This thesis presents an intelligent early packet discard (I-EPD) for real-time video streaming over a multi-hop ad hoc network. In a multi-hop ad hoc network, the quality of transferring real-time video streams could be seriously degraded, since every intermediate node (IN) functionally like forwarding device does not possess large buffer and sufficient bandwidth. Even worse, a selected forwarding node could leave or power off unexpectedly which breaks the route to destination. Thus, a video packet temporarily buffered in intermediate nodes may exceed its time constraint when either a congested or failed link occurs; a stale video packet is useless even if it can reach destination after network traffic becomes smooth or failed route is reconfigured. In the proposed I-EPD, an IN can intelligently determine whether a buffered video packet should be discarded based on an estimated time constraint which is calculated from the RTP timestamps and the round trip time (RTT) measured by RTCP. For the purpose of validation, we implement the I-EPD scheme on a Linux-based embedded system. We compare the quality of video streams under different bit rates and different route repair time. In addition, we use PSNR to validate the quality of pictures from the aspect of application layer. The experimental results demonstrate that with I-EPD buffer utilization on IN can be more effectively used and unnecessary bandwidth wastage can be avoided.

Peer-to-Peer (P2P) technology is a kind of decentralized framework which is applied to file sharing and video streaming in Internet predominantly. However, there are some problems, like free-riding, fairness, unexpected fluctuation, in P2P systems. Especially, without free riding controlling, many users stop contributing data and only want to receive data. To avoid these problems, a credit-based incentive mechanism using sneak, public-key cryptography and multi-source technique is proposed to make peers in system have no choice but to share data. In the proposed incentive mechanism, the concept of credit is employed. If a peer can contribute more bandwidth for sharing stream, it is arranged closer to video server, to get more credit in return, have less packet loss risk and receive better stream quality. Free-rider is a peer who denies contributing or contributes less resource than it registers to the system. In the proposed mechanism, free-riders are judged by the credits of each peer. In order to prevent illegal behaviors, the public-key cryptography is used for differentiating real and fake credit, also used for tracing credits. Experimental results show that when 90% peers are zero uploading peers, using the proposed mechanism can reduce 85% of zero uploading peers. When 90% peers are cheating peers, adopting the proposed mechanism can reduce 94% of cheating peers. System utility is usually about 100%. Maximum number of hops to video server is about five to six which is much smaller than without using the proposed mechanism, about 40 hops. The loads in managing system topology and delivering sub-streams are reduced. Besides, the resource of the system is used effectively.

The number and scope of data driven streaming applications is growing. Such streaming applications are promising targets for effectively utilizing multi-cores because of their inherent amenability to pipelined parallelism. While existing methods of orchestrating streaming programs on multi-cores have mostly been static, real-world applications show ample variations in execution time that may cause the achieved speedup and throughput to be sub-optimal. One of the principle challenges for moving towards dynamic pipeline balancing has been the lack of approaches that can predict upcoming dynamic variations in execution efficiently, well before they occur. In this thesis, we propose an automated dynamic execution behavior prediction approach based on compiler analysis that can be used to efficiently estimate the time to be spent in different pipeline stages for upcoming inputs. Our approach first uses dynamic taint analysis to automatically generate an input-based execution characterization of the streaming program, which identifies the key control points where variation in execution might occur with respect to the associated input elements. We then automatically generate a light-weight emulator from the program using this characterization that can predict the execution paths taken for new streaming inputs and provide execution time estimates and possible dynamic variations. The main challenge in devising such an approach is the essential trade-off between accuracy and overhead of dynamic analysis. We present experimental evidence that our technique can accurately and efficiently estimate dynamic execution behaviors for several benchmarks with a small error rate. We also showed that the error rate could be lowered with the trade-off of execution overhead by implementing a selective symbolic expression generation for each of the complex conditions of control-flow operations. Our experiments show that dynamic pipeline balancing using our predicted execution behavior can achieve considerably higher speedup and throughput along with more effective utilization of multi-cores than static balancing approaches.

<p>Examensarbetet bakom denna rapport utgör en undersökande studie om hur transmission av mediadata i nätverk kan göras effektivare. Det kan åstadkommas genom att viss tilläggsinformation avsedd för att jämna ut datatakten adderas i det realtidsprotokoll, Real Time Protocol, som används för strömmande media. Genom att försöka skicka lika mycket data under alla konsekutiva tidsintervall i sessionen kommer datatakten vid en godtycklig tidpunkt med större sannolikhet att vara densamma som vid tidigare klockslag. En streamingserver kan tolka, hantera och skicka data vidare enligt instruktionerna i protokollets sidhuvud. Datatakten jämnas ut genom att i förtid, under tidsintervall som innehåller mindre data, skicka även senare data i strömmen. Resultatet av detta är en utjämnad datataktskurva som i sin tur leder till en jämnare användning av nätverkskapaciteten.</p><p>Arbetet inkluderar en översiktlig analys av beteendet hos strömmande media, bakgrundsteori om filkonstruktion och nätverksteknologier samt ett förslag på hur mediafiler kan modifieras för att uppfylla syftet med examensarbetet. Resultat och diskussion kan förhoppningsvis användas som underlag för en framtida implementation av en applikation ämnad att förbättra trafikflöden över nätverk.</p>

<p>Abstract</p><p>This paper investigates how to individualize the tuition in English by using ability groupings. The study is primarily focusing on the group of under-achieving ability pupils since teachers have a special responsibility for those pupils who experience difficulties in attaining the goals that have been set for their education, according to the Curriculum for compulsory school Lpo 94. The main research question was if the groupings were positive or negative for the under-achieving pupils at school X, i.e. the compulsory school under investigation. In addition I also wanted to find out if the ability groupings at school X responded to the steering documents, i.e. the Curriculum in compulsory school and the Education Act. The study is based on primary data in the forms of interviews, with four English teachers, and questionnaires, answered by the under-achieving ability pupils.</p><p>Conclusions were drawn that the ability groupings seemed to be mostly negative concerning the under-achieving pupils. The groupings at school X are not supported in any steering documents either since the groupings are not temporary, they do to some extent put the pupils at a certain grade level and on a certain track which violates the Education Act. One of the conclusions in this paper is also that there should be clearer guidelines on how to approach this phenomenon.</p>

This presentation provides an overview of the current production and broadcasting system of Radio UNiCC, the University Radio in Chemnitz. Furthermore, several significant technologies are discussed. In the end there are some more examples of audio streaming. This presentation was broadcasted into the internet and was recorded as well. The recording file is available. / Der Vortrag beschreibt das aktuelle Produktions- und Sendesystem von Radio UNiCC, dem Uniradio an der TU Chemnitz. Weiterhin werden Schlüsseltechnologien des Audiostreamings gegenübergestellt. Abgerundet wird der Vortrag durch weitere Anwendungsbeispiele von Audiostreaming. Der Vortrag wurde live übertragen und aufgezeichnet. Der Mitschitt steht zur Verfügung.

Für Übertragung von Live-Videodaten an mehrere Empfänger gibt es als etablierte Technologien die Übertragung per Multicast an mehrere Empfänger sowie die Unicast-Übertragung an jeden Empfänger von einem Verteilpunkt aus. Nachteilig ist im ersten Fall, dass viele ISPs kein Multicast unterstützen, im zweiten Fall der hohe Bandbreitenbedarf am Verteilpunkt. Die Unterschiede in der verfügbaren Bandbreite von ISP-Zugängen (DSL, ISDN) und Campusnetzwerken (z.B. Studentennetzen) erfordern es, das Videomaterial in mehreren Qualitätsstufen anzubieten. Im Team wurde ein System entworfen und realisiert, das aus einer Quelle effizient verschiedene Qualitätsstufen des Videomaterials erzeugt ("Videoteil") und dieses mittels Peer-To-Peer-Technologie verteilt ("Netzwerkteil"). Diese Arbeit enthält den den Netzwerkteil.

Der Vortrag zeigt Möglichkeiten und Wege der Verteilung von Audio und Video im Web. Die Betrachtung und Einordnung verschiedener Streamingverfahren erfolgt an Beispielen. Ausgewählte Serverlösungen und geeignete Player werden steckbrieflich vorgestellt.

Includes bibliographical references (p. 34-37).

Includes bibliographical references (p. 31-34).