O aumento crescente no número de processadores das arquiteturas paralelas que estão no topo dos rankings de desempenho, apesar de permitir uma maior capacidade de processamento, também traz consigo um aumento na taxa de falhas diretamente proporcional ao número de processadores. Atualmente, as técnicas de tolerância a falhas com recuperação retroativa são as mais empregadas em aplicações MPI, principalmente a técnica de checkpoint coordenado. No entanto, previsões afirmam que essa última técnica será inadequada para as arquiteturas emergentes. Em contrapartida, as técnicas de log de mensagens possuem características que as tornam mais apropriadas no novo cenário que se estabelece. O presente trabalho consiste em uma proposta de log de mensagens pessimista remoto com checkpoint não-coordenado e a avaliação de desempenho da comunicação MPI sobre Publish/Subscriber no qual se baseia o log de mensagens. O trabalho compreende: um estudo das técnicas de tolerância a falhas mais empregadas em ambientes de alto desempenho e a motivação para a escolha dessa variante de log de mensagens; a proposta de log de mensagens; uma implementação de comunicação Open MPI sobre OpenAMQ e sua respectiva avaliação de desempenho com comunicação tradicional TCP/IP e com o log de mensagens pessimista local da distribuição do Open MPI. Os benchmarks utilizados foram o NetPIPE, o NAS Parallel Benchmarks e a aplicação Virginia Hydrodynamics (VH-1). / The growing number of processors in parallel architectures at the top of performance rankings allows a higher processing capacity. However, it also brings an increase in the fault rate which is directly proportional to the number of processors. Nowadays, coordinated checkpoint is the most widely used rollback technique for system recovery in the occurrence of faults in MPI applications. Nevertheless, projections point that this technique will be inappropriate for the emerging architectures. On the other hand, message logging seems to be more appropriate to this new scenario. This work consists in a proposal of pessimistic message logging (remote based) with non-coordinated checkpoint and the performance evaluation of an MPI communication mechanism that works over Publish/Subscriber channels in which the proposed message logging is based. The work is organized as following: an study of fault tolerant techniques used in HPC and the motivation for choosing this variant of message logging; a message logging proposal; an implementation of Open MPI communication over OpenAMQ; performance evaluation and comparision with the tradicional TCP/IP communication and a pessimistic message logging (sender based) from Open MPI distribution. The benchmark set is composed of NetPIPE, NAS Parallel Benchmarks and Virginia Hydrodynamics (VH-1).

Submitted by Luiz Felipe Barbosa (luiz.fbabreu2@ufpe.br) on 2015-03-09T13:59:48Z No. of bitstreams: 2 DISSERTAÇÃO Gilson Medeiros de Oliveira Júnior.pdf: 2514749 bytes, checksum: 5cb30bafc7935aa9accc6f4541777d22 (MD5) license_rdf: 1232 bytes, checksum: 66e71c371cc565284e70f40736c94386 (MD5) / Made available in DSpace on 2015-03-09T13:59:48Z (GMT). No. of bitstreams: 2 DISSERTAÇÃO Gilson Medeiros de Oliveira Júnior.pdf: 2514749 bytes, checksum: 5cb30bafc7935aa9accc6f4541777d22 (MD5) license_rdf: 1232 bytes, checksum: 66e71c371cc565284e70f40736c94386 (MD5) Previous issue date: 2014-02-26 / A mobilidade urbana é uma questão preocupante que vem ganhando bastante atenção nos últimos anos, principalmente, devido a alguns fatores como, a proximidade da realização de eventos internacionais como a Copa do Mundo de Futebol de 2014 e as Olimpíadas de 2016, além do aumento no número de carros, a má qualidade do transporte público urbano, a falta de infra-estrutura viária e o aumento da demanda do transporte público, graças ao aumento populacional. Aliado a resolução ou a amenização desses problemas está o crescimento tecnológico, principalmente no que diz respeito a tecnologia móvel, através da popularização e do uso de dispositivos como: smartphones e tablets, por exemplo. Esse crescimento tecnológico torna a computação cada vez mais presente na vida e no cotidiano das pessoas e, com isso, surge um novo paradigma da computação chamado de Computação Ubíqua. Nesse cenário de utilização de tecnologia móvel e da Computação Ubíqua, destacam-se as aplicações sensíveis ao contexto, que podem ser acessadas em qualquer lugar e a qualquer momento, levando em consideração informações estáticas e dinâmicas dos seus usuários. Essas aplicações, na maioria das vezes, precisam de uma infra estrutura distribuída especializadas provida através de middleware para o gerenciamento (processamento, aquisição e armazenamento) de informações contextuais. Este trabalho propõe uma arquitetura de middleware que busca prover suporte ao desenvolvimento de aplicações ubíquas e sensíveis ao contexto. A arquitetura proposta foi definida após o estudo do estado da arte e dos requisitos necessários para o desenvolvimento desse novo tipo de aplicações. A arquitetura é baseada, principalmente, na criação e acoplamento de componentes, tornando fácil sua extensão e manutenção. A implementação da arquitetura proposta é validada através da desenvolvimento de um middleware chamado UbiMid que atua no domínio dos sistemas inteligentes de transporte público, desenvolvido no âmbito do projeto UbiBus.

Made available in DSpace on 2014-06-12T15:59:37Z (GMT). No. of bitstreams: 2 arquivo5349_1.pdf: 1338520 bytes, checksum: 967ea568492e0071544a3a6f547aed59 (MD5) license.txt: 1748 bytes, checksum: 8a4605be74aa9ea9d79846c1fba20a33 (MD5) Previous issue date: 2006 / A TV Digital Interativa traz consigo uma série de mudanças que vai muito além da melhoria de qualidade em áudio e vídeo. Ela trará o aumento da diversidade e disponibilidade de conteúdo multimídia e fará com que as pessoas deixem de ser meros espectadores para se tornarem também usuários de aplicativos multimídia interativos. Estes aplicativos virão com o conteúdo multimídia e irão permitir que o usuário possa interagir ativamente com o conteúdo exibido ou até mesmo com as emissoras, pois os padrões de TV digital prevêem a existência de um canal de interação através do qual o aparelho de TV possa se comunicar numa rede de computadores, como a Internet, por exemplo. Desta forma, as aplicações de TV digital poderão acessar serviços como e-mail, Internet-banking e comércio eletrônico. No entanto, estes são serviços centralizados e não exploram a interatividade entre as pessoas. É fácil prever que, num futuro próximo, existirá demanda por aplicações que explorem a interatividade entre as pessoas. Porém, fazer com que estas aplicações se comuniquem é um verdadeiro desafio, pois o uso de servidores para intermediar mensagens está sujeito a problemas de escalabilidade, e a troca direta de mensagens é prejudicada pela heterogeneidade da rede em que o canal de interação é disponibilizado. Este trabalho propõe um Middleware para este tipo de aplicação distribuída. O Middleware proposto foi baseado no modelo de comunicação peer-to-peer, pois, após a análise dos requisitos deste tipo de software distribuído, verificou-se que este modelo é o que melhor se adequa. O mesmo foi construído sobre a plataforma JXTA, que resolve os problemas de infra-estrutura da rede, compreendendo um serviço avançado de busca e um mecanismo de invocação remota de método, semelhante à tecnologia Java RMI

Orientadores: Luiz Eduardo Buzato, Edmundo Roberto Mauro Madeira / Tese (doutorado) - Universidade Estadual de Campinas, Instituto de Computação / Made available in DSpace on 2018-08-16T20:16:50Z (GMT). No. of bitstreams: 1 Cicerre_FabioRodrigodeLima_D.pdf: 1417286 bytes, checksum: 642d37f1cb522ec4a356bbb4e9b55b96 (MD5) Previous issue date: 2010 / Resumo: O conceito de grade surgiu com a necessidade crescente de se aproveitar recursos computacionais disponíveis em uma ou mais organizações para resolver problemas que exigem compartilhamento de dados e um grande poder de processamento. Uma grade computacional tem como objetivo principal permitir a execução distribuída e paralela de tarefas em recursos compartilhados. Uma grade é constituída de infra-estrutura física, composta de uma ou mais redes autônomas de computadores, e de um sistema de suporte (middleware), que provê serviços de gerenciamento de informações sobre os recursos da grade, controle de acesso e execução de tarefas sobre esse recursos e mecanismos de comunicação. Atualmente existem diversos sistemas que suportam a execução de tarefas independentes em uma grade computacional, mas poucos consideram a execução de processos de workflow, que permitem a definição de dependência explícita de dados e controle entre tarefas, o que impede um melhor aproveitamento de recursos, escalabilidade, desempenho de execução e recuperação automática de processos com manutenção de consistência. O sistema Xavantes, proposto e descrito nesse trabalho, procura suprir essas deficiências, tendo como principal objetivo suportar a execução distribuída de processos de workflow em máquinas heterogêneas, em uma ou mais organizações autônomas e dinâmicas, provendo um middleware que forneça uma melhor escalabilidade, desempenho e confiabilidade para a execução de aplicações em grades computacionais / Abstract: The grid concept has emerged from the increasing necessity of using available computational resources in one or more organizations in order to solve problems that require data sharing and large processing power. The main goal of a computational grid is to allow the distributed and parallel execution of tasks in shared resources. A grid is composed of a physical infra-structure, with one or more autonomous networks of computers, and a middleware, which provides services of information management about the grid resources, access control and tasks execution in these resources and communication mechanisms. Nowadays, there are some systems that support the execution of independent tasks in a computational grid, but only ones consider the execution of workflow processes, which allow the explicit definition of data and control dependencies among tasks, and this restricts a better use of available resources, scalability, execution performance, and automatic recovery of processes with the correct consistency maintenance. The Xavantes system, proposed and described in this work, is designed to reduce these deficiencies, having as its main goal the supports to the distributed execution of workflows in heterogeneous resources of one or more autonomous and dynamic organizations, providing a middleware that delivers a better scalability, performance and reliability to the application execution in grid computing / Doutorado / Sistemas Distribuídos e Redes de Computadores / Doutor em Ciência da Computação

Orientador: Luis Geraldo Pedroso Meloni / Dissertação (mestrado) - Universidade Estadual de Campinas, Faculdade de Engenharia Eletrica e de Computação / Made available in DSpace on 2018-08-09T08:48:37Z (GMT). No. of bitstreams: 1 Santos_DaviTrindadedos_M.pdf: 2563209 bytes, checksum: 37ca5bdcc2186881af9bae40d99ff148 (MD5) Previous issue date: 2007 / Resumo: O Brasil preparase para a migração de seu sistema de difusão de sinais de televisão da tecnologia analógica para a tecnologia digital, também chamada de TV Digital (TVD). Dentre os vários objetivos a serem alcançados com esta migração está a inclusão digital de milhões de brasileiros, visto que a televisão encontrase presente em mais de 90% dos domicílios do país. A implantação de programas de EAD via TVD apresentase como uma ferramenta poderosa que pode possibilitar o acesso à educação a grande parte da população. Este trabalho apresenta pesquisas realizadas dentro do projeto Sistema Brasileiro de TV Digital (SBTVD) visando o desenvolvimento de aplicativos de TV Digital interativa (TVDi) para utilização em cursos de EAD. São apresentados dois tipos de aplicativos de TVDi que foram concebidos, além de um ambiente para criação de conteúdo e um sistema para coleta e exibição de estatísticas acerca da utilização destes aplicativos / Abstract: Brazil has been conducing the migration of your television broadcasting system from the analogical to the digital system, also called Digital TV (DTV). Among the goals to be achieved with this migration is the digital inclusion of millions of brazilians that are reached by this media, which is present in more than 90% of the houses. The development of Distance Learning (DL) prograns through DTV may be seen as a tool that, using the powerfull DTV media, may help the brazilian population to engage in Distance Learning programs. This work presents researches conducted during the studies for the Brazilian System of Digital TV (SBTVD) aiming the development of interactive DTV (iDTV) applications that may be used in DL courses. Two types of applications are presented, besides a tool for content authoring and a system to collect and create usage reports from the teachers using these applications / Mestrado / Telecomunicações e Telemática / Mestre em Engenharia Elétrica

Abstract Mobile peer-to-peer (P2P) networks have emerged as a substrate for distributed wireless Internet applications. With P2P systems, it is possible to share resources such as data storage space, media files, network bandwidth, or computing power among the devices, which participate in the network, without large and expensive centralised server machines. However, the special characteristics of the mobile environment such as the low computational power, changing network conditions, and limited battery life pose several challenges for the fluent operation of mobile devices in P2P networks; this is also in part affected by the complexity of distributed P2P systems. Software development for mobile devices and the design of mobile networking systems are challenging due to the limited device resources and heterogeneous software platforms. Moreover, the energy consumption of a mobile device, and the network-wide routing efficiency that affects for example the resource lookup performance, depend on multiple variable parameters. P2P application development can be facilitated, however, by using middleware, which hides the complexity of networking from the application programmers. The research contributions of this thesis can be classified into three categories: (1) Novel functionalities of mobile middleware are proposed. One of these is a cross-layer connectivity management framework, which aims to select the best combination of network technology entities in a specific usage situation; the selection is made by logic contained in a replaceable state machine. Another new functionality is a system, which installs a missing application for the user, when the user receives a session request from another user. A signalling system based on the cellular USSD protocol is also presented; the system facilitates the establishment of a P2P connection with a mobile device, whose IP-based network connection is off. Moreover, the suitability of the P2PSIP protocol for implementing wireless distributed services is analysed. (2) P2P-related measurement studies are presented. In them, the message routing efficiency of a P2PP protocol implementation and the network traffic load caused by the messaging are observed, and the energy consumption incurred by the same implementation in a mobile device is measured. In addition, a server-based testbed system used in these measurements is described. (3) Experience-backed guidelines for mobile middleware development are presented. These include practical instructions for software development on a restricted mobile platform, and guidelines and observations related to cross-platform software development. / Tiivistelmä Mobiilit vertaisverkot (P2P) ovat uusi alusta hajautetuille langattomille Internet-sovelluksille. P2P-järjestelmien avulla on mahdollista jakaa resursseja kuten tiedon tallennustilaa, mediatiedostoja, verkon kaistanleveyttä tai laskentatehoa laitteiden kesken, jotka osallistuvat verkkoon, ilman suuria ja kalliita keskitettyjä palvelinkoneita. Mobiiliympäristön erityispiirteet kuten vähäinen laskentateho, vaihtelevat verkko-olosuhteet ja rajallinen akkukesto asettavat kuitenkin useita haasteita sujuvalle mobiililaitteiden vertaisverkkokäytölle; tähän vaikuttaa osaltaan myös hajautettujen P2P-järjestelmien monimutkaisuus. Mobiililaitteiden ohjelmistokehitys ja mobiiliverkkojärjestelmien suunnittelu on haastavaa rajallisten laiteresurssien ja epäyhtenäisten ohjelmistoalustojen vuoksi. Lisäksi mobiililaitteen energiankulutus ja koko P2P-verkon reititystehokkuus, joka vaikuttaa esimerkiksi resurssien hakutehokkuuteen, riippuvat useista muuttuvista parametreista. P2P-sovelluskehitystä voidaan kuitenkin helpottaa käyttämällä välikerrosohjelmistoja, jotka kätkevät verkon käytön monimutkaisuuden sovellusohjelmoijilta. Tämän väitöskirjan tutkimuksellinen uutuusarvo voidaan jakaa kolmeen osa-alueeseen: (1) Uusia toiminnallisuuksia mobiileihin välikerrosohjelmistoihin esitellään. Eräs näistä on verkkokerrosten välinen yhteydenhallintajärjestelmä, jossa pyritään valitsemaan tietyssä käyttötilanteessa paras verkkotekniikoiden yhdistelmä; valinnan tekee vaihdettavien tilakoneiden sisältämä logiikka. Toinen uusi toiminnallisuus on järjestelmä, joka asentaa puuttuvan sovelluksen käyttäjälle, kun käyttäjä saa toiselta käyttäjältä pyynnön ryhtyä yhteysjaksoon tämän kanssa. Myös soluverkkojen USSD-protokollaan perustuva signalointijärjestelmä esitellään; järjestelmän avulla helpotetaan P2P-yhteydenmuodostusta sellaisen mobiililaitteen kanssa, jonka IP-pohjainen verkkoyhteys on poissa päältä. Lisäksi P2PSIP-protokollan soveltuvuutta langattomien hajautettujen palvelujen toteuttamiseen analysoidaan. (2) Esitellään P2P-mittaustutkimuksia. Näissä havainnoidaan P2PP-protokollatoteutuksen viestien reititystehokkuutta ja viestien aiheuttamaa verkkoliikennekuormaa sekä mitataan saman toteutuksen aiheuttamaa energiankulutusta mobiililaitteessa. Lisäksi kuvaillaan mittauksissa hyödynnetty palvelinpohjainen testausjärjestelmä. (3) Esitetään kokemuksiin perustuvia neuvoja mobiilien välikerrosohjelmistojen kehitystyötä varten. Nämä sisältävät käytännöllisiä ohjeita rajoitetulla mobiilialustalla tapahtuvaan ohjelmistokehitykseen sekä neuvoja ja havaintoja liittyen yhtäaikaiseen useiden kohdealustojen ohjelmistokehitykseen.

Abstract The popularity of social networking services emphasises the importance of user communities. As the next evolutionary step, social networking will evolve towards user communities that are dynamically established according to contextual aspects such as the shared location and/or activities of people. In this vision, distributed peer-to-peer (P2P) technologies provide a respectable alternative to the centralised client-server technologies for implementing the upcoming community-centric mobile services. This thesis focuses on the development of enabling techniques and methods for a community-centric mobile service environment and on the evaluation of users’ perceptions and user acceptance of novel community-centric mobile services. The research was carried out by first examining the applicability of different P2P system architectures for implementing the management of user communities. Next, an experimental prototype of a community-centric mobile service environment was implemented. The experimental prototype was used for evaluating the performance of the community-centric mobile service environment and for conducting the user studies of novel community-centric mobile services. The evaluation was conducted in a real deployment environment with test users. The main results are the following: (1) a conceptual analysis of the advantages and disadvantages of P2P group management systems that are implemented using architecturally distinct P2P system architectures; (2) a performance evaluation of a P2P community management system, in which each community is implemented as an independent structured P2P overlay network; (3) a novel community-centric mobile service environment and its core component, a mobile middleware, which enables the use of P2P technologies and context information as part of Web-based applications running on a Web browser; (4) an evaluation of users’ perceptions and user acceptance of novel community-centric mobile services that are related to making a selection of an entertainment premise based on its music style and voting for music in a user community. / Tiivistelmä Käyttäjäyhteisöiden tärkeys on korostunut sosiaalista verkottumista edistävien palveluiden suosion kautta. Seuraavaksi sosiaalinen verkottuminen tulee laajentumaan yhteisöihin, jotka perustetaan dynaamisesti muun muassa ihmisten yhteisen sijainnin ja/tai toiminnan perusteella. Tulevaisuuden yhteisöllisten mobiilipalveluiden toteutuksessa hajautetut vertaisverkkoteknologiat tarjoavat varteenotettavan vaihtoehdon keskitetyille asiakas-palvelin -teknologioille. Tämä väitöskirjatyö keskittyy kehittämään tarvittavia tekniikoita ja menetelmiä yhteisöllisen mobiilipalveluympäristön toteuttamiseen sekä arvioimaan käyttäjien kokemuksia uusista yhteisöllisistä mobiilipalveluista. Tutkimuksessa tarkasteltiin aluksi arkkitehtuurillisesti erilaisten vertaisverkkojärjestelmien sopivuutta käyttäjäyhteisöjen hallinnan toteutettamiseksi. Seuraavaksi tutkimuksessa toteutettiin kokeellinen prototyyppi yhteisöllisestä mobiilipalveluympäristöstä. Kokeellista prototyyppiä hyödynnettiin sekä yhteisöllisen mobiilipalveluympäristön suorituskyvyn että uusien yhteisöllisten mobiilipalveluiden käyttäjäkokemuksen arvioinnissa. Arviointi suoritettiin aidossa käyttöympäristössä testikäyttäjien avulla. Väitöskirjatyön keskisimmät tulokset ovat seuraavat: (1) käsitteellinen analyysi arkkitehtuurillisesti erilaisten vertaisverkkojärjestelmien vahvuuksista ja heikkouksista ryhmänhallintajärjestelmän toteuttamisessa; (2) sellaisen yhteisönhallintajärjestelmän arviointi, jossa jokainen yhteisö on toteutettu erillisen ja itsenäisen rakenteellisen vertaisverkon avulla; (3) uusi yhteisöllinen mobiilipalveluympäristö, jonka keskeisin osa mobiili välikerrosohjelmisto mahdollistaa vertaisverkkoteknologioiden ja kontekstitiedon käyttämisen selaimessa toimivissa Web-sovelluksissa; (4) sellaisten uusien yhteisöllisten mobiilipalveluiden käyttäjäkokemuksen arviointi, jotka liittyvät vapaa-ajan viettopaikan valitaan musiikin perusteella ja soitettavan musiikin äänestämiseen käyttäjäyhteisöissä.

Enterprise Systems like customer-billing systems or financial transaction systems are required to process large volumes of data in a fixed period of time. Those systems are increasingly required to also provide near-time processing of data to support new service offerings. Common systems for data processing are either optimized for high maximum throughput or low latency. This thesis proposes the concept for an adaptive middleware, which is a new approach for designing systems for bulk data processing. The adaptive middleware is able to adapt its processing type fluently between batch processing and single-event processing. By using message aggregation, message routing and a closed feedback-loop to adjust the data granularity at runtime, the system is able to minimize the end-to-end latency for different load scenarios. The relationship of end-to-end latency and throughput of batch and message-based systems is formally analyzed and a performance evaluation of both processing types has been conducted. Additionally, the impact of message aggregation on throughput and latency is investigated. The proposed middleware concept has been implemented with a research prototype and has been evaluated. The results of the evaluation show that the concept is viable and is able to optimize the end-to-end latency of a system. The design, implementation and operation of an adaptive system for bulk data processing differs from common approaches to implement enterprise systems. A conceptual framework has been development to guide the development process of how to build an adaptive software for bulk data processing. It defines the needed roles and their skills, the necessary tasks and their relationship, artifacts that are created and required by different tasks, the tools that are needed to process the tasks and the processes, which describe the order of tasks.

Climate science or climatology is the scientific study of the earth’s climate, where climate is the term representing weather conditions averaged over a period of time. Climate models are mathematical models used to quantitatively describe, simulate and study the interactions among the components of the climate system -atmosphere, ocean, land and sea-ice. CCSM (Community Climate System Model) is a state-of-the-art climate model, and a long-running coupled multicomponent parallel application involving component models for simulating the components of the climate system. Each of the component models is a large-scale parallel application, and the parallel components exchange climate data through a specialized component called coupler. Typical multi-century climate simulations using CCSM take several weeks or months to execute on most parallel systems. In this thesis, we study the applicability of a computational grid for effective execution of long-running coupled multi-component climate applications like CCSM. Initial studies of the application characteristics led us to develop a dynamic component extension strategy for temporal inter-component load-balancing. By means of experiments on different parallel platforms with different number of processors, we showed that using our strategy can lead to about 15% reduction and savings of several days in execution times of CCSM for 1000-year simulation runs. Our initial studies also indicated that unlike typical grid applications, CCSM has limits on scalability to very large number of processors and hence cannot directly benefit from the large number of processors on a computational grid. However, its long-running nature and the limits of execution imposed on jobs on most multi-user batch queueing systems, led us to investigate the benefits of its execution on a grid of batch systems. The idea is that multiple batch queues can improve the processor availability rate with respect to the application thereby possibly improving its effective throughput. We explored this idea in detail with simulation studies involving various system and application characteristics, and execution models. By conducting large number of simulations with different workload characteristics and queuing policies of the systems, processor allocations to components of the application, distributions of the components to the batch systems and inter-cluster bandwidths, we showed that multiple batch executions lead to upto 55% average increase in throughput over single batch executions for long-running CCSM. Having convinced ourselves of possible advantages in performance, we then ventured to construct an application-level middleware framework. Our framework supports long duration execution of multi-component applications spanning multiple submissions to queues on multiple batch systems. It coordinates the distribution, execution, rescheduling, migration and restart of the application components across resources on different sites. It also addresses challenges including execution time limits for jobs, and differences in job-startup times corresponding to different components. Further, within the framework, we developed robust rescheduling policies that decide when and where to reschedule the components to the available resources based on the application execution characteristics and queue dynamics. Our grid middleware framework resulted in multi-site executions that provided larger application throughput than single-site executions, typically performed by climate scientists, and also removed the bottlenecks associated with a single system execution. We used this framework for long-running executions of CCSM to study the effect of increased black carbon aerosols and dust aerosols on the Indian monsoons. Black Carbon aerosols are essentially of anthropogenic origin and occur due to improper burning of fossil fuels, and dust is a naturally occurring aerosol. The concentrations of both these aerosols is high over the Indian region. We study the impact of these aerosols on precipitation and sea surface temperature (SST) through multi-decadal simulations conducted with our grid-enabled climate system model. Our observations indicated that increasing the concentrations of aerosols leads to an increase in precipitation in the central and eastern parts of India, and a decrease in SST over most of Indian ocean.

L'Internet des objets (ou IdO) se traduit à l'heure actuelle par l'accroissement du nombre d'objets connectés, c'est-à-dire d'appareils possédant une identité propre et des capacités de calcul et de communication de plus en plus sophistiquées : téléphones, montres, appareils ménagers, etc. Ces objets embarquent un nombre grandissant de capteurs et d'actionneurs leur permettant de mesurer l'environnement et d'agir sur celui-ci, faisant ainsi le lien entre le monde physique et le monde virtuel. Spécifiquement, l'Internet des objets pose plusieurs problèmes, notamment du fait de sa très grande échelle, de sa nature dynamique et de l'hétérogénéité des données et des systèmes qui le composent (appareils puissants/peu puissants, fixes/mobiles, batteries/alimentations continues, etc.). Ces caractéristiques nécessitent des outils et des méthodes idoines pour la réalisation d'applications capables (i) d'extraire des informations utiles depuis les nombreuses sources de données disponibles et (ii) d'interagir aussi bien avec l'environnement, au moyen des actionneurs, qu'avec les utilisateurs, au moyen d'interfaces dédiées. Dans cette optique, nous défendons la thèse suivante : en raison de la nature continue des données (mesures physiques, évènements, etc.) et leur volume, il est important de considérer (i) les flux comme modèle de données de référence de l'Internet des objets et (ii) le traitement continu comme modèle de calcul privilégié pour transformer ces flux. En outre, étant donné les préoccupations croissantes relatives à la consommation énergétique et au respect de la vie privée, il est préférable de laisser les objets agir au plus près des utilisateurs, si possible de manière autonome, au lieu de déléguer systématiquement l'ensemble des tâches à de grandes entités extérieures telles que le cloud. À cette fin, notre principale contribution porte sur la réalisation d'un système distribué de gestion de flux de données pour l'Internet des objets. Nous réexaminons notamment deux aspects clés du génie logiciel et des systèmes distribués : les architectures de services et le déploiement. Ainsi, nous apportons des solutions (i) pour l'accès aux flux de données sous la forme de services et (ii) pour le déploiement automatique des traitements continus en fonction des caractéristiques des appareils. Ces travaux sont concrétisés sous la forme d'un intergiciel, Dioptase, spécifiquement conçu pour être exécuté directement sur les objets et les transformer en fournisseurs génériques de services de calcul et de stockage.Pour valider nos travaux et montrer la faisabilité de notre approche, nous introduisons un prototype de Dioptase dont nous évaluons les performances en pratique. De plus, nous montrons que Dioptase est une solution viable, capable de s'interfacer avec les systèmes antérieurs de capteurs et d'actionneurs déjà déployés dans l'environnement. / The Internet of Things (IoT) is currently characterized by an ever-growing number of networked Things, i.e., devices which have their own identity together with advanced computation and networking capabilities: smartphones, smart watches, smart home appliances, etc. In addition, these Things are being equipped with more and more sensors and actuators that enable them to sense and act on their environment, enabling the physical world to be linked with the virtual world. Specifically, the IoT raises many challenges related to its very large scale and high dynamicity, as well as the great heterogeneity of the data and systems involved (e.g., powerful versus resource-constrained devices, mobile versus fixed devices, continuously-powered versus battery-powered devices, etc.). These challenges require new systems and techniques for developing applications that are able to (i) collect data from the numerous data sources of the IoT and (ii) interact both with the environment using the actuators, and with the users using dedicated GUIs. To this end, we defend the following thesis: given the huge volume of data continuously being produced by sensors (measurements and events), we must consider (i) data streams as the reference data model for the IoT and (ii) continuous processing as the reference computation model for processing these data streams. Moreover, knowing that privacy preservation and energy consumption are increasingly critical concerns, we claim that all the Things should be autonomous and work together in restricted areas as close as possible to the users rather than systematically shifting the computation logic into powerful servers or into the cloud. For this purpose, our main contribution can be summarized as designing and developing a distributed data stream management system for the IoT. In this context, we revisit two fundamental aspects of software engineering and distributed systems: service-oriented architecture and task deployment. We address the problems of (i) accessing data streams through services and (ii) deploying continuous processing tasks automatically, according to the characteristics of both tasks and devices. This research work lead to the development of a middleware layer called Dioptase, designed to run on the Things and abstract them as generic devices that can be dynamically assigned communication, storage and computation tasks according to their available resources. In order to validate the feasability and the relevance of our work, we implemented a prototype of Dioptase and evaluated its performance. In addition, we show that Dioptase is a realistic solution which can work in cooperation with legacy sensor and actuator networks currently deployed in the environment.