Lisandro Z Granville

In view of the challenges encountered in management the flows that transit over a traffic exchange network, the identification of called elephants flows can contribute to the quality of services provided to its participants. In this perspective, taking advantage of the resources found in switches with pro-grammable support, this work presents a mechanism to realize the identification of elephant flows directly in the programmable data plane of the network. Each packet is analyzed by a switch when it ingresses the network and the identification process occurs in immediate. The prototype developed in P4 showed up significantly more efficient than the state-of-the-art approaches implemented with the OpenFlow protocol. The results showed that it is possible to identify elephant flows quickly and efficiently, with less than 35% false positives and less than 10% false negatives, in a scenario where memory resources were scarce.

With the purpose of lessening the software production efforts in computer networks area, the use of tools that help the system specification and implementation process is needed. This work presents a development tool based on the NSOMA methodology, intended to help the designers in the formal definition of specifications. It is composed of a graphical editor, an animation module that allows the visualization of specification dynamic behavior and a source code automatic generating module.

Design of Distributed Multimedia Applications (DAMD) is a multi-institutional cooperative project aiming the development of a methodology, based on the Formal Description Technique (FDT) Enhancements to Language of Temporal Ordering Specification (E-LOTOS) and supported by a set o appropriate tools, for the specification, validation, implementation, and testing of distributed multimedia applications. This paper presents the main results of this project.

This work presents the E-DART (Enhancements to DART), a graphical environment for the specification of systems based on graphic constructions for the formal description technique E-LOTOS. The use of diagrams reduces the complexity of the specifications turning E-LOTOS accessible even to users not familiarized with a formal description technique. A new graphic syntax was created as well as a specification tool, the E-DART Editor, that supports the new syntax. A translating module, from graphic specifications to the E-LOTOS syntax, is also part of the environment. The main purpose is to provide a new way to allow the user to specify time-dependent systems in a graphical approach, abstracting the largest complexities of the E-LOTOS textual syntax.

This work presents the E-DART (Enhancements to DART), a graphical environment for the specification of systems based on graphic constructions for the formal description technique E-LOTOS. The use of diagrams reduces the complexity of the specifications turning E-LOTOS accessible even to users not familiarized with a formal description technique. A new graphic syntax was created as well as a specification tool, the E-DART Editor, that supports the new syntax. A translating module, from graphic specifications to the E-LOTOS syntax, is also part of the environment. The main purpose is to provide a new way to allow the user to specify time-dependent systems in a graphical approach, abstracting the largest complexities of the E-LOTOS textual syntax.

RESUMO Estando na Era da Informação, a Internet é o meio de comunicação que apresenta o maior índice de crescimento e aceitação, tornando-se uma enorme fonte de marketing a ser explorada na utilização do comércio eletrônico. Diante disto, restam poucas alternativas às empresas que não aceitarem a Internet como um novo canal para comercialização de seus produtos. Mas a entrada neste novo cenário, principalmente por parte de estabelecimentos de médio e pequeno porte, não é simples pois, nem todos possuem a infra-estrutura necessária. Objetiva-se disponibilizar um acesso ao comércio eletrônico aos estabelecimentos que queiram comercializar seus produtos na Internet, mas não possuem a infra-estrutura para a implantação de um site de comércio eletrônico. ABSTRACT As we are in the Age of the Information, Internet is the means of communication that presents the highest rate of growth and acceptance, becoming a big source of marketing to be explored in the eletronic business therefore, there are left few alternatives for the enterprises that do not accept the Internet as a new canal for the commercialization of their products. However, the entrance is this new setting is not easy, mainly for the medium and small establishments that do not have a switable infrastructure yet. It's aimed at providing an access to the eletronic business for the stores that want to commercialize their products by the Internet, but dont't have the infrastructure to implant na eletronic business site.

As the Internet grew and evolved, it became more broadly used by everyone. What once was destined to military and academic putposes is now used as a marketing strategy, a means to sell products and perform electronic business. Electronic business exists in two distinct forms. One is between suppliers and consumers of goods or services. The other is between a seller (company) and a final consumer (home user). Nowadays, many companies see the Internet as a niche to be explored. For some of them, implementing an e-commerce web site is not a *financial challenge, whereas smaller businesses have it differently. This work defines the most appropriate structure for a given business taking into account its own resources combined with those found on the Web and pre.rents an e-commerce site generating tool that will solve small businesses troubles.

Offering QoS on TCP/IP networks is the object of intense research. New applications, such as telemedicine, distance learning, videoconference and others can only be implemented on environments that ensure QoS for the existing services. Usually, routers are the network resources that undergo changes in order to implement it – hosts have little or no influence. This paper presents an implementation of QoS structures offered directly from the host. It also presents a marking process that is run on the end systems and can be remotely managed via SNMP. A signaling application is also described in order to allow QoS requests to inherited applications that do not perceive QoS service on the networks.

At the current development stage of computer networks, the emergence of new applications that use the high performance available is unavoidable. In this context, any service that requires high performance also requires network QoS (Quality of Service). To help the maintenance of QoS services and QoS provisioning mechanisms, this work proposes a QoS monitor. This monitor operates on IP-based networks and its goal is to measure current QoS parameters observed on the network and compare them with the negotiated QoS parameters. Thus, network managers can be informed about degradations, and proceed with proper actions in order to provide adequate conditions to applications that require strict time warranties in order to operate properly.

Providing QoS-guaranteed services in current installed networks is an important issue, but only deploying QoS services is not enough to guarantee their success: QoS management must also be provided. Nowadays, policy-based network management (PBNM) addresses this need, but such management is not enough either. Network managers often deal with QoS tasks that cannot be performed using only PBNM. This paper describes six important QoS management-related tasks (QoS installation, operation maintenance, discovery, monitoring, analysis and visualization) and shows solutions that can help managers proceed with these tasks. Unfortunately, these solutions are independent from each other, leading to a scenario where integration is difficult. To solve this lack of integration, QAME (QoS-Aware Management Environment) is proposed. QAME provides support to allow the execution of the defined QoS tasks in an integrated fashion.

Providing QoS-guaranteed services in current installed networks is an important issue, but only the deploying QoS services is not enough to guarantee their success: QoS management must also be provided Nowadays, policy based management addresses this need, but such management is not enough either Network managers often deal with QoS tasks that cannot be performed using only policy-based management. This paper describes six important QoS management-related tasks (QoS installation, operation maintenance, discoveT, monitoring, analysis and visualization) and shows solutions that can help managers proceed with these tasks. Unfortunately, these solutions are independent from each other: leading to a scenario where integration is diflcult. To solve this lack of integration, QAME (QoS-Aware Management Environment) has been developed. QAME provides support to allow the execution of the defined QoS tasks in an integrated fashion.

Providing QoS-guaranteed services in current installed networks is an important issue, but only the deploying QoS services is not enough to guarantee their success: QoS management must also be provided. Nowadays, police-based management addresses this need, but such management is not enough either. Network managers often deal with QoS tasks that cannot be performed using only policy-based management. This paper describes six important QoS management-related tasks (QoS implementation, operation maintenance, discovery, monitoring, analysis and visualization) and shows solutions that can help managers proceed with these tasks. Unfortunately, these solutions are independent from each other, leading to a scenario where integration is difficult. To solve this lack of integration, QAME (QoS-Aware Management Environment) has been developed. QAME provides support to allow the execution of the defined QoS tasks in an integrated fashion.

Providing QoS-guaranteed services in current installed networks is an important issue, but only deploying QoS services is not enough to guarantee their success: QoS management must also be provided. Nowadays, police-based network management (PBNM) addresses this need, but such management is not enough either. Network managers deal with QoS tasks that cannot be performed using only PBNM. Other solutions, besides PBNM, have to be used to proceed with QoS management-related tasks. Unfortunately, these solutions are independent from each other, leading to a scenario where integration is difficult. This paper introduces QAME (QoS-Aware Management Environment) which main goal is the provisioning of facilities to allow a common and integrated Web-based management of QoS-enabled networks.

Because of the growing interest in videoconferencing on personal computers as a new means of communication and because a number of problems prevent it, this work presents a new, low-cost vid-eoconference model that runs on personal computers and IP networks. This model collects control data in order to create a structure that is able to adapt applications to network traffic conditions. Given the traffic conditions and some user-defined quality parameters , applications must be able to adapt their behavior during a session , thus allowing the videoconference software to optimize its use of the network. Adapt is a tool that aims at testing the proposed model.

This paper introduces the notion of PoP (Policy of Policies) used to define standard policy replacement strategies in a policy-based network. We also propose an architecture to support PoP within PDPs (Policy Decision Points originally defined by the IETF). The notion of PoP, and the proposed architecture allow the automation of the policy replacement task currently manually executed by the network administrator based on the network business plan.

This work presents a classification for the management of QoS-enabled networks. This classification shows that a solution for an integrated management is required. Thus, a model for integrated QoS management is proposed. Such model is analyzed through the implementation of the QAME (QoS-Aware Management Environment) prototype. Resumo. Este trabalho apresenta uma classificação para as pesquisas relacionadas à gerência de redes com QoS. Tal classificação permite verificar que uma solução de gerência integrada é necessária. Assim, é também apresentado um modelo para gerência integrada de QoS que é analisado através da implementação do protótipo QAME (QoS-Aware Management Environment).

QoS and multicast are facilities that several modern applications require from networks. However, the management of such facilities is complex and not integrated, when based on traditional management architectures. In this paper we present a policy-based management architecture and system for the integrated management of QoS and multicast-enabled networks. The definition of policies for this architecture is also presented showing how a network administrator can use such policies in order to manage the QoS and multicast facilities. The proposed architecture is based on the IETF approach. However, we discuss how such approach had to be adapted in order to support not only QoS management, but also multicast management, in an integrated fashion. Finally, we also present the implementation of our proposal providing some management example scenarios.

Policy-based management and QoS monitoring are both tasks related to the management of modern QoS-enabled network. Although related to each other, these tasks are currently executed in a non integrated fashion. This paper presents an architecture that integrates policy-based management and QoS monitoring through the extension of the original IETF policy-based management architecture. The main advantage of using our proposed approach is that network administrators are freed to execute other tasks, while the QoS-enabled network is still monitored. Another advantage is that we are monitoring policies and verifying if they are respected in the network even after its deployment, which is a new feature absent in the IETF solution.

QoS and multicast are facilities that several modern applications require from networks. However, the management of such facilities is complex and not integrated. In this paper we present a policy-based management architecture and system for the integrated management of QoS and multicast-enabled networks. The definition of policies for this architecture is also presented. The proposed architecture is based on the IETF approach. However, we discuss how such approach had to be adapted in order to support not only QoS management but also multicast management. Finally, we also present the implementation of our proposal providing some management example scenarios.

The development of PDPs is a hard task because currently there is not a widely accepted set of protocols to implement the communication between PDPs and other elements of the PBNM solutions. This paper presents a framework to support the development of PDPs in order to easy the time spent in coding new ones. The framework is based on the IETF approach for policies, which helps in the dissemination of the IETF view of policies.

Dynamic domains are domains that need be quickly created, used and discarded. Today, there are no facilities available to support dynamic domains in most network management systems. This paper introduces two new languages to deal with dynamic domains. The first language is used to create new domains through the selection of managed objects. The second language, on its turn, is used to visualize the dynamic domains created by the first language. Both languages are explained through examples and implementations details are presented. Resumo. Domínios dinâmicos são domínios que necessitam ser rapidamente criados, utilizados e descartados. Atualmente, na maioria dos sistemas de gerenciamento de redes não existem facilidades disponíveis para suportar o conceito de domínios dinâmicos. Este artigo apresenta duas novas linguagens relacionadas à manipulação de domínios dinâmicos. A primeira linguagem é utilizada para criar novos domínios a partir da seleção de objetos gerenciáveis. A segunda linguagem, por sua vez, é utilizada para visualizar os domínios dinâmicos criados através da primeira linguagem. Ambas as linguagens são explicadas através de exemplos. Além disso, detalhes de implementação são apresentados.

Dynamic domains are domains quickly created, used and discarded. Today, there are no facilities available to support dynamic domains in most network management systems. This paper introduces two new languages to deal with dynamic domains. The first language is used to define new domains through the selection of managed objects. The second language, on its turn, is used to visualize already created dynamic domains. Both languages are explained through examples and implementations details are presented.

Keywords: The network management area has some proposals to use XML to encode information models and managed object instances. In this paper we present a solution to dynamically create SNMP to XML proxies using a SAX parser and the translation facilities from the libsmi tools. We also present an analysis system that uses the management information provided by the proxies in XML.

Abrmt-Poli~).-bared management m d QOS monitorlag are both taru misted to the m n n y e " t of modem Qos4.abled network Al. thoogh related to urh other, lbpw takr are " o U y exemted in a non hlepted fashion Tbls paper p-k an prchitstM ulat integrates pliq-bared manammt and QoS monitodng throogb the atension ol the originni IETP policy-baced managancot srrhitstum. The d n ad-wtagr of using our proposed appm~m k that network ;ulmllnlrtraton are f m d to execute other mk& nbile the Qosvlabled network k stili monitord. Another advantage k Ihat we art monitorbg poUdcs aod *U-ilyiagillheyarrmpwledin I l l e ~ ~ t a r ~ r * r v e ~ n I t. r i k d ~ l ~ ~ ~ ~ ~ w h i ~ kamalcsforeatarntinLcfETPsoluUo~. Kyaords: Nctarork and System Monitodng. D(strIbuted Monitoring, End-to-cnd QOs h IP nelwork.

Computing grids require the underlying network infrastructure to be properly configured in order to have appropriate communications among the grids' nodes. The management of networks and the management of grids are currently executed by different tools operated by different administrative personnel. Eventually, the grid communication requirements will need corresponding support from the network management tools, but such requirements are fulfilled only when grid administrators manually asks network administrators for corresponding configurations. In this paper we propose a policy translation mechanism that creates network policies given grid requirements expressed in grid policies. We also present a system prototype that allows (a) grid administrators to define grid policies, and (b) network administrators to define translating rules. These rules are used by the proposed translation mechanism to generate the necessary underlying network configuration policies.

The management of the underlying network infrastructure , which supports the grid communications, is not proceeded through the same management systems used for the grid management. In this scenario, an integrated management of grids and networks could turn the maintenance processes easier. This paper proposes an hierarchical policy-based architecture, whose goal is to allow such desired integration, where grid policies are translated to network policies following mapping rules defined by network administrators. It also describes a prototype implemented based on the architecture.

Resumo. No gerenciamen.to de redes de compUTadores, alarmes são f unda-menrais para a percepção de erros ou fa lhas. Com a UTilização de SNMP, agenres são config urados pa ra enviarem traps aos gerem es, alertando sobre a constatação de alguma anormalidade em dispositivos ou na própria rede. Porém, vários agemes em uma mesma rede podem constatar a mesma fa lha e enviarem o mesmo alerta ao gerente. Ou ainda, se a mesma fal ha não é re-solvida em detenninado tempo, inúmeros alertas são gerados. Quando há ape-nas um gerente ou apenas um nível de gerentes, não há como eliminar alertas repetidos ou considerados inúteis para o administrado!; mas quando há mais de um nível de gerentes, uma correlação de alarm es pode serfeita. Esta correlação Jaz com que o número de alarmes seja reduzido, o que aj uda o administrador na visualização das f alhas em seu soft ware de gerenciamento. Este trabalho propõe uma arquitetura onde é feita uma correlação dos alarmes e a comunicação entre geren tes é f eita através de Web Services. 1. Introdução Alarmes são muito úteis no gerenciamento de redes de computadores, pois a iniciati va é do dispos itivo, av isando ao gerente que há algum problema. Dessa forma o gerente não precisa consultar todos os dispos itivos da rede em um intervalo de tempo muito pequeno. Além di sso, o alarme é di sparado seguindo limiares definidos pelo administrador, como por exempl o, em determinado momento o admini strador defin e que um determinado per-centual de colisões de pacotes na rede é mu ito alto e deve ser gerado um alarme, e, em outro momento o admini strador dec ide baixar esse percentual. Os alarmes mais utili zados são os traps, transportados pelo SNMP. Cada alarme é transportado em uma mensagem, não havendo correlação entre alarmes. Es ta possível correlação poderi a reduzir o número de mensagens. Se um problema persiste por muito tempo, os alarmes são gerados continuamente, relatando o mesmo probl ema, o que pode "inundar"a rede com tráfego desnecessário. Outro problema é que determinadas fal-has ocasionam outras, se não forem corri gidos os problemas. Como não há correlação levando-se em consideração determinado interva lo de tempo, toda Falhas geram alarmes, o que " inunda" mais ainda a rede. A partir do SNMPv2 , a comunicação entre gerentes se tornou possível, porém os alarrnes são repassados sem nenhuma correlação, e o tráfego continua sem ser reduzido.

Advanced applications, such as video conference, video on-demand, distance learning, among others, make use of computer networks as infrastructure support. However, in order to achieve an effective usage, the computer networks, on their turn, must offer some special capabilities to handle the applications needs. Quality of service (QoS) and multicast transmission support are some of the capabilities that must be supported. Besides the QoS and multicast support, it is necessary to supply a network management suitable to the applications expectations. Proposals that provide an individual solution for each of the capabilities have already been implemented. However, these solutions are not integrated, which makes the task of network management difficult and complex, due to the limitations inherent to non integrated systems. In this paper is showed a policy-based solution for integrated QoS and Multicast management using PCIMe. Resumo. Aplicações como videoconferência, vídeo sob-demanda, aplicações de ensino a distância, entre outras, utilizam-se das redes de computadores como infra-estrutura de apoio. Mas para que tal uso seja efetivo, as redes de computadores, por sua vez, devem fornecer algumas facilidades especiais para atenderàs necessidades dessas aplicações. Dentre as facilidades que devem ser fornecidas estão os suportesà qualidade de serviço (QoS-Quality of Service) e as transmissões multicast. Além do suporte a QoS e multicast nas redes, ´ e necessário fornecer um gerenciamento da rede adequadò as expectativas de tais aplicações. Soluções que fornecem gerenciamento de forma individual para tais facilidades, já foram propostas e implementadas. Entretanto, estas soluções não conseguem agir de modo integrado, o que torna a tarefa do gerente da rede extremamente complexa e difícil de ser executada, pois possibilitam um fornecimento não adequado das facilidades desejadasàs aplicações. Neste artigó e apresentada uma solução baseada em políticas para gerenciamento integrado de QoS e multicast, usando PCIMe.

High performance clusters, like any network element, require management. Cluster management area has no standard protocols or de facto management tools like network management. For this reason the interaction with other tools is so hard. To obtain interoperability, cluster management particu-larities should be adapted to the management architecture used in the network. This work presents the experiences obtained with SNMP-based cluster management. Moreover, a cluster management tool based on SNMP is proposed.

Computing grids require the underlying network infrastructure to be properly configured in order to have appropriate communications among the grids' nodes. The management of networks and the management of grids are currently executed by different tools operated by different administrative personnel. Eventually, the grid communication requirements will need corresponding support from the network management tools, but such requirements are fulfilled only when grid administrators manually asks network administrators for corresponding configurations. In this paper we propose a policy translation mechanism that creates network policies given grid requirements expressed in grid policies. We also present a, system prototype that allows (a) grid administrators to define grid policies, and (b) network administrators to define translating rules. These rules are used by the proposed translation mechanism to generate the necessary underlying network configuration policies.

It is a great challenge to analyze all the data obtained from the variety of measurement tools existing today. This task is even more complicated when we need to extract useful information for managing networks belonging to different domains due to systems and equipment heterogeneity. The utilization of database systems coupled with Web Services, can bring significant contribution to managing application that make use of those measurement. This article proposes a new approach for traffic flow measurement managing which provides heterogeneous systems integration using standard based technologies. Resumo. Analisar os dados advindos das medições realizadas através da vari-ada quantidade de ferramentas existentes hojé e um grande desafio. A tarefa torna-se ainda mais complicada quando se pretende extrair informaçõesúteis ao gerenciamento das redes que pertencem a domínios diferentes. Neste sen-tido, a utilização de sistemas de banco de dados em conjunto com Serviços Web, pode trazer contribuições significativasàs aplicações de gerenciamento que fazem uso dos dados destas medições. O presente artigo propõe uma nova abordagem para o gerenciamento das medições por fluxos de tráfego, prevendo a integração entre sistemas heterogêneos utilizando tecnologias baseadas em padrões.

NetFlow is a solution for network metering based on the capture of flow descriptors issued by network devices, such as routers and switches. In larger networks, such as backbones, it is not rare to be forced to use several collectors in order to retrieve the global flow distribution. This paper presents a Web Services-based solution for the management of NetFlow collectors distributed along different administrative domains. The objective is, not only collect and co-relate the data captured by collectors, but also to control, through configuration operations, the process of data capturing and storage executed within collectors. Resumo. NetFlow é uma solução para medição de rede baseada na captura de descritores de fluxo transmitidos por dispositivos, como roteadores e switches. Em redes de grandes dimensões, como num backbone, não é raro ter de se utilizar vários coletores NetFlow. Este artigo apresenta uma solução baseada em Web Services para o gerenciamento de coletores NetFlow espalhados ao longo de vários domínios administrativos diferentes. O objetivo é, não apenas coletar e correlacionar os dados capturados pelos coletores, mas também controlar, através de operações de configuração, o modo como os coletores de uma rede executam a captura e armazenam os dados recuperados. 1. Introdução No gerenciamento de redes, a medição é uma das tarefas mais importantes porque permite aos administradores compreender o uso e o comportamento da infra-estrutura de comunicação. Existem diversas soluções para medição de redes, sendo que os sistemas que permitem a análise de redes locais internas aos domínios administrativos são sistemas mais maduros e estáveis. Por outro lado, os sistemas para medição de enlaces de longa distância, por exemplo, enlaces que interligam POPs de um grande backbone, são soluções mais recentes e que ainda podem evoluir consideravelmente. Um modelo comum e amplamente aceito para medições utiliza coletores de tráfegos em pontos críticos da rede. Os coletores recebem as informações relativas aos fluxos de equipamentos (tipicamente roteadores) e processam essas informações que são então analisadas pelos administradores. Ferramentas gráficas ajudam muito nesse processo, e é comum se encontrar nas soluções a exibição das informações processadas através de gráficos históricos em páginas Web.

This paper describes an architecture able to support automatic policy replacement when defined network events occur. Thus, an automated system capable to react to network behavior changes is achieved, deploying a new policy better suited to this new network behavior. The architecture main objective is to help network managers with network operations maintenance task when the occurrence of special events leads to network malfunction. A formal architecture able to provide this functionality and new related concepts are presented, as well as a functional prototype of this architecture. Resumo. Este artigo descreve uma arquitetura capaz de suportar a substituição automática de políticas quando determinados eventos ocorrem na rede. Desta forma, obtém-se um sistema autônomo capaz de reagir a mudanças no comportamento da rede, aplicando uma nova política mais adequada a esta nova situação em que a rede se encontra. O objetivo principal desta arquitetura é auxiliar o gerente na tarefa de manutenção da operação de sua rede quando da ocorrência de eventos que possam perturbar o seu funcionamento. A arquitetura genérica capaz de prover esta funcionalidade e os novos conceitos envolvidos são apresentados, bem como a implementação desta arquitetura em um sistema de gerência.

High perfonnance clusters, like any network element, require management. Cluster management tools do not follow a standard operation method, making the interaction with other tools hard. To obtain interoperability, cluster management particularities should be adapted to the management architecture used in the network. This work presents the experiences obtained with SNMP-based cluster management. Moreover, U cluster management tool based on SNMP is proposed.

The management of computing grids is required in order to allow the proper operation of the grid services offered to users. However, the management of the underlying network infrastructure, which supports the grid communications , is proceeded through different management systems than those used for the grid management. In this scenario, an integrated management of grids and networks could turn the maintenance processes easier. This paper proposes an hierarchical policy-based architecture whose goal is to allow such desired integration. In the architecture proposed grid policies are translated to network policies following mapping rules defined by network administrators. The paper also describes a prototype implemented based on the architecture.

The traditional management by delegation model has lead the IETF, more recently, to define its Script MIB, which allows management entities to transfer, control and retrieve results generated by management scripts. These operations, however, can also be accomplished by Web Services. Web Services has been pointed out as an interesting approach for network management in general, but in this paper we present the use of Web Services in the specific context of management by delegation. We also present a Web Services-based system prototype that allows us to compare a Script MIB implementation against SNMP to Web Service gateways for management by delegation.

Currently, Web Services for network management have been an intensive field of investigation. However, the investigations carried up to do not properly address the issue related to event notifications. The Simple Network Management Protocol, which is the widely deployed and accepted solution, has an interesting but also limited support for event notification through its trap messages. In this paper we present a distributed notification correlation architecture that allowed us to closely investigate the use of Web Services as a tool in managing networks considering the specific case of notification support. Other aspect explored in this paper is the use of policies in a notification correlation architecture. This study complements the investigations under development showing aspects of Web Services and Policy Based Network Management (PBNM) for network management that were unknown in the field of event notification.

The management of high-performance clusters, in several situations, needs to be integrated with the management of the remaining of the computer network. In order to achieve such an integration, a network management architecture and protocol, such as the SNMP, can also be used for the management of clusters. This paper investigates the required management integration, through the implementation of SNMP agents deployed in the infrastructure of a cluster and, as well, also investigates the development of a management system for use over SNMP agents, themselves. Resumo. O gerenciamento de clusters de alto desempenho necessita, em diversas situações, ser integrado ao gerenciamento do restante da rede de computadores. Para se obter essa integração pode-se adotar uma arquitetura e um protocolo padronizado de gerenciamento de redes, tal como o SNMP, para o gerenciamento de clusters. Este artigo investiga essa integração de gerenciamento, através da implementação de agentes SNMP, para emprego na infra-estrutura de clusters, bem como, também o desenvolvimento de um sistema para uso no gerenciamento destes mesmos agentes SNMP.

—The management of long distance backbones based on high speed optical networks requires new solutions for challenging tasks. For instance, operators and users located at different administrative domains must communicate with each other in order to configure and monitor agreed quality of service levels. This paper proposes a novel peer-to-peer (P2P) management architecture for optical networks, focused initially on the new RNP Giga backbone. In the proposed architecture, peers provide, in a ubiquitous fashion, management information to modules that interface with both the optical infrastructure and network users. This paper also presents preliminary experiments carried out in order to evaluate the processing overhead imposed by the P2P infrastructure on the SNMP-based management platform.

Device configuration is an important management task that requires proper protocols for its appropriate accomplishment. Considering the currently available configuration protocols, NETCONF distinguishes itself from the others. On the other hand, SOAP can also be used in configuration and today plays an important role on Web Services popularization, technology that is responsible for Web applications interoperability. This paper discusses the usage of NETCONF and SOAP for device configuration. In addition, we present system prototypes that implement such protocols and allow their performance evaluation in terms of response time an bandwidth consumption. Resumo. A configuração de dispositivos é uma importante tarefa de gerenciamento que requer protocolos que atendam suas necessidades básicas. Dentre os protocolos de configuração disponíveis na atualidade, destaca-se o NETCONF. Por outro lado, SOAP também pode ser usado para configuração e vem ganhando importância com a atual popularização dos Web Services, os quais proporcionam interoperabilidade entre aplicações Web. Neste artigo é discutido o uso de NETCONF e SOAP para a configuração de dispositivos. Além disso, são apresentados protótipos que implementam tais protocolos e que permitiram a realização de avaliações de desempenho dos mesmos em relação ao tempo de resposta e banda consumida.

AbstrAct Large network security companies often report websites, called Booters, that offer DDoS attacks as a paid service as the primary reason for the increase in occurrence and power of attacks. Although hundreds of active Booters exist today, only a handful of those that promoted massive attacks faced mitigation and prosecution actions. In this tutorial article we focus our attention on Booters that are " under the radar " of security initiatives, by advertising high attack power and being very popular on the Internet. We discuss and provide grounds for critical thinking on what should be further done toward Booter mitigation.

AbstrAct National research and education networks (NRENs), such as ESnet, GéANT, and RNP, currently promote the employment of dynamic circuit networks (DCNs) to improve scientific communications beyond the capabilities of today's Internet. In spite of their alleged benefits to materialize user-initiated, ad hoc dedicated end-to-end circuits for high-demanding applications , DCNs also pose important challenges. Two examples are dealing with the end-user's lack of ability or willingness to understand low-level technicalities of virtual circuit establishment, and accommodating NRENs' local policies throughout the life-cycle of DCN. In this article, we seek to improve the usability of DCN services with a focus on the inexperienced end-user and the skilled network operator alike. We introduce MEICAN, a platform to manage the life-cycle of DCN from definition to provisioning of virtual circuits. We also present a case study of inter-domain circuit reservation with mixed manual and automated policy checking, and discuss lessons learned, relevant challenges, and open issues still to be overcome.

—Many aspects of the management of computer networks, such as quality of service and security, must be taken into consideration to ensure that the network meets the users and clients demands. Fortunately, management solutions were developed to address these aspects, such as Intent-Based Networking (IBN). IBN is a novel networking paradigm that abstracts network configurations by allowing administrators to specify how the network should behave and not what it should do. In this paper, we introduce an IBN solution called INSpIRE (Integrated NFV-based Intent Refinement Environment). INSpIRE implements a refinement technique to translate intents into a set of configurations to perform a desired service chain in both homogeneous environments (VNFs only) and heterogeneous environments (VNFs and physical middleboxes). Our solution is capable of (i) determining the specific VNFs required to fulfill an intent, (ii) chaining these VNFs according to their dependencies, and (iii) presenting enough low-level information to network devices for posterior traffic steering. Finally, to assess the feasibility of our solution we detail a case study that reflects real-world management situations and evaluate the scalability of the refinement process.

—Internet Exchange Points (IXPs) play a key role in the Internet architecture, enabling cost-effective connections among multiple autonomous systems (ASes). The management of IXP networks includes the activity of taking care of elephant flows; they represent a small number of the total flows of an IXP, but have high impact on the overall network traffic. Managing elephant flows involves adequate identification and eventually rerouting of such flows to more appropriate locations, to minimize the possible negative impact on the other (mice) flows. Elephant flow management becomes even more important in SDN-based IXPs that require controllers to have a consistent view of the underlying network to allow fine-grained adjustment. In this paper, we propose, develop, and evaluate a recommendation system to suggest alternative configurations to previously identified elephant flows in an SDN-based IXP network. In our solution, the IXP operator can define templates that ultimately define how elephant flows can be rerouted to achieve a specific objective. We demonstrate that our system can help IXP operators to mitigate the impact of elephant flows on the IXP network.

Industry and academia have increased the deployment of Network Functions Virtualization (NFV) on their environments, either for reducing expenditures or taking advantage of NFV flexibility for service provisioning. In NFV, Service Function Chainings (SFC) composed of Virtualized Network Functions (VNF) are defined to deliver services to different customers. Despite the advancements in SFC composition for service provisioning, there is still a lack of proposals for ensuring the integrity of NFV service delivery, i.e., detecting anomalies in SFC operation. Such anomalies could indicate a series of different threats, such as DDoS attacks, information leakage, and unauthorized access. In this PhD, we propose a framework composed of an SFC Integrity Module (SIM) for the standard NFV architecture, providing the integration of anomaly detection mechanisms to NFV orchestrators. We present recent results of this PhD regarding the implementation of an entropy-based anomaly detection mechanism using the SIM framework. The results presented in this paper are based on the execution of the proposed mechanism using a realistic SFC data set.

—Programmable Virtual Networks (PVNs) make the network more flexible and allow the fast introduction of new services. However, several shortcomings hamper their wider adoption, including: (i) the extensive knowledge required to configure and manage the NetApps; (ii) the lack of descriptors to detail all nuances of the NetApps; and (iii) there is no solution that enables to distribute and configure NetApps over distinct technologies. Therefore, we propose iMPROVE to simplify the introduction of services in PVNs, enhancing the distribution of NetApps. We also extend the ETSI network service descriptor to support distinct technologies as well as to represent conflict issues. We demonstrate evidence of iMPROVE's feasibility in a case study and compare it with the main solutions for distributing and deploying applications over multiple platforms.

— Information Visualization (InfoVis) is a powerful tool to assist network administrators in daily tasks, and several authors report investigations about visualization techniques for network management. However, we have observed a fundamental issue that has not been addressed yet: how to promote the adoption of visualizations by network administrators, focusing on improving productivity and reducing costs? We claim that, in general, administrators do not have expertise in the InfoVis domain. Thus, the adoption of visualizations tends to be expensive and decrease the productivity of administrators, in special, because they lose focus on their core tasks. To overcome the raised issue, in this paper, we introduce a reuse-based approach (named as VisNSM) that aims to promote and encourage the adoption of visualizations for network management. To analyze and evaluate the feasibility of VisNSM, we have conducted a case study, and the obtained results show that our approach can significantly reduce costs and improve productivity.

—Network Functions Virtualization (NFV) is driving a paradigm shift in telecommunications networks, fostering new business models and creating innovation opportunities. In NFV-enabled networks, Service Providers (SPs) have the opportunity to build a business model where tenants can purchase Virtual Network Functions (VNFs) that provide distinct network services and functions. However, the chance to negotiate VNFs requires a change in traditional network planning strategies to accommodate tenants demands. In this context, the planning tasks perform a critical role in the introducing of business strategies that encompass both profit and health of services, which requires operators to have a broad understanding of the environment. In this paper, we propose the usage of two interactive visualization techniques to help NFV network operators in planning and strategic decisions. We advocate that our visualizations can aid in NFV planning tasks, such as infrastructure investment, resources allocation, and service pricing. We present three case studies to provide evidence of the feasibility and effectiveness of our visualizations.