Thanks to their increasing sophistication and popularity, mobile devices, in the form of smartphones and tablets, have become the fastest growing contributors to Internet traffic. Indeed, smartphones are projected to account for 50% of global Internet traffic by 2017, with the share of mobile video increasing to about 40% of total Internet traffic. As users embrace Internet streaming of video, several studies have found that a small decrease in video quality leads to a substantial increase in viewer abandonment and disengagement rates. To handle the explosive growth in video traffic, Adaptive HTTP streaming, which exploits the prevalence of commodity web servers and content distribution networks, has emerged as the key technology for delivering video to end users. Although a number of systems have been proposed for HTTP video streaming in traditional environments and for fixed clients, existing platforms for video streaming on mobile devices are still in their infancy and do not address the additional challenges often experienced by mobile clients: high fluctuations in network conditions, heterogeneous networking interfaces, multiple form-factors, and limited battery life. In this dissertation, we propose a number of solutions for improving the Quality of Experience of HTTP video streaming on mobile devices. We begin by evaluating the performance of several existing video quality adaptation schemes when deployed on mobile platforms. Through experiments with smartphones in wide-area environments, we assemble several key findings. First, we show that the high fluctuations in network throughput on cellular and Wi-Fi networks impose significant challenges for efficiently architecting the video adaptation scheme. Second, we find significant differences between the performance of the current state-of-the-art schemes in controlled experimental settings and their performance in mobile settings on key quality metrics such as inefficiency, instability, rebuffering ratio, and startup latency. We also find noticeable differences in the behavior of the schemes under Wi-Fi and cellular network access, with most of the schemes performing worse when the network access is cellular. Given these observations, we hypothesize on the possible causes of these inefficiencies. We also identify the best practices of existing schemes and key insights from experimental results that can serve as foundations for addressing many of the limitations. Armed with these measurement-driven insights, we propose a novel video quality adaptation scheme, called MASS, which is more robust to the vagaries of the wireless networking conditions. We implement and evaluate our solution on commodity Android smartphones, and demonstrate significant performance gains over existing schemes. To further improve the streaming experience, we introduce an extension to HTTP video streaming that leverages the synergy between social network participation and video streaming to optimize end-user Quality of Experience. Our system, called SDASH, integrates and applies well-known concepts such as cooperative caching, prefetching, and P2P streaming for reducing bitrate fluctuations and optimizing the viewing experience. Finally, we develop a general infrastructure for constructing temporally and spatially localized P2P communities of mobile devices sharing similar interests. The platform enables on-demand cooperation among mobile clients based on device context and client preferences. We use a concrete implementation of the mobile P2P infrastructure for evaluating the performance of SDASH. This dissertation addresses the challenges facing Adaptive HTTP Streaming under mobile networking conditions. Through experimentation with commodity mobile devices, we show that the proposed techniques for bitrate adaptation and cooperative streaming can significantly improve the video viewing experience.

Det senaste året har jag arbeta med analys av digital mediehantering, digital media som ett informationsverktyg samt tekniken bakom det. ’Hur kan man göra saker mer effektivt’, ’är webb-tv något för Försvarsmakten att använda sig av i syfte att göra sitt informationsflöde mer effektivt’, både internt och mot allmänheten. Föreliggande arbete tar upp tekniken bakom dessa områden, jag försöker även förklara hur relativt enkelt det kan vara att integrera dem i en existerande organisation utan att göra dyra omorganisationer. Jag har även inventerat AMF och FM media på deras utrustning, teknik och gjort en lista på vad som är användbart och på vad som bör tas bort inför en omstrukturering. Vidare föreslår jag var en viss utrustning bör lämpligast placeras för att göra mest nytta. Slutligen handlar detta arbete om att skapa en medvetenhet om digital media, och visa exempel på vad man kan göra med denna teknik i en organisation som Försvarsmakten och AMF.

Surface acoustic wave (SAW) devices find uses in a plethora of applications including but not limited to chemical, biological sensing, and microfluidic actuation. The primary aim of this dissertation is to develop a SAW biosensor, capable of simultaneous detection of target biomarkers in fluid media at concentrations of picogram/ml to nanogram/ml levels and removal of non-specific proteins from sensor surface using the process of acoustic streaming, for potential chemical sensing, medical, and clinical diagnostic applications. The focus is on the development of three dimensional finite element structural and fluid-structure interaction models to study wave propagation and acoustic actuation of fluids in a SAW biosensor. This work represents a significant improvement in understanding fluid flow over SAW devices, over the currently available continuum model of Nyborg. The developed methodology includes use of a novel substrate, namely, Langasite coupled with various combinations of novel multidirectional interdigital transducer (IDT) configurations such as orthogonal, focused IDTs as well as sensor surface modifications, such as micro-cavities. The current approach exploits the capability of the anisotropic piezoelectric crystal to launch waves of different characteristics in different directions, which can be put to the multiple uses including but not limited to sensing via shear horizontal waves and biofouling elimination via Rayleigh wave induced acoustic streaming. Orthogonal IDTs gives rise to constructive interference, thereby enhancing the magnitudes of device displacements and fluid velocities. The net effect is an increase in device sensitivity and acoustic streaming intensity. The use of micro-cavities in the delay path provides a synergistic effect, thereby further enhancing the device sensitivity and streaming intensity. Focused IDTs are found to enhance the device displacements and fluid velocities, while focusing the device displacements and fluid motion at the device focal point, thereby enhancing the SAW device biosensing performance. The work presented in this dissertation has widespread and immediate use for enhancing sensor sensitivity and analyte discrimination capabilities as well as biofouling removal in medical diagnostic applications of SAW sensors. This work also has a broad relevance to the sensing of multiple biomarkers in medical applications as well as other technologies utilizing these devices such as microfluidic actuation.

Submarine groundwater discharge (SGD) represents a significant portion of the total discharge from coastal aquifers through diffuse seepage and point source springs, but can be difficult to locate. SGD is important as it can be a source of nutrients to estuaries and other coastal ecosystems. In an effort to evaluate geophysical and thermal methods for detecting SGD on the Florida Gulf coast, a suite of water-borne surveys were run in conjunction with aerial thermal imagery over the lower Suwannee River and estuary in March and September 2005. Thermal imagery exploits temperature differences between discharging groundwater and surface water. Thermal images were collected in March (dry season), at the end of winter, and at night to maximize the differences between warm groundwater and colder surface water. Generally pore waters in zones of concentrated SGD should be fresher, and hence more resistive than "background" values. Marine streaming resistivity data can detect pore water resistivity variations and were collected alongside continuous 222Rn and CH4 sampling from surface waters. Naturally occurring tracers, 222Rn and CH4, are used as the "standard" against which resistivity and thermal images are compared. Based on the expected properties of discharging groundwater, we hypothesize that in zones of elected tracer concentrations, increased thermal image temperatures and increased terrain resistivities will be observed. The data set as a whole supports this hypothesis. However, regional-scale correlations are clearly and significantly influenced by factors other than SGD including thermal-image noise, the presence of the fresh/salt water interface, and a large regional tracer gradient generated by a first-order spring. At local scales (tens to hundreds of meters) there are no significant correlations between thermal image temperature and tracer concentrations, due at least in part to flight-line edge effects that dominate the thermal imagery. After correcting for regional trends,significant correlations between tracer concentration and log resistivity exist only in a subset of the data that lies offshore. Because neither thermal imagery nor streaming resistivity data consistently support the hypotheses, this study suggests that neither method by itself is reliable for detecting SGD in this area

With the advent of mobile Internet the amount of time people spend with multimedia applications in the mobile environment is surging and demand for high quality multimedia data over the Internet in the mobile environment is growing rapidly. However the mobile environment is significantly more unfriendly than the wired environment for multimedia applications in many ways. Network resources are limited and the condition is harder to predict. Also multimedia applications are generally delay intolerant and bandwidth demanding, and with users moving, their demand could be much more dynamic and harder to anticipate. Due to such reasons many existing mobile multimedia applications show unsatisfactory performance in the mobile environment. We target three multimedia content delivery applications and optimize with limited and unpredictable network conditions typical in the mobile Internet environment. Vehicular networks have emerged from the strong desire to communicate on the move. We explore the potential of supporting high-bandwidth applications such as video streaming in vehicular networks. Challenges include limited and expensive cellular network, etc. Internet video conferencing has become popular over the past few years, but supporting high-quality large video conferences at a low cost remains a significant challenge due to stringent performance requirements, limited and heterogeneous client. We develop a simple yet effective Valiant multicast routing to select application-layer routes and adapt streaming rates according to dynamically changing network condition in a swift and lightweight way enough to be implemented on mobile devices. Bitrate adaptive video streaming is rapidly gaining popularity. However recent measurements show weaknesses in bitrate selection strategies implemented in today's streaming players especially in the mobile environment. We propose a novel rate adaptation scheme that classifies the network condition into stable and unstable periods and optimizes video quality with different strategies based on the classification. / text

Video traffic is growing rapidly in wireless networks. Different from ordinary data traffic, video streams have higher data rates and tighter delay constraints. The ever-varying throughput of wireless links, however, cannot support continuous video playback if the video data rate is kept at a high level. To this end, adaptive video transmission techniques are employed to reduce the risk of playback interruptions by dynamically matching the video data rate to the varying channel throughput. In this dissertation, I develop new models to capture viewers' quality of experience (QoE) and design adaptive transmission algorithms to optimize the QoE. The contributions of this dissertation are threefold. First, I develop a new model for the viewers' QoE in rate-switching systems in which the video source rate is adapted every several seconds. The model is developed to predict an important aspect of QoE, the time-varying subjective quality (TVSQ), i.e., the up-to-the-moment subjective quality of a video as it is played. I first build a video database of rate-switching videos and measure TVSQs via a subjective study. Then, I parameterize and validate the TVSQ model using the measured TVSQs. Finally, based on the TVSQ model, I design an adaptive rate-switching algorithm that optimizes the time-averaged TVSQs of wireless video users. Second, I propose an adaptive video transmission algorithm to optimize the Overall Quality (OQ) of rate-switching videos, i.e., the viewers' judgement on the quality of the whole video. Through the subjective study, I find that the OQ is strongly correlated with the empirical cumulative distribution function (eCDF) of the video quality perceived by viewers. Based on this observation, I develop an adaptive video transmission algorithm that maximizes the number of video users who satisfy given constraints on the eCDF of perceived video qualities. Third, I propose an adaptive transmission algorithm for scalable videos. Different from the rate-switching systems, scalable videos support rate adaptation for each video frame. The proposed adaptive transmission algorithm maximizes the time-averaged video quality while maintaining continuous video playback. When the channel throughput is high, the algorithm increases the video data rate to improve video quality. Otherwise, the algorithm decreases the video data rate to buffer more videos and to reduce the risk of playback interruption. Simulation results show that the performance of the proposed algorithm is close to a performance upper bound. / text

This thesis targets at modeling and resolving practical problems using mathematical tools in two representative networking systems nowadays, i.e., peer-to-peer (P2P) video streaming system and cloud computing system. In the first part, we study how to mitigate the following tussle between content service providers and ISPs in P2P video streaming systems: network-agnostic P2P protocol designs bring lots of inter-ISP traffic and increase traffic relay cost of ISPs; in turn, ISPs start to throttle P2P packets, which significantly deteriorates P2P streaming performance. First, we investigate the problem in a mesh-based P2P live streaming system. We use end-to-end streaming delays as performance, and quantify the amount of inter-ISP traffic with the number of copies of the live streams imported into each ISP. Considering multiple ISPs at different bandwidth levels, we model the generic relationship between the volume of inter-ISP traffic and streaming performance, which provides useful insights on the design of effective locality-aware peer selection protocols and server deployment strategies across multiple ISPs. Next, we study a similar problem in a hybrid P2P-cloud CDN system for VoD streaming. We characterize the relationship between the costly bandwidth consumption from the cloud CDN and the inter-ISP traffic. We apply a loss network model to derive the bandwidth consumption under any given chunk distribution pattern among peer caches and any streaming request dispatching strategy among ISPs, and derive the optimal peer caching and request dispatching strategies which minimize the bandwidth demand from the cloud CDN. Based on the fundamental insights from our analytical results, we design a locality-aware, hybrid P2P-cloud CDN streaming protocol. In the second part, we study the profit maximization and cost minimization problems in Infrastructure-as- a- Service (IaaS) cloud systems. The first problem is how a geo-distributed cloud system should price its datacenter resources at different locations, such that its overall profit is maximized over long-term operation. We design an efficient online algorithm for dynamic pricing of VM resources across datacenters, together with job scheduling and server provisioning in each datacenter, to maximize the cloud's profit over the long run. Theoretical analysis shows that our algorithm can schedule jobs within their respective deadlines, while achieving a time-averaged overall profit closely approaching the offline maximum, which is computed by assuming perfect information on future job arrivals is freely available. The second problem is how federated clouds should trade their computing resources among each other to reduce the cost, by exploiting diversities of different clouds' workloads and operational costs. We formulate a global cost minimization problem among multiple clouds under the cooperative scenario where each individual cloud's workload and cost information is publicly available. Taking into considerations jobs with disparate length, a non-preemptive approximation algorithm for leftover job migration and new job scheduling is designed. Given to the selfishness of individual clouds, we further design a randomized double auction mechanism to elicit clouds' truthful bidding for buying or selling virtual machines. The auction mechanism is proven to be truthful, and to guarantee the same approximation ratio to what the cooperative approximation algorithm achieves. / published_or_final_version / Computer Science / Doctoral / Doctor of Philosophy

Clustering algorithms are an important tool for data mining and data analysis purposes. Clustering algorithms fall under the category of unsupervised learning algorithms, which can group patterns without an external teacher or labels using some kind of similarity metric. Clustering algorithms are generally iterative in nature and computationally intensive. They will have disk accesses in every iteration for data sets larger than memory, making the algorithms unacceptably slow. Data could be processed in chunks, which fit into memory, to provide a scalable framework. Multiple processors may be used to process chunks in parallel. Clustering solutions from each chunk together form an ensemble and can be merged to provide a global solution. So, merging multiple clustering solutions, an ensemble, is important for providing a scalable framework. Combining multiple clustering solutions or partitions, is also important for obtaining a robust clustering solution, merging distributed clustering solutions, and providing a knowledge reuse and privacy preserving data mining framework. Here we address combining multiple clustering solutions in a scalable framework. We also propose algorithms for incrementally clustering large or very large data sets. We propose an algorithm that can cluster large data sets through a single pass. This algorithm is also extended to handle clustering infinite data streams. These types of incremental/online algorithms can be used for real time processing as they don't revisit data and are capable of processing data streams under the constraint of limited buffer size and computational time. Thus, different frameworks/algorithms have been proposed to address scalability issues in different settings. To our knowledge we are the first to introduce scalable algorithms for merging cluster ensembles, in terms of time and space complexity, on large real world data sets. We are also the first to introduce single pass and streaming variants of the fuzzy c means algorithm. We have evaluated the performance of our proposed frameworks/algorithms both on artificial and large real world data sets. A comparison of our algorithms with other relevant algorithms is discussed. These comparisons show the scalability and effectiveness of the partitions created by these new algorithms.

Surface acoustic wave (SAW) sensors detect chemical and biological species by monitoring the shifts in frequency of surface acoustic waves generated on piezoelectric substrates. Incorporation of nanomaterials having increased surface area as sensing layer have been effective in improving the sensitivity as well as miniaturization of SAW sensors. Selectivity, sensitivity and speed of response are the three primary aspects for any type of sensor. This dissertation focuses on design and development of SAW devices with novel transducer configurations employing nanomaterial sensing layers for enhanced sensing, improved selectivity, and speed of response. The sensing mechanism in these SAW sensors is a complex phenomenon involving interactions across several different length and time scales. Surface acoustic wave propagation at the macro-scale is influenced by several kinetic phenomena occurring at the molecular scale such as adsorption, diffusion, reaction, and desorption which in turn depend on the properties of nanomaterials. This suggests the requirement of a multi-scale model to effectively understand and manipulate the interactions occurring at different length scales, thereby improving sensor design. Sensor response modeling at multiple time and length scales forms part of this research, which includes perturbation theories, and simulation techniques from finite element methods to molecular-level simulations for interpreting the response of these surface acoustic wave chemical and biosensors utilizing alloy nanostructures as sensing layers. Molecular modeling of sensing layers such as transition metal alloy nanoclusters and nanowires is carried out to gain insights into their thermodynamic, structural, mechanical and dynamic properties. Finite element technique is used to understand the acoustic wave propagation at the macroscale for sensing devices operating at MHz frequencies and with novel transducer designs. The findings of this research provide insights into the design of efficient surface acoustic wave sensors. It is expected that this work will lead to a better understanding of surface acoustic wave devices with novel transducer configurations and employing nanomaterial sensing layers.

ITC/USA 2012 Conference Proceedings / The Forty-Eighth Annual International Telemetering Conference and Technical Exhibition / October 22-25, 2012 / Town and Country Resort & Convention Center, San Diego, California / We propose a spectrally-efficient multiple-access technique that is particularly suitable for aeronautical telemetry applications involving serial streaming of data from multiple test articles to a ground station. Unlike conventional frequency-division multiple access, we assign overlapping frequency bands to different users with a minimum carrier separation corresponding to the symbol rate. We utilize multiuser detection strategies at the ground station to separate the transmissions from different test articles. As shown by the simulation results, the proposed scheme is robust to large frequency offsets due to oscillator offsets and Doppler shifts commonly encounters in aeronautical telemetry applications.