VoIP technology is one of the important applications of the network. In addition to using traditional PSTN telephone, people can choose favorable VoIP telephone because of the Internet popularization and High speed Internet. RTP (Real-time Transfer Protocol) is the foundation of VoIP, and it is suitable for applications to manage real-time transmission of multimedia data. But RTP does not provide any confidentiality, the eavesdrop of VoIP is easy to realize. In this paper, we discuss how to design and implement an embedded SIP user agent that provide an fast secure encoding of RTP voice streaming. We use limited resources to implement RTP encoding mechanism to prevent anyone from eavesdropping .

none

none

L'avènement de la VoIP (Voice over IP) a déclenché une période de profonds changements dans le marché des télécommunications. En particulier, dans le secteur de la téléphonie résidentielle, cette technologie s'est consolidée, rapidement et pour de nombreuses raisons, comme l'évolution de la téléphonie traditionnelle. Dès les tous débuts, et afin d'établir une base de compatibilité permettant l'interconnexion de plusieurs réseaux téléphoniques et la convergence entre les systèmes traditionnels analogiques et leur évolution numérique, l'industrie a demandé l'établissement de cadres normatifs. En réponse à ces besoins, plusieurs standards et protocoles, avec de successives modifications et corrections, ont été publiés dans une période relativement brève. Parmi les plus populaires, SIP (Session Initiation Protocol), un protocole de signalisation, et RTP (Real-Time Transport Protocol), un protocole de transport de flots temps réel, se démarquent et ils sont au coeur de la majorité des applications conçues actuellement. Bien que, aujourd'hui, SIP et RTP sont liés fortement à la téléphonie sur IP, leur portée et leurs possibilités sont beaucoup plus vastes, ce qui déclenche un grand intérêt et justifie l'effort mis dans la conception des implémentations plus performantes et orientées plus spécifiquement à divers serveurs mandataires UA (User Agent). Dans ce contexte, le but du présent projet de maîtrise est de concevoir des piles de protocoles SIP et RTP orientées vers des applications de téléphonie sur IP, dans un environnement embarqué. Des conditions additionnelles sont que les piles doivent être codées en langage C et s'appuyer sur le système d'exploitation en temps réel MicroC/OS-II. Afin de faciliter la portabilité, il doit se prévoir des couches d'abstraction du matériel et du système d'exploitation. Même si les applications ciblées pour le projet sont, principalement, celles de VoIP, la pile SIP doit viser d'autres domaines, notamment des applications de domotique et de contrôle à distance. Cette dernière condition impose, de façon indirecte, d'autres conditions sur la taille du code et la puissance de calcul demandée, car le matériel pour ces types d'applications est d'habitude plus simple et moins puissant que les ordinateurs qui sont souvent utilisés dans les applications professionnelles de communication. Ce mémoire, qui décrit le travail effectué, est organisé en deux parties. La première fait une introduction théorique à la téléphonie sur IP, et sert de fondement à la deuxième partie, où la mise en oeuvre des protocoles SIP et RTP est décrite en détail.L'accent a été mis sur les justifications des décisions prises pendant toute la conception afin d'aider à mieux comprendre la logique appliquée et de permettre sa reconsidération et analyse dans de futures itérations. Comme résultat des contraintes et limitations imposées dans le cadre de ce projet, les piles de protocoles conçues se sont révélées très compactes et performantes, ce que justifie pleinement la continuité du travail dans l'avenir.

This experiment was growing CZTSe (Cu2ZnSnSe4) single phase thin film by using rapid thermal selenization process on Se/Cu/Sn/Zn/SLG thin film. It can complete the reaction to avoid Cu2SnSe3 appearing during the RTP. To discuss the effect of nitrogen and selenium flow rates on the thin film quality and adhesion, and to confirm the composition of the CZTSe single phase thin film. And I also annealed and changed the Se/Cu/Zn/Sn/SLG stacked layer to improve the thin film uniformity. Finally, it was stacked on the Mo/SLG and annealed by varying raising rate of temperature in order to enhance the adhesion. The results indicated that the various flow of nitrogen could cause different conditions. The element selenium easily escaped due to lower nitrogen flow could not provide enough outer pressure; larger nitrogen flow carried the extra high vapor pressure gas go through the surface of substrate and lead to the worse adhesion with the substrate. Unfortunately, using the analysis of X-ray diffraction and Raman spectrum couldn¡¦t determine the single phase of CZTSe. However, it has been confirmed by optical bandgap. In this experiment, the compositions of CZTSe single phase are found to be Cu-poor and Zn-poor, the optical bandgap (Eg) is 0.88~1.04 eV, and the resistivity (£l) is 1~10-2 £[-cm. By the rapid thermal selenization process, because of the rapid gradient of temperature, it brought out the diffused non-uniformly among the precursors. Therefore, the uniformity of thin films would not be perfect. As the result, re-annealing and change are the efficient methods to improve the uniformity of the thin film. The problems are un-wetting effect exists in the stacked liquid phase Se/Cu/Zn/Sn and the sodium glass substrates. The morphologies of the thin films are island connection. Finally, increasing the temperature in 15 oC per second and annealing the thin film for one minute at 250 oC with the stacked layer of Se/Cu/Sn/Zn/Mo/SLG are successful to promote the adhesion between the CZTSe and Mo.

Multicast is an effective technology to save network bandwidth while transmitting data packets in WiFi networks. However, it lacks of a mechanism to recover lost packets. When interference or congestion occurs in wireless networks, the receiver cannot receive complete data. In this Thesis, we propose a multicast retransmission mechanism which integrates NACK in WiFi networks. Also, the sender can dynamically adjust retransmission parameters according to the network status reported from the receiver. To demonstrate our proposed mechanism, three parts are implemented on the Linux platform. First, the RTP header is added in each packet and the buffer is maintained in both the sender and the receiver before the sender can transmit data. Second, a module is installed in the wireless card driver to examine the CRC errors caused by interference. Then the sender can be notified by NACK such that the sender may know the reasons of packet loss and the numbers of dropped packets. Third, the sender utilizes RTCP to calculate RTT and derive RTO. The retransmission time is calculated based on RTO, interferences, and congestion. Finally, our measurement results show that the sender can effectively use multicast retransmission according to various background traffics. Besides, the number of retransmitted packets can be significantly reduced.

The lithium tantalite [LiTaO3,abbreviated to LT] thin films were deposited on Pt/SiO2/Si substrates by spin coating with sol-gel processing and rapid thermal processing in this thesis. 1,3 propanediol was used as solvent to minimize the number of cycles of spin coating and drying processes to obtain the desired thickness of thin film. By changing the heating rate (600~3000¢J/min) and the heating temperature (500~800¢J), the effects of various processing parameters on the thin films growth are studied. The effects of various heating rate on the response of pyroelectric IR detector devices are studied also. Experimental results reveal that the heating rate will influence strongly on grain size, dielectricity, ferroelectricity and pyroelectricity of LT thin films. With the increase of heating rate, the grain size of LT thin film decreases slightly, and the C-axis orientation is enhanced. The relative dielectric constant of LT thin film increases from 28 up to 45.6, the tand increases from 0.033 to 0.134, Ec increases from 122 KV/cm to 183 KV/cm, Pr increases from 7.45 mC/cm2 to 12.12 mC/cm2, and g increases from 9.33´10-9 C/cm2K up to 2.66´10-8 C/cm2K, respectively, as the heating rate increases form 600 up to 3000¢J/min. In addition, the results also show that the LT thin film possesses the largest figures of merit Fv (2.19¡Ñ10-10 Ccm/J) and Fm (4.01¡Ñ10-9 Ccm/J) at the heating temperature of 700¢J and heating rate of 1800¢J/min. The voltage responsivities (Rv) measured at 80 Hz increase from 5496 to 8455 V/W and the specific detecivities (D*) measured at 300 Hz increase from 1.94¡Ñ108 to 2.38¡Ñ108 cmHz1/2/W with an increase of heating rate from 600 to 1800¢J/min. However, the voltage responsivity and the specific detecivity decrease with heating rate in excess of 1800¢J/min. The results show that LT1800 pyroelectric thin film detector exists both the maximums of voltage responsivity and specific detecivity. Therefore, LT1800 thin film exhibits the best IR characteristics for detector material.

Multimedia streaming is one of the killer applications for cellular communications. Although the 3GPP proposes the Packet-Switched Streaming (PSS) protocol to support the multimedia streaming services, the performances are still not good enough. To improve the performances, most of the researches divide the cellular networks into two parts: the wired network and the wireless network, and focus on the wireless network part. Hence the adaptive streaming was proposed. It utilizes the RTCP feedback in RTP to monitor the wireless network, and makes appropriate transmission parameters adjustments to prevent from buffer underflow and packet loss. But the overall performance may not be only limited by the wireless network part. Also, with high mobility, the link quality may be influenced severely by many factors. Hence, we will evaluate the performance of streaming services over 3G cellular networks on the train of Taiwan high speed railway in this paper and propose a mechanism to improve the performance.

Made available in DSpace on 2014-06-12T15:58:56Z (GMT). No. of bitstreams: 2 arquivo4774_1.pdf: 2496019 bytes, checksum: feac7bc4e882b9034f0e91a01ab90df8 (MD5) license.txt: 1748 bytes, checksum: 8a4605be74aa9ea9d79846c1fba20a33 (MD5) Previous issue date: 2004 / Com o advento das Redes de Computadores e a sua popularização, principalmente nas décadas de 80 e 90, a tradicional, segura e robusta arquitetura centralizada personificada nos mainframes, deu lugar rapidamente a uma arquitetura descentralizada que além correr atrás do prejuízo quando se analisa os benefícios já consolidados da arquitetura centralizada, ainda se volta a partir da década de 90 para a disponibilização de serviços ditos de nova geração, caracterizados por tratar além da mídia textual, ainda áudio e vídeo. Comparativamente ao ambiente complexo, amplo e heterogêneo observado na Internet, o ambiente de uma Intranet é bem comportado para se disponibilizar os serviços de nova geração uma vez que o universo de equipamentos e tecnologias envolvidos, para se garantir os requisitos necessários aos novos serviços, é previsível. A presente dissertação possui como principal contribuição o desenvolvimento de uma infra-estrutura de transcodificação e a implementação de uma aplicação para distribuição de áudio em uma rede corporativa (Intranet) adaptável ao estado da rede que se utiliza da infra-estrura elaborada. Embora focada na distribuição apenas da mídia de áudio, a infraestrutura proposta pode evoluir para permitir a criação de aplicações de distribuição de áudio e vídeo e ainda áudio e vídeo conferências. Usa-se o protocolo de Aplicação/Transporte RTP - Real Time Transport Protocol, disponível em vários formatos de áudio existentes na API JMF - Java Media Framework, além de se abordar as necessidades a nível de rede, tais como Multicast e mecanismos de priorização de quadros e pacotes nas camadas 2 e 3 do modelo de referência OSI (i.e. IEEE 802.1P e DiffServ). São identificados quatro Casos de Uso básicos na infra-estrutura proposta: Transmitir ao Transcodificador; Iniciar Sessões de Áudio no Transcodificador; Receber Stream de Áudio RTP no Receptor; e, Trocar a Sessão Multicast de Recepção. Os três primeiros são especificados, implementados e validados; o quarto, é apenas especificado, sendo a sua implementação e validação em cenário com filas diferenciadas de priorização de pacotes na camada de rede, um trabalho futuro. O trabalho explora o conceito de transcodificação como ideal para lidar com a heterogeneidade de recursos de rede existentes em uma Intranet

Because Vehicular Ad Hoc Networks (VANETs) often have a high packet loss rate, the formerly used protocol for video streaming, Real-time Transport Protocol (RTP), is no longer suitable for this specific environment. Previous conducted research has offered many new protocols to solve this problem; however, most of them cannot make full use of the existing Internet video streaming resources like RTP servers. Our work proposes two models to solve this compatibility issue. The first model is called the converter model. Based on this model, we first modify RTP using Erasure Coding (EC) technique in order to adapt it to the high packet loss rate of VANETs. This newly developed protocol is called EC-RTP. And, we then developed two converters. The first converter stands on the boundary between the Internet and VANETs. It receives the RTP packets which sent from Internet. And then it translates them to the EC-RTP packets. These packets are transported over the VANETs. The second converter receives these EC-RTP packets, translates them back to the RTP packets. It then sends them to the RTP player, so that the RTP player can play these packets. To make EC-RTP can carry more kinds of video streams other than RTP, we proposed a second model. The second model is called the redundancy tunnel. Based on this model, we let the protocol between the two converters carry RTP protocol as its payload. We use the same technique as we have used to modify RTP. At last, we did some experiments with Android tablets. The experiment results show our solution can use the same player to play the same video resources as RTP does. However, unlike RTP, it can reduce packet loss rate.