, using Palaka et al.’s framework, we consider the unit production cost to be a decreasing function in quoted lead time. In most published papers, the unit production cost was assumed to be constant. In practice, the unit production cost generally depends on the quoted lead time. Indeed, the firm can manage better the production process and reduce the production cost by quoting longer lead time to the customers.In the second contribution, we still consider Palaka et al.’s framework but model the firm as an M/M/1/K queue, for which demand is rejected if there are already K customers in the system. In the literature on single firm setting following Palaka et al.’s research, only the M/M/1 queue was used, i.e., where all customers are accepted, which might lead to long sojourn times in the system. Our idea is based on the fact that rejecting some customers, might help to quote shorter lead time for the accepted ones, which might finally lead to a higher profitability, even if in the first glance we lose some demand.In the third contribution, we study a new framework for the lead time quotation and pricing problem under endogenous demand as we model the supply chain by two production stages in a tandem queue (M/M/1-M/M/1). In the literature with multi-firm setting, all papers considered that only one actor has production operations and the other actor has zero lead time. We investigated both the centralized and decentralized decision settings.For each problem studied, we formulated a profit-maximization model, where the profit consists of a revenue minus the production, storage and lateness penalty costs, and provides the optimum result (analytically or numerically). These resolutions led us to demonstrate new theoretical results (such as the expected lateness in an M/M/1/K, and the sufficient condition required to satisfy the global service constraint in a tandem queue by only satisfying the local service constraints). We also conducted numerical experiments and derived managerial insights.

Thesis (Ph. D.)--Massachusetts Institute of Technology, Dept. of Electrical Engieering and Computer Science, 1982. / MICROFICHE COPY AVAILABLE IN ARCHIVES AND ENGINEERING. Engineering copy is in leaves. / Bibliography: p. 156-157. / by Daniel Uri Friedman. / Ph.D.

Thesis (M.S.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 1981. / MICROFICHE COPY AVAILABLE IN ARCHIVES AND ENGINEERING. / Bibliography: leaves 120-124. / by John Nikolaos Tsitsiklis. / M.S.

Made available in DSpace on 2015-04-22T22:10:16Z (GMT). No. of bitstreams: 1 alexandre.pdf: 1847704 bytes, checksum: 0a0ce0d01ec14506d096a04795e0d60e (MD5) Previous issue date: 2009-09-30 / This academic work intends to use simulation modeling to study the behavior of the queue of attendance of voters who request registration, which can range from 50 voters per day, in periods of small demand, to 8000 voters per day, in electoral years. For such, a study of case in the Central Office of Attendance of Voters (CATE) of the Electoral Regional Court of Amazon (TRE-AM) was made, emphasizing the process of arrival and attendance of voters, regarding the operations of enlistment, revision, transference and second copies. To develop this study, a bibliographical survey was made on Services of Electoral Attendance, Queuing Theory, Modeling and Simulation of Systems, and Arena® Simulator. In addition to the collecting of data of the field research, queries to the historical information stored in the database of TREAM were made, with the purpose of validation of the collected data. From these data, a mathematical model was developed, whose model of computational simulation was implemented with the software Arena®, and validated for the current operational conditions of small demand. From this model, nine alternative scenes for the operational conditions of the central office of attendance of voters were generated, allowing comparisons considering the amount of attendants, their level of experience, operations of attendance, and demand of voters. The analysis of the results shows that a modification in the operational configuration considerably influences the average time that the voter waits in line and the size of the line, being the main influence factors the number of attendants and their level of experience, which suggests that some measures, such as training, standardization of the procedures and implantation of new provisory attendance centers a longer time in advance, will guarantee quality improvement of the services / Este trabalho acadêmico busca utilizar a modelagem de imulação para estudar o comportamento da fila de atendimento aos eleitores que solicitam títulos eleitorais, a qual pode variar de 50 eleitores por dia, em período de demanda pequena, até chegar a 8.000 eleitores por dia, em períodos de anos eleitorais. Para tal, foi feito um estudo de caso na Central de Atendimento ao Eleitor (CATE) do Tribunal Regional Eleitoral do Amazonas (TRE-AM), dando ênfase ao processo de chegada de eleitores e ao processo de atendimento, no que diz respeito às operações de alistamento, revisão, transferência e segunda via. Para desenvolvimento deste estudo, foi feito um levantamento bibliográfico sobre Serviços de Atendimento Eleitoral, Teoria das Filas, Modelagem e Simulação de Sistemas, e Simulador Arena®. Além da coleta de dados da pesquisa de campo, foram feitas consultas às informações históricas armazenadas no banco de dados do TRE-AM para fins de validação dos dados coletados. A partir desses dados, foi desenvolvido um modelo matemático, cujo modelo de simulação computacional foi implementado no software Arena®, e validado para as condições operacionais atuais de demanda pequena. Com esse modelo, foram gerados nove cenários alternativos para as condições operacionais da central de atendimento ao eleitor, fazendo comparações de acordo com a quantidade de atendentes, nível de experiência dos atendentes, operações de atendimento, e demanda de eleitores. A análise dos resultados mostra que a alteração na configuração operacional influencia consideravelmente no tempo médio que o eleitor espera na fila, assim como no tamanho da fila, sendo os fatores principais a quantidade e o nível de experiência dos atendentes, sugerindo-se assim treinamentos, padronização dos procedimentos e implantação de novas Centrais de Atendimento provisórias, com um tempo de antecedência maior que o atual, garantindo-se assim melhor qualidade nos serviços de atendimento.

We consider a class of queueing systems that consist of server pools in parallel and multiple customer classes. Customer service times are assumed to be exponentially distributed. We study the asymptotic behavior of these queueing systems in a heavy traffic regime that is known as the Halfin and Whitt many-server asymptotic regime. Our main contribution is a general framework for establishing state space collapse results in the Halfin and Whitt many-server asymptotic regime for parallel server systems having multiple customer classes. In our work, state space collapse refers to a decrease in the dimension of the processes tracking the number of customers in each class waiting for service and the number of customers in each class being served by various server pools. We define and introduce a state space collapse function, which governs the exact details of the state space collapse. Our methodology is similar in spirit to that in Bramson (1998); however, Bramson studies an asymptotic regime in which the number of servers is fixed and Bramson does not require a state space collapse function. We illustrate the applications of our results in three different parallel server systems. The first system is a distributed parallel server system under the minimum-expected-delay faster-server-first (MED-FSF) or minimumexpected- delay load-balancing (MED-LB) policies. We prove that the MED-FSF policy minimizes the stationary distribution of total number of customers in the system. However, under the MED-FSF policy all the servers in our distributed system except those with the lowest service rate experience 100% utilization but under the MED-LB policy, on the other hand, the utilizations of all the server pools are equal. The second system we consider is known as the N-model. We show that when the service times only depend on the server pool providing service a static priority rule is asymptotically optimal. Finally, we study two results conjectured in the literature for V-systems. We show for all of these systems that the conditions on the hydrodynamic limits can easily be checked using the standard tools that have been developed in the literature to analyze fluid models.