Volumetric video enables fully immersive free-viewpoint experiences but imposes bandwidth and computation costs that prevent real-time deployment on everyday devices. Existing codecs achieve strong compression but struggle to meet live-streaming latency budgets, while super-resolution methods often depend on GPU acceleration or costly training. We present Super-VV, an end-to-end streaming framework that is light on bits yet rich on quality. Super-VV combines three key components: (i) a quality-preserving downsampler that prunes redundant geometry while retaining perceptual detail, (ii) a Fast Encoder/Decoder that extends Draco with tile-parallel and in-memory optimizations, and (iii) a lightweight, CPU-based upsampler that uses k-nearest-neighbors interpolation to restore dense geometry and color without GPUs or training. Together, these modules sustain real-time throughput of up to 73 FPS on commodity CPUs, while reducing bandwidth by 98% compared to raw transmission and maintaining compelling visual quality (PSNR of 42.75 dB) compared to the original volumetric videos.

Volumetric video promises rich 6DoF experiences, but current 3D Gaussian Splatting (3DGS) pipelines are inefficient for dynamic content: each frame is trained independently, ignoring temporal redundancy, which leads to high storage and bandwidth requirements. In this paper, we propose CubeGS, a lightweight, streaming-oriented compression mechanism for 3DGS video that preserves visual fidelity and rendering efficiency of 3DGS while substantially reducing the data footprint. CubeGS organizes frames using a Group-of-Pictures (GoP) structure, introducing reference frames and predictive frames within dynamic 3DGS sequences. Central to our design is cube-wise motion residual capturing, which reuses Gaussian parameters across time and encodes only localized changes in appearance and geometry. To further reduce storage and transmission cost, we integrate a modified Draco geometry encoder tailored to Gaussian parameterization. Together, these components make on-demand Gaussian Splatting streaming feasible over commercial networks with CubeGS averaging 193 Mbps for streaming a 24 fps video. CubeGS achieves approximately 7.7× data reduction and up to 16.9% higher video quality compared to 3DGStream, the state-of-the-art dynamic 3DGS method. Importantly, CubeGS retains the rendering speed of 3DGS while scaling efficiently to complex dynamic scenes with much lower storage and bandwidth overhead than existing approaches.

Recent improvements in adaptive video streaming have been significant in the areas of adaptation algorithms and encoding (e.g., emergence of AV1 codec). A promising but yet unexplored feature of AV1 is the switching frames (S-frames). In this paper, we explore the pros and cons of S-frames in DASH streaming and demonstrate that S-frames can improve the compression and quality switching capabilities of AV1. We conduct a measurement study to understand the extent of benefits of S-frames and find that 10% – 50% data savings are achievable for many encoding configurations where S-frame use is maximized. We further propose a stream structure where the number of S-frames used balances QoE improvement with the ability to play or seek randomly within the video in challenging network conditions. We evaluate the streams using S-frames in a DASH player under a variety of conditions. We find that when S-frame use and switching opportunities are both maximized, they significantly improve stall performance, the critical aspect of the Quality of Experience (QoE), by reducing stall time by at least 88 % in realistic bandwidth conditions.

360° video streaming is central to immersive applications such as virtual reality, education, and telepresence, yet delivering stall-free playback with high viewport quality remains difficult under fluctuating bandwidth and inaccurate viewport prediction. Prior solutions either fail to guarantee stall-free playback, introduce prohibitive overhead, or neglect prediction inaccuracies. We propose Opt360, a DASH-compliant optimization framework that generalizes tile assignment into multi-tier priority zones, incorporates prediction accuracy and window length directly into the optimization, and enforces hard constraints on stalls and quality switches. The resulting mixedinteger formulation, coupled with a segment-internal tile scheduler, adapts to diverse viewport models while remaining real-time feasible. Extensive evaluations demonstrate that Opt360 ensures smooth playback, remains resilient to viewport variations, and effectively utilizes bandwidth for improved video quality, even under challenging network conditions.

Volumetric video streaming, which presents dynamic 3D objects captured from multiple angles, significantly increases video data size and complexity compared to traditional 2D video, posing major challenges for efficient transmission. Current encoding methods either introduce excessive latency (projectionbased encoding) or inefficiency in bandwidth usage (direct 3D encoding). To address these limitations, this paper introduces TARS, a temporal-spatial adaptive streaming solution that exploits inter-frame correlations in dynamic point-cloud videos. TARS employs a Field-of-View (FoV)-aware approach to intelligently avoid the retransmission of redundant regions across consecutive frames, leveraging a specialized Point Cloud Structural Similarity metric for precise similarity assessment. Experimental results demonstrate that TARS effectively achieves bandwidth savings of up to 66 % while maintaining visual fidelity and achieving minimal quality degradation-with a Mean Square Error (MSE) as low as 0.145 under conservative settings. Additionally, the approach significantly optimizes decoding efficiency, resulting in up to 2.42 times increase in decoding frame rates compared to transmitting and decoding videos with independently coded frames (e.g., the default Draco coding), highlighting its suitability for real-time and bandwidth-constrained volumetric streaming scenarios.

The growing demand for interactive and personalized visual experiences, particularly in the metaverse era, highlights the importance of three-dimensional volumetric video. However, its large size and complex processing requirements pose significant challenges for practical streaming, often leading to resource waste and quality issues. To address this, we propose OptVV, an optimization framework for high-quality and bandwidth-efficient volumetric video streaming. The framework comprises three key components: (1) Adaptive QoE optimization, dynamically adjusting video quality based on network conditions, playback requirements, and volumetric video characteristics to ensure smooth playback; (2) an optimized DASH scheme for enhanced quality adaptation and bandwidth efficiency; and (3) optimal resource scheduling for real-time decoding, efficiently managing computational resources to maintain playback smoothness. Evaluation results show that OptVV significantly improves streaming performance, achieving up to 80% bandwidth savings, 177% higher viewport quality, 83% fewer playback stalls, and 72% faster decoding time compared to existing methods.

With the increasing demand for immersive experiences, volumetric video has emerged as a critical technology, offering users six degrees of freedom (6DoF) to fully explore three-dimensional scenes. However, despite significant advancements, there remains a lack of a comprehensive and flexible adaptive streaming framework capable of delivering volumetric video over dynamic network conditions. To address this gap, we present VV-DASH, an end-to-end framework for adaptive volumetric video streaming over DASH (Dynamic Adaptive Streaming over HTTP). Our framework covers the entire streaming pipeline, from video source to video playback. We propose a codec-agnostic DASH Volumetric Video (DVV) segment format that consolidates compressed video content into DASH-ready segments. This segmentation improves achievable streaming throughput by 13.2%, effectively reduces bandwidth demand, and enhances the achievable streaming bitrate by up to 37.8%. In summary, VV-DASH provides a practical, high-performance framework for scalable and adaptive volumetric video streaming.

Live, real-time volumetric video streaming enables immersive remote communication by transmitting 3D representations of participants, but faces significant bandwidth and computational challenges on consumer hardware. We introduce a distance-aware volumetric video streaming method that prunes point cloud data in real time based on both real-world and virtual viewing distances before any conventional compression or tiling is applied. Our prototype, RealityStream, uses a single Kinect v1 for capture, referencing a precomputed VMAF-driven distance table to guide adaptive downsampling for transmission. This effectively reduces bandwidth demand by up to 65% while maintaining acceptable visual fidelity. Implemented on a commodity laptop and streamed to a standalone VR headset, RealityStream is shown to be practical, achieving end-to-end latencies of around 75 ms. These findings highlight that taking into account both real and virtual distance is a low-complexity approach for efficient, real-time volumetric video capture and streaming. RealityStream will benefit existing volumetric streaming platforms, making real-time 3D communication more efficient over networks.

Volumetric video is transforming immersive media and remote interaction, enabling applications ranging from holographic telepresence to augmented reality concerts. Despite its potential across various industries, real-time delivery is challenged by substantial data volume and stringent requirements for low-latency, high-quality streaming. Previous research has addressed these issues through improved compression efficiency, faster encoding and decoding processes, and adaptive streaming protocols to address network variability. However, a practical solution for live volumetric video streaming remains elusive. This paper introduces LiV, a live Dynamic Adaptive Streaming over HTTP (DASH) system specifically designed for general volumetric video. LiV effectively balances bandwidth demands and computational efficiency, facilitating stall-free playback and high visual quality. Leveraging a parallelized execution of the Draco encoder and decoder, our evaluations demonstrate that LiV enables smooth, real-time streaming with enhanced visual fidelity within the user’s field-of-view (FoV). LiV also significantly reduces the bandwidth demand by up to 30%. These findings mark a significant step toward practical live volumetric video with broad implications for immersive media.

The demand for data science innovation is growing across all sectors. There is no shortage of new big-data applications and tools for deploying data science projects. Data scientists typically create data workflows manually or by writing scripts to automate parts of the process. However, basic scripting is insufficient for large-scale projects and performance optimizations. This poses great technical challenges for data scientists. In this paper, we present the Automated Low-code Flexible (ALF) Framework, designed to remove technical barriers and streamline data workflows. ALF also provides a holistic view of workflow construction and end-to-end workflow management. Its design is supported by a case study, demonstrating ALF’s ability to integrate a diverse array of applications, alongside performance enhancement tools and workflow inspection capabilities.

In recent years, live streaming of 360° videos has increased in popularity due to the emergence of virtual and mixed reality (VR/MR) applications. The high quality and bird-eye view characteristics of VR/MR pose real-time challenges for 360° video streaming. The tile-based 360° video streaming has been created based on Dynamic Adaptive Streaming over HTTP (DASH), mainly focusing on adaptation algorithms but ignoring the impact of coding properties and variations in video content on streaming quality. In this paper, we take on the challenge in a new direction through the exploration of the potential of CRF rate control. Our deep quality inspection of CRF transcoded videos lead to the proposal of a dynamic quality ladder for a more continuous quality provisioning in contrast to the conventional discrete quality levels. Our quality selection is based on visual quality and CRF rate control rather than just the resolution and bitrate in conventional DASH. We propose TYLE, a Tile-based dynamic quality enhancement for 360° video streaming. Without increasing the bandwidth demand, TYLE serves the content in the Field-of-View (FoV) at the highest quality level and content in the near-FoV region with improved visual quality compare to conventional DASH. TYLE maintains smoother and stabler playback, especially under the condition challenged by bandwidth under-provisioning and high motion-activity videos.

About one-third of autistic people are nonspeaking, and most are never provided access to an effective alternative to speech. Thoughtfully designed AR applications could provide members of this population with structured learning opportunities, including training on skills that underlie alternative forms of communication. A fundamental step toward creating such opportunities, however, is to investigate nonspeaking autistic people’s ability to tolerate a head-mounted AR device and to interact with virtual objects. We present the first study to examine the usability of an interactive AR-based application by this population. We recruited 17 nonspeaking autistic subjects to play a HoloLens 2 game we developed that involved holographic animations and buttons. Almost all subjects tolerated the device long enough to begin the game, and most completed increasingly challenging tasks that involved pressing holographic buttons. Based on the results, we discuss best practice design and process recommendations. Our findings contradict prevailing assumptions about nonspeaking autistic people and thus open up exciting possibilities for AR-based solutions for this understudied and underserved population.

Approximately one-third of individuals with autism are nonspeaking: They cannot communicate effectively using speech. Some traditional accounts suggest that these individuals cannot talk because they lack the symbolic capacity for language. And yet, recent studies have shown that these individuals’ cognitive abilities are vastly underestimated by standardized tests, and that difficulties with motor skills and movement contribute to their difficulty with speech. One consequence of the traditional accounts of nonspeaking autism is that life skills (rather than academic content) tend to be emphasized in schooling. Without access to meaningful academic content, their educational and vocational opportunities are significantly limited. Recent studies have proposed the use of head-mounted Augmented Reality (AR) applications as a means of providing engaging, customizable, and age-appropriate content to this population. Specifically, such applications can address the unique sensory and motor needs of nonspeaking autistic students, e.g., allow them to move freely around the room as they interact with lessons in the application. This paper describes the design and evaluation of the first AR application aimed to facilitate tailored educational experiences for nonspeaking autistic students. After extensive consultations with nonspeaking people, parents, and professionals, we developed our application to run on HoloLens 2 offering lessons and multiple-choice comprehension and spelling questions. We conducted a study involving five nonspeaking autistic participants and two specialized educators. Through a design critique process and an iterative design refinement approach, we show that most of our participants successfully interacted with the application and completed different types of lesson tasks. Based on quantitative data from the study sessions and qualitative feedback from participants and educators, we provide recommendations for UI and UX design that will promote the development and use of such software for this under-served and under-researched population.

The demand for real-time video streaming is increasing due to emerging live and interactive applications like virtual conferencing/collaboration and augmented/mixed reality. Real-time video transcoding and streaming face challenges, as inefficient transcoding can cause delays and hinder viewer QoE. In this paper, we propose Antifreeze, a complete end-to-end solution for real-time transcoding and streaming. Antifreeze includes a transcoding-aware adaptation algorithm that considers visual quality, bandwidth, buffer dynamics, and transcoding time to maximize client QoE. By dynamically adapting through on-the-fly transcoding, Antifreeze provides a personalized streaming experience. Results demonstrate that Antifreeze reduces playback stalls and improves visual quality in live video streaming sessions across different bandwidth profiles.

Traditional Adaptive BitRate (ABR) streaming faces a challenge of providing smooth experience under highly variable network conditions, especially when low latency is required. Effective adaptation techniques exist for deep-buffer scenarios, such as streaming long-form Video-on-Demand content, but remain elusive for short-form or low-latency cases, when even a short segment may be delivered too late and cause a stall. Recently proposed temporal adaptation aims to mitigate this problem by being robust to losing a part of the video segment, essentially dropping the tail of the segment intentionally to avoid the stall. In this paper, we analyze this approach in the context of a recently adopted codec AV1 and find that it does not always provide the promised benefits. We investigate the root causes and find that a combination of codec efficiency and TCP behavior can defeat the benefits of temporal adaptation. We develop a solution based on QUIC, and present the results showing that the benefits of temporal adaptation that still apply to AV1, including reduced stall time up to 65% compared to the original TCP-based approach. In addition, we present a novel way to use the stream management features of QUIC to benefit Quality-of-Experience (QoE) and reduce wasted data in video streaming.

The increasing demand for video streaming in all forms draws significant research and development attention, especially on the client-side for adaptive streaming services like DASH and HLS. However, the implementation challenges in developing and validating new client-side solutions within a full-stack video player pose a major obstacle. State-of-the-art open-source video players, such as DASH.js, VLC, and GPAC, were designed for specific purposes and are difficult to extend and modify for video streaming research. To address this issue, we propose iStream Player, a versatile video player framework featuring fully extendable and independent micro-modules similar to Lego blocks. Constructing a video player in iStream Player is as simple as assembling Lego pieces. Our case studies demonstrate that it is effortless to create a diverse range of players by making only minor changes, such as extending or replacing only one or two micro-modules. As a result, iStream Player significantly reduces the time and effort required to develop and validate new solutions, providing researchers and developers in the video streaming field with a shared platform to explore and to share their innovative ideas.

Official statistics indicate that internet users all around the world watch more videos and play more games during the COVID-19 pandemic than at any time [18]. This unprecedented, challenging situation demands solutions to accommodate rapid growth while maintaining and/or enhancing the video quality. This paper proposes SODA-Stream, an SDN-based optimization framework for enhancing Quality-of-Experience (QoE) in DASH streaming. The optimization framework max-imizes the number of concurrent streaming sessions that can be accommodated in a network and maximize streaming quality. The practical implementation of the framework utilizes the dynamic routing and bandwidth allocation enabled by Software Defined Networking (SDN). The evaluation results show that SODA-Stream significantly outperforms the conventional network routing and resource allocation algorithms, accepting 52% more sessions, 45% improvement in bandwidth allocation, and 70% reduction in bandwidth wastage, smoother playback, and better viewing experience.

Internet of Things (IoT) and its applications are proliferating. Performance and scalability are key aspects of an IoT system. A scalable IoT system can gracefully accommodate a growth in the number of IoT devices while still meeting performance requirements. This motivates the need for tools that allow the performance and scalability of an IoT system to be evaluated prior to deployment. Implementing the actual system at full scale and then evaluating it through measurements is ideal from the point of view of realism. However, such an approach can be expensive and inflexible. Emulating an IoT deployment on commodity hardware is an attractive alternative since it allows various system design alternatives to be evaluated in a flexible way without the need for expensive full scale implementations of the alternatives. Recently, many such IoT emulation platforms have been proposed. However, none of these platforms are appropriate for performance and scalability testing since they do not provide a mechanism to match the performance characteristics of the IoT devices being emulated. We present Oasis, a system that addresses this limitation. Oasis uses Docker containers to emulate IoT devices. It leverages Docker's resource allocation and network emulation capabilities to match the performance characteristics of an emulated device to that of its native counterpart. We also show that the ability to accurately match performance improves scalability as well.

Media content has become a primary source of information, entertainment, and even education. The ability to provide video content querying as well as interactive experiences is a new challenge. To this end, question answering (QA) systems such as Alexa and Google Assistant have become quite established in consumer markets but are limited to general information and lack context awareness. In this paper, we propose Context-QA, a light-weight context-aware QA framework, to provide QA experiences on multimedia content. The context awareness is achieved through our innovative Staged QA Controller algorithm that keeps the search for answers in the context most relevant to the question. Our evaluation results show that Context-QA improves the quality of the answers by up to 49% and uses up to 56% less time compared to the conventional QA model. Subjective tests show Context-QA improved results over conventional QA models, with 90% reporting enjoying this new media form.

Cloud computing provisions scalable resources for high performance industrial applications. Cloud providers usually offer two types of usage plans: reserved and on-demand. Reserved plans offer cheaper resources for long-term contracts while on-demand plans are available for short or long periods but are more expensive. To satisfy incoming user demands with reasonable costs, cloud resources should be allocated efficiently. Most existing works focus on either cheaper solutions with reserved resources that may lead to under-provisioning or over-provisioning, or costly solutions with on-demand resources. Since inefficiency of allocating cloud resources can cause huge provisioning costs and fluctuation in cloud demand, resource allocation becomes a highly challenging problem. In this paper, we propose a hybrid method to allocate cloud resources according to the dynamic user demands. This method is developed as a two-phase algorithm that consists of reservation and dynamic provision phases. In this way, we minimize the total deployment cost by formulating each phase as an optimization problem while satisfying quality of service. Due to the uncertain nature of cloud demands, we develop a stochastic optimization approach by modeling user demands as random variables. Our algorithm is evaluated using different experiments and the results show its efficiency in dynamically allocating cloud resources.

Interactive video streaming applications are becoming increasingly popular. To maintain the Quality of Experience (QoE) of an end user, interactive streaming platforms need to transcode a video stream, i.e., adapt the quality of the video content, to match the network conditions between the platform and the user as well as the device capabilities of the end user. Modern video codecs such as High Efficiency Video Coding (HEVC) require significant computational resources for transcoding operations. Consequently, there is a need for systems that can perform transcoding quickly at runtime to sustain the real-time performance required for interactive streaming while at the same time using just the right amount of computational resources for the transcoding operations. This paper addresses this need by designing and implementing CONTRAST, a Container- based Distributed Transcoding Framework for Interactive Video Streaming. For any given stream and transcoding resolution, CONTRAST exploits a profiling technique to automatically determine the degree of parallelism, Le., the number of processing cores, demanded by the transcoding process to sustain the stream’s frame rate. It then launches Docker containers configured with the required number of cores to perform the transcoding. Experiments using a set of realistic video streams show that CONTRAST is able to sustain the frame rate requirements for interactive streams in a more resource efficient manner compared to baseline techniques that do not consider the degree of parallelism. To the best of our knowledge, our paper is the first to establish best practices for implementing transcoding platforms for interactive streaming videos encoded using a modem video codec.

Users and operators of cloud-based Spark clusters often require quick insights on how the execution time of an application is likely to be impacted by the resources allocated to the application, e.g., the number of Spark executor cores assigned, and the size of the data to be processed. Existing techniques typically require extensive prior executions of the application under various resource allocation settings and data sizes to obtain an accurate model. In this paper, we explore the accuracy of a model with less prior executions of the application. Such a model can be useful for situations where quick predictions are required and little cluster resources are available for building a model. We use logs from two executions of an application with small sample data and different resource settings and explore the accuracy of the predictions for other resource allocation settings and input data sizes.

The Apache Spark cluster computing platform is being increasingly used to develop big data analytics applications. There are many scenarios that require quick estimates of the execution time of any given Spark application. For example, users and operators of a Spark cluster often require quick insights on how the execution time of an application is likely to be impacted by the resources allocated to the application, e.g., the number of Spark executor cores assigned, and the size of the data to be processed. Job schedulers can benefit from fast estimates at runtime that would allow them to quickly conFigure a Spark application for a desired execution time using the least amount of resources. While others have developed models to predict the execution time of Spark applications, such models typically require extensive prior executions of applications under various resource allocation settings and data sizes. Consequently, these techniques are not suited for situations where quick predictions are required and very little cluster resources are available for the experimentation needed to build a model. This paper proposes an alternative approach called PERIDOT that addresses this limitation. The approach involves executing a given application under a fixed resource allocation setting with two different-sized, small subsets of its input data. It analyzes logs from these two executions to estimate the dependencies between internal stages in the application. Information on these dependencies combined with knowledge of Spark's data partitioning mechanisms is used to derive an analytic model that can predict execution times for other resource allocation settings and input data sizes. We show that deriving a model using just these two reference executions allows PERIDOT to accurately predict the performance of a variety of Spark applications spanning text analytics, linear algebra, machine learning and Spark SQL. In contrast, we show that a state-of-the-art machine learning based execution time prediction algorithm performs poorly when presented with such limited training data.

With the ever growing popularity of video services, maintaining high Quality of Experience (QoE) of the end users with heterogeneous devices and network conditions is becoming more challenging. Each user requires content that matches with their device capability and network conditions. This motivates the need for flexible video transcoding, which enables changing the properties of videos on-the-fly to fit different users. However, the transcoding process is compute intensive especially when handling modern video coding standards such as High Efficiency Video Coding (HEVC) and supporting emerging applications such as live broadcasts. Consequently, there is a need for lightweight and resource-efficient systems that can perform transcoding quickly at real-time to sustain desired user QoE requirements. We design and implement a container-based video transcoding system to address this need. We experimentally show that our system can meet real-time transcoding while using less computational resources than a native transcoding approach. Our work also identifies container and transcoder parameters that can impact the overall performance of the proposed system.

To cope with diverse network conditions, HTTP Adaptive Streaming (HAS) enables video players to dynamically change the video quality throughout the video stream. However, effective adaptation that minimizes stalls and start-up time while maximizing quality and stability remains elusive, especially when available bandwidth is variable or multiple players compete for the bottleneck capacity. Conventional approach to adaptation is to make a decision on the next video segment quality based on hysteresis of prior throughput measurements. This approach is not robust to bandwidth fluctuation at small time scales, which can consequently lead to stalls, bandwidth waste, and unstable quality, mainly due to the inability to mitigate significant bandwidth reduction during the segment download. We propose BETA- Bandwidth-Efficient Temporal Adaptation, an agile approach that allows HAS players to refine the quality level within video segments on the fly, according to the actual bandwidth conditions experienced while downloading each segment. We define a new HAS-oriented transmission order of video frames within segments that facilitates decodability of partial frames and paves the way for changing the paradigm from discrete to continuous bitrate ladders for HAS. BETA can work with any adaptation algorithm or HAS player to significantly improve robustness and efficiency in dynamic network environments and for low-latency streams, as well as to dramatically reduce content storage and encoding infrastructure requirements. Our evaluation using the real player implementation shows that BETA improves video quality by up to 20%, reduces number of stalls by 20%-100% in nearly 80% of cases, and cuts down wasted bandwidth by 22%-100%.

On one hand, since the introduction of UHD (ultra-high definition) videos, e.g., 4K and 8K videos, it is becoming more resource and time intensive to transcode videos. On the other hand, the increasing demand for video streaming implies more videos need to be transcoded. These two facts motivate the need for techniques to speedup coding and transcoding time. In this paper, we propose Stride, the first distributed video transcoding system that leverages the Apache Spark big data platform. The design of Stride is transcoder agnostic, meaning it can adopt any transcoder implementation (e.g., FFMPEG) without any modification. We provide an experimental characterization of the impact of video transcoding and Spark configuration parameters to identify the optimal settings. We also compare Stride with competing approaches. Our results show that Stride achieves 3.27 times speedup when the computing power (i.e., the number of vCPUs in a cloud) is increased by a factor of 4, which is significantly higher than the other alternatives we explore. In particular, Spark's dynamic task scheduler allows Stride to reduce transcoding time by 19.86% compared to an implementation without Spark. Our benchmark study suggests that Stride can support transcoding from 4K to 1080p (full HD) at a rate matching the video bitrate using approximately only 24 virtual cores.

Along with the increasing demand for video stream ing, network service providers and video content providers are challenged to maximize their service in terms of both quantity and quality while guaranteeing Quality-of-Service (QoS). In this paper, we propose MaxStream, a SDN-based flow maximization framework. The framework exploits the global view of the network available at the SDN controller to formulate integer multi-commodity flow problems with objectives to maximize the number of streaming sessions accommodated in a network to improve providers revenue (quantity of the service) and to maximize bandwidth provisioning for QoS enhancement (quality of the service). Our simulation results confirm that on average, MaxStream accepts 13% more streaming sessions and offers 50% increase in bandwidth provisioning compared to variations of the widest-shortest path algorithm.

Cloud computing is a new era of computing that offers resources and services for Web applications. Selection of optimal cloud resources is the main goal in cloud resource allocation. Sometimes, customers pay more than required since cloud providers' pricing strategy is designed for the interest of the providers. Nonetheless, cloud customers are interested in selecting cloud resources to meet their quality of service (QoS) requirements. Thus, for the interest of both providers and customers, it is vital to balance the two conflicting objectives of deployment cost and QoS performance. In this paper, we present a cost-effective and runtime friendly algorithm that minimizes the deployment cost while meeting the QoS performance requirements. In other words, the algorithm offers an optimal choice, from customers' point of view, for deploying a Web application in cloud environment. The multi-objective optimization algorithm minimizes cost and maximizes QoS performance simultaneously. The proposed algorithm is verified by a series of experiments on different workload scenarios deployed in two distinct cloud providers. The results show that the proposed algorithm finds the optimal combination of cloud resources that provides a balanced trade-off between deployment cost and QoS performance in relatively low runtime.

Web services are now increasingly being hosted on public cloud infrastructure as a service platforms such as the Amazon Web service elastic compute cloud (EC2). However, previous studies have shown that the virtualized infrastructure used in public clouds can introduce contention among virtual machines (VMs) for shared physical host resources eventually leading to performance problems. Subscribers in a public cloud platform typically do not have access to metrics that can directly quantify the adverse impact of such inter-VM interference on Web service response times. We present a software probe based system to address this limitation. The probe is a lightweight application that runs on each Web service VM that needs to be monitored. We periodically measure the probe's response time on a monitored VM. We then compare this response time with the probe's previously recorded baseline no-interference response time when it executes in isolation on a VM of the same type. Statistically significant increase in the probe's response time from the baseline is used to detect interference. The probe also indicates the type of contention at the physical host that causes the interference. This information can be exploited by a subscriber to mitigate the problem. Results show that our approach is quite effective over two different cloud platforms and a wide variety of workload scenarios. In particular, results indicate that Web service instances hosted on EC2 suffer from interference. Our probe was able to detect 93% of performance degradations triggered by such interference. In all these cases, the probe imposed an average overhead of only 3%-4% on the mean response time of the Web service being monitored.

Cloud storage systems are emerging as the primary solution for online storage and information sharing. As the demand for such a service is increasing at a phenomenal rate, the cost for maintaining and delivering content concerns the cloud providers and ISPs. As in other distributed systems, e.g., file sharing and multimedia streaming, network coding can significantly simplify the process for content distribution and retrieval. However, it also raises difficulties in updating portions of a file, as any change in the file will impact all coded content in the system. In this paper, we present the differential update model and its optimization for updating coded blocks by delivering only the changes in a file. We complete the design with an update algorithm and a communication protocol among all participants in the system. Our experimental results verify that our design makes network coding practical for file updates in cloud storage systems. The proposed update model leads to bandwidth saving, compared to conventional update mechanisms, with minimal computational costs.

HTTP-based adaptive protocols dominate today's video streaming over the Internet, and operate using multiple quality levels that video players request one segment at a time. Despite their popularity, studies have shown that performance of video streams still suffers from stalls, quality switches and startup delay. In wireless networks, it is well-known that high variability in network bandwidth affects video streaming. MultiPath TCP (MPTCP) is an emerging paradigm that could offer significant benefits to video streaming by combining bandwidth on multiple network interfaces, in particular for mobile devices that typically support both WiFi and cellular networks. In this paper, we explore whether MPTCP always benefits mobile video streaming. Our experimental study on video streaming using two wireless interfaces yields mixed results. While beneficial to user experience under ample and stable bandwidth, MPTCP may not offer any advantage under some network conditions. We find that when additional bandwidth on the secondary path is not sufficient to sustain an upgrade in video quality, it is generally better not to use MPTCP. We also identify that MPTCP can harm user experience when an unstable secondary path is added to the stable primary path.

Cloud computing provides a reliable and cost- effective setting for deploying large-scale web applications. However, choosing and configuring an appropriate cloud Infrastructure-as-a-Service (IaaS), e.g., the appropriate database and computing instances and acceptable service rates, is a daunting task. The task is also challenging when trying to optimize the IaaS for conflicting objectives such as performance and cost. Furthermore, due to lack of understanding of the pricing model and the cloud IaaS, a cloud consumer may pay more than necessary or may not fully utilize the purchased resources. For this reason, we propose an algorithm that suggests the most cost-effective configuration meeting the QoS requirements and budget constraint. In contrast to existing cost optimization proposals, our proposed algorithm maps the minimum requirements of the to- be-deployed web application to deployment costs according to the price model set by cloud providers. The algorithm also considers QoS requirements for different resource types in the cloud, namely, database servers, computing servers, storage, and service rate. The proposed algorithm is evaluated by a series of experiments on a web application with seven different workload scenarios. The experimental results show the effectiveness of the proposed algorithm in achieving a solution with the minimum deployment cost for each scenario while satisfying all customer's requirements.

To improve the quality of video streaming subject to video bitrate or communication channel capacity, a high-quality video is encoded into multiple layers of unequal importance. Layers that provide higher quality rely on the previous layers for successful reconstruction of transmitted video packets. Hence, if a video packet in a reference layer is corrupted or lost during transmission, the dependent layers cannot be reconstructed successfully, and the resources consumed to transmit them are wasted. To address this problem, unequal error protection (UEP) techniques have been proposed to provide appropriate level of protection to each layer according to their importance. Nonetheless, the importance of a piece of video content is determined by not only the layering structure, but also coding dependency imposed by encoding decisions. In this paper, based on a deep inspection of coding and prediction in SVC (a layered video coding standard) and an analysis of seven real SVC videos, we conclude that macro block-level coding dependency will provide a more accurate importance measure when applying UEP to protection video packets in noisy channels.

In recent years, the demand for high quality video streaming services has been growing significantly. Distributed video transcoding in cloud, i.e., re-encoding the source video to best match the capabilities of the network connection and the playback device in cloud and sending each user a tailored version of the video, is a recent solution for fast and high quality video streaming services. In such a transcoding scheme, video is segmented into chunks of equal size and the chunks are distributed among multiple virtual machines for parallel transcoding. The transcoded chunks are then merged together to create the new transcoded video appropriate for playback on specific end-user devices. In this paper, we conduct a performance analysis of the impact of chunk size on coding efficiency and transcoding time. We observe that transcoding with larger chunks leads to better coding efficiency (i.e., lower bitrate) by trading off the transcoding time. The improvement in coding efficiency and the level of trade-off in transcoding time highly depend on the visual similarity among frames of a video sequence. From the analysis, we suggest that for better coding efficiency and faster transcoding, the chunk size should be dynamically adjusted according to the visual similarity.

To improve the quality of experience of video streaming services, content providers are challenged by the need to prepare videos at different quality levels appropriate to the network infrastructure and device hardware specification. Distributed video transcoding in the cloud has received many research attentions to address this challenge. Such a cloud-based solution segments a video into multiple video chunks and distributes chunks to virtual machines in the cloud for parallel transcoding. However, by inspecting video codec standards, we learn that important inter-dependency among video frames is broken if the video is segmented into fixed-size chunks, which leads to increasing bitrate and transcoding time. In this paper, we propose a distributed video transcoding scheme that exploits dependency among GOPs by preparing video chunks of variable size. Experimental results from real video sequences with diverse visual features show that the proposed transcoding scheme effectively reduces bitrate and transcoding time.

The number of network systems and network applications is rapidly increasing with the wide deployment of broadband network access. Network protocols and algorithms, as key components of network systems, are usually developed and evaluated in a simulated network, an emulated network and/or a real network. In this paper, we propose evalBox, a cross-platform evaluation framework that provides support to deploy network systems in various evaluation platforms. The framework supports network simulation, network emulation, and real deployment. Hence, a network system once implemented can then be directly deployed in any platform supported by evalBox for evaluation. The design of evalBox effective mitigates the effort involved in the learning/training phase of these platforms and eliminates the redundant tasks involved in porting a network system across different platforms for evaluation.

Layered video coding standards encode a high-quality video into multiple layers of unequal importance. Dependent layers that provide higher quality rely on their respective reference layers for successful reconstruction of transmitted video frames. Hence, if a video packet in a reference layer is corrupted or lost during transmission, all its dependent layers cannot be reconstructed successfully, and the resources consumed to transmit them are wasted. To address this problem, unequal error protection (UEP) techniques have been proposed to provide protection to each layer according to their importance. Nonetheless, the importance of a piece of video content is determined by not only the layering structure, but also visual features and encoding decisions. In this paper, we look deeper into the coding and prediction structure of layered encoded videos and model the the dependency among macroblocks and submacroblocks (the finest processing units of H.264 video coding standard) as a weighted graph. Based on this graph, we propose a dependency-aware UEP model that protects macroblocks according to their importance. Our simulation results show that the proposed UEP model outperforms the conventional UEP models for layered SVC videos by 3.76 dB of peak signal-to-noise ratio (PSNR) when the channel packet loss rate is as high as 28%.

With increasing deployment of 3G/4G technologies and growing computing power of modern smartphones, video multicast over mobile networks is becoming very popular. Due to mobility and interference, the channel quality fluctuates frequently over time, making it challenging to deliver live streaming services to mobile devices. In this paper, we propose an adaptive video streaming system that delivers layered video content to mobile devices over wireless channels of different and fluctuating qualities. To this end, the proposed system employs a novel design to prepare the redundant coded blocks required for forward error correction. This design makes the system flexible to dynamic changes in loss rates with minimum coding overhead. Specifically, the employed coding scheme is empowered by a novel fine granular coefficient matrix that decreases the delay and the computational complexity of coding operations. Furthermore the simulation results show that the new system offers a flexible streaming service, decreases the computational cost of preparing the FEC codes, significantly decreases the video transmission delay, and finally conserves energy on mobile devices.

While the demand for multimedia streaming, especially from mobile devices, is growing rapidly, mobile devices are challenged by the variable transmission rate in wireless channels and limited battery capacity. In this paper, we propose an energy-efficient cooperative streaming system. In the proposed system, mobile devices collectively stream a copy of the multimedia content from the source over cellular links. The devices form a cooperative group and share received content, in coded form for efficient loss detection and recovery, with each other over short-range links. The sharing is guided by the optimal rate allocation and scheduling (RAS) algorithm that determines the amount of data and the data to be transmitted on each link. Overall, the system minimizes both the streaming traffic in the cellular network and the energy consumed by streaming applications on mobile devices. Our experimental results also show that the RAS algorithm prolongs the streaming session for the entire cooperative group.

The continuous developments and improvements of network infrastructure along with the growing number of modern smartphones, tablets and smart TVs have led to an increasing popularity of multimedia applications, such as video conferencing, video streaming and mobile TV. Towards delivering video stream to diverse devices over a heterogamous network, scalable video coding (SVC) has received many research attentions. SVC is an extension of H.264/AVC that allows a video streaming service provider to encode a high quality video into a number of scalable layers. The receivers of the stream may decode the video at the appropriate quality level that is suitable for their hardware/software capabilities and network connections. Nevertheless, compared to single layer H.264/AVC, the de facto standard for many commercial streaming service providers such as YouTube, SVC is not widely deployed, mostly due to its overhead in terms of bit rate and complexity. In this paper, we conduct a thorough study on the performance of SVC for full HD video streaming. Our performance analysis identifies good and bad uses of SVC, quantifies the coding overhead, and benchmarks the SVC video quality under different spatial, temporal, and quality settings. Through the use a set of carefully selected and diverse video sequences, we also identify the types of video that can benefit from SVC.

Due to the increasing deployment of 3G/4G technologies and the growing computing power of modern smart-phones, video streaming is one of the most popular applications on these devices. In this paper, we propose an energy-efficient multimedia streaming system for smartphones. This system simultaneously utilizes two communication channels: cellular links carrying streaming content from the media Cloud to smartphones and WiFi links enabling cooperation among smartphones. There are three key contributions. First, the system copes with the variable loss rate and bandwidth fluctuation, common problems in wireless communication, by taking advantage of a novel two-level coding scheme. Second, the system reduces energy consumption due to network coding operations, by utilizing the copious computing resources in the Cloud. Consequently, XOR-only network coding is sufficient for sharing among smartphones. Third, we propose a light-weight distributed scheduling algorithm to manage collaboration and content sharing among the neighbouring nodes. The proposed local dissemination scheme, along with the proposed two-level coding scheme, outperforms state-of-the-art streaming systems in cellular networks. Our experimental results show that our proposed system saves up to 73% of battery usage on each phone compared to the current streaming mechanism over a 3G network. In addition, the system is capable to support higher quality streaming and decrease the streaming delay.

In recent years, cloud storage systems have emerged as the primary solution for online storage and information sharing. Due to efficient storage and bandwidth utilization, the use of erasure codes and network coding is proven to effectively provide fault tolerance and fast content retrieval in cloud storage systems. In a nutshell, coded blocks are distributed among storage nodes, and file retrieval is accomplished by downloading sufficient coded blocks from any group of storage nodes. However, due to high correlation between coded blocks and the original file, even a single-byte update invalidates all coded blocks in the system. In this paper, we introduce DeltaNC, a new differential update algorithm that keeps all coded blocks in a network-coding-based cloud storage system synchronized by transmitting only the changes in the file. Our experimental results, from a trace-driven simulator, show that DeltaNC significantly reduces the bandwidth and CPU usage and its performance is comparable to that offered by the Diff program, the common tool for updating files.

While every network node only relays messages in a traditional communication system, the recent network coding (NC) paradigm proposes to implement simple in-network processing with packet combinations in the nodes. NC extends the concept of “encoding” a message beyond source coding (for compression) and channel coding (for protection against errors and losses). It has been shown to increase network throughput compared to traditional networks implementation, to reduce delay and to provide robustness to transmission errors and network dynamics. These features are so appealing for multimedia applications that they have spurred a large research effort towards the development of multimedia-specific NC techniques. This paper reviews the recent work in NC for multimedia applications and focuses on the techniques that fill the gap between NC theory and practical applications. It outlines the benefits of NC and presents the open challenges in this area. The paper initially focuses on multimedia-specific aspects of network coding, in particular delay, in-network error control, and media-specific error control. These aspects permit to handle varying network conditions as well as client heterogeneity, which are critical to the design and deployment of multimedia systems. After introducing these general concepts, the paper reviews in detail two applications that lend themselves naturally to NC via the cooperation and broadcast models, namely peer-to-peer multimedia streaming and wireless networking.

In recent years, the demand for multimedia streaming is soaring over the Internet. Due to the lack of a centralized administrative point, Peer-to-Peer (P2P) streaming system is vulnerable to pollution attacks, in which video segments might be altered by any peer before being shared. Among existing proposals, reputation-based defence mechanisms are the most effective and practical solutions. In this paper, we perform a measurement study on the effectiveness of this class of solutions. We implement a framework that allows us to simulate different variations of the reputation rating systems, from the global approach to the decentralized local approach, under different parameter settings and pollution models. In order to ensure the framework and the simulated solution is representative enough, we dissect existing proposals and implement a flexible defence mechanism, in which different components may be enabled and disabled by simply tuning certain parameters. Our results reveal that global knowledge of the content flow in the network does not necessarily improve the performance. It is often susceptible under collaborative attacks. We also find that expelling misbehaving peers is often more useful to prevent attacks than limiting their likelihood to be connected, although this can lead to poor playback quality.

In recent years, Peer-to-Peer (P2P) multimedia streaming has become an alternative to cable/satellite TV services. Many P2P streaming applications further provide users with DVD-like operations: play, pause, chapter selection, fast-forward, and rewind. Such a real-time interactive multimedia streaming is commonly referred to as the P2P Video-on-Demand (VoD). To further improve the streaming quality, recent research employs network coding as the key enabling technology. However, the practicality and implementation challenges of network coding received very little attention. In this paper, we present a practical implementation of network coding in a P2P VoD system, VoD+NC, based on which we conduct a measurement study on the actual performance gain provided by network coding. In the meantime, we identify design pitfalls when incorporating network coding into a P2P VoD system. Our study shows that, unlike P2P live streaming, directly applying network coding to a P2P VoD system does not necessarily lead to an immediate improvement in playback quality. With the proper configuration, network coding not only brings the same benefits as it does in P2P live streaming, but also better accommodates the asymmetric interests among peers and simplifies the neighbourhood management.

The demand for cloud services is growing at a phenomenal rate, and so is the energy cost of the data centres powering those services. This is pressing cloud service providers to look for ways of reducing energy consumption. One approach is to utilize energy-efficient hardware and/or software in the data centres, the other approach is to relocate some services, e.g., personal files and data rendering, to end-host computers, a.k.a. peers. In the later approach, peers contribute their communication and computation resources to exchange data and provide services, while the data centre performs central administration and authentication, as well as backend processing. In this paper, we model the energy consumption for both approaches and then perform analytical studies. Our analysis shows that (1) making the data centre energy efficient can reduce the energy cost significantly; (2) the number of hops from the data centre to the peers and among peers directly influences the energy saving; (3) it is preferred to utilize peers that are already online for other purposes; (4) introducing content delivery network (CDN) servers and enabling proxy service on home modems are the keys to make a hybrid P2P-cloud network go green. We further verified these findings by simulation.

Peer-to-Peer (P2P) live streaming traffic has been growing at a phenomenal rate over the past few years. When the original streaming content is mixed with bogus data, the corresponding P2P streaming network is being subjected to a “pollution attack.” As the content is shared by peers, the bogus data can be spread widely in minutes. In this paper, we study the impact of a pollution attack in popular streaming models, under various network settings and configurations. The study was conducted in SPoIM, our emulation of real-world P2P streaming systems under pollution attacks, through which we observed that the feasibility of the attack is sensitive to the speed at which an attacker can modify content. Our experimental results showed that different streaming approaches are more vulnerable in one network configuration than the others, and that the impact and effectiveness of the attack is not dependent on the network size, but does highly depend on the network stability and the bandwidth availability of the polluters and the source. Based the experimental results, we suggested possible improvements in streaming models to defend themselves against the pollution attack. Finally, we examined possible defense mechanisms and demonstrated the effectiveness of a reputation-based defense mechanism against a typical pollution attack.

Homomorphic hash functions (HHFs) have been applied into peer-to-peer networks with erasure coding or network coding to defend against pollution attacks. Unfortunately HHFs are computationally expensive for contemporary CPUs, This paper to exploit the computing power of graphic processing units (GPUs) for homomorphic hashing. Specifically, we demonstrate how to use NVIDIA GPUs and the computer unified device architecture (CUDA) programming model to achieve 38 times of speedup over the CPU counterpart. We also develop a multi-precision modular arithmetic library on CUDA platform, which is not only key to our specific application, but also very useful for a large number of cryptographic applications.

To rapidly evolve new designs of peer-to-peer (P2P) multimedia streaming systems, it is highly desirable to test and troubleshoot them in a controlled and repeatable experimental environment in a local cluster of servers, as it is risky to integrate untested protocols in live production and mission-critical peer-to-peer sessions, such as live P2P streaming. Though it is possible to construct such controlled experiments with virtual machine monitors, there are a number of challenges and roadblocks: (1) The deployment of such resource-hungry virtual machine environments are complicated and time-consuming for researchers without prior systems expertise; (2) The system designer needs to implement many basic streaming elements, such as playback buffers and message switches. In this paper, we seek to address these challenges by introducing Crystal, an emulation framework for practical P2P multimedia streaming systems, which provides support for developing, testing, and troubleshooting new streaming system designs in a controlled server cluster environment. It is our imperative design objective that Crystal offers ease of use, rapid experimental turnaround, and the capability of emulating realistic P2P environments.

Network coding has recently been widely applied in various networks for system throughput improvement and/or resilience to network dynamics. However, the computational overhead introduced by the network coding operations is not negligible and has become the cornerstone for real deployment of network coding. In this paper, we exploit the computing power of contemporary Graphic Processing Units (GPUs) to accelerate the network coding operations. We proposed three parallel algorithms that maximize the parallelism of the encoding and decoding processes, i.e., the power of GPUs is fully utilized. This paper also shares our optimization design choices and our workarounds to the challenges encountered in working with GPUs. With our implementation of the algorithms, we are able to achieve up to 12 times of speedup over the highly optimized CPU counterpart, using the NVIDIA GPU and the Computer Unified Device Architecture (CUDA) programming model.

Wireless spectrum is a scare resource, while media streaming usually requires high end-to-end bandwidth. Media streaming in wireless ad hoc networks is therefore a particularly challenging problem, especially for the case of streaming to multiple receivers. In this paper, we design linear optimization models for computing a high-bandwidth routing strategy for media multicast in wireless networks, which targets near-optimal throughput, given constraints including network topology, radio capacity, and link contention. We study both the directional antenna and omni-directional antenna cases and point out their connections. We also combine the classic forward error correction techniques with the novel network coding techniques to provide error control in a timely fashion. Simulation results show that our solutions indeed achieve high streaming rates, and prompt error recovery under a wide range of link failure patterns.

In recent literature, network coding has emerged as a promising information theoretic approach to improve the performance of both peer-to-peer and wireless networks. It has been widely accepted and acknowledged that network coding can theoretically improve network throughput of multicast sessions in directed acyclic graphs, achieving their cut-set capacity bounds. Recent studies have also supported the claim that network coding is beneficial for large-scale peer-to-peer content distribution, as it solves the problem of locating the last missing blocks to complete the download. We seek to perform a reality check of using network coding for peer-to-peer live multimedia streaming. We start with the following critical question: How helpful is network coding in peer-to-peer streaming? To address this question, we first implement the decoding process using Gauss-Jordan elimination, such that it can be performed while coded blocks are progressively received. We then implement a realistic testbed, called Lava, with actual network traffic to meticulously evaluate the benefits and tradeoffs involved in using network coding in peer-to-peer streaming. We present the architectural design challenges in implementing network coding for the purpose of streaming, along with a pull-based peer-to-peer live streaming protocol in our comparison studies. Our experimental results show that network coding makes it possible to perform streaming with a finer granularity, which reduces the redundancy of bandwidth usage, improves resilience to network dynamics, and is most instrumental when the bandwidth supply barely meets the streaming demand.

In information theory, it has been shown that network coding can effectively improve the throughput of multicast communication sessions in directed acyclic graphs. More practically, random network coding is also instrumental towards improving the downloading performance in BitTorrent-like peer-to-peer content distribution sessions. Live peer-to-peer streaming, however, poses unique challenges to the use of network coding, due to its strict timing and bandwidth constraints. In this paper, we revisit the complete spectrum in the design space of live peer-to-peer streaming protocols, with a sole objective of taking full advantage of random network coding. We present R2, our new streaming algorithm designed from scratch to incorporate random network coding with a randomized push algorithm. R2is designed to improve the performance of live streaming in terms of initial buffering delays, resilience to peer dynamics, as well as reduced bandwidth costs on dedicated streaming servers, all of which are beyond the basic requirement of stable streaming playback. On an experimental testbed consisting of dozens of dual-CPU cluster servers, we thoroughly evaluate R2with an actual implementation, real network traffic, and emulated peer upload capacities, in comparisons with a typical live streaming protocol (both without and with network coding), representing the current state-of-the-art in real-world streaming applications.

In recent literature, network coding has emerged as a promising information theoretic approach to improve the performance of both peer-to-peer (P2P) and wireless networks. It has been widely accepted and acknowledged that network coding can theoretically improve network throughput of multicast sessions in directed acyclic graphs, achieving their cut-set capacity bounds. Recent studies have also supported the claim that network coding is beneficial for large-scale P2P content distribution, as it solves the problem of locating the last missing blocks to complete the download. We seek to perform a reality check of using network coding for P2P live multimedia streaming. We start with the following critical question: How helpful is network coding in P2P streaming? To address this question, we first implement the decoding process using Gauss-Jordan elimination, such that it can be performed while coded blocks are progressively received. We then implement a realistic testbed, called Lava, with actual network traffic to meticulously evaluate the benefits and tradeoffs involved in using network coding in P2P streaming. We present the architectural design challenges in implementing network coding for the purpose of streaming, along with a pull-based P2P live streaming protocol in our comparison studies. Our experimental results show that network coding makes it possible to perform streaming with a finer granularity, which reduces the redundancy of bandwidth usage, improves resilience to network dynamics, and is most instrumental when the bandwidth supply barely meets the streaming demand.

With network coding, intermediate nodes between the source and the receivers of an end-to-end communication session are not only capable of relaying and replicating data messages, but also of coding incoming messages to produce coded outgoing ones. Recent studies have shown that network coding is beneficial for peer-to-peer content distribution, since if eliminates the need for content reconciliation, and is highly resilient to peer failures. In this paper, we present our recent experiences with a highly optimized and high-performance C++ implementation of randomized network coding at the application layer. We present our observations based on an extensive series of experiments, draw conclusions from a wide range of scenarios, and are more cautious and less optimistic as compared to previous studies

Existing research work towards the composition of complex federated services has assumed that service requests and deliveries flow through a particular service path or tree. Here, we extend such a service model to a directed acyclic graph, allowing services to be delivered via parallel paths and interleaved with each other. Such an assumption of the service flow model has apparently introduced complexities towards the development of a distributed algorithm to federate existing services, as well as the provisioning of the required quality in the most resource-efficient fashion. To this end, we propose sFlow, a fully distributed algorithm to be executed on all service nodes, such that the federated service flow graph is resource efficient, performs well, and meets the demands of service consumers.