Assessing media streaming performance traditionally requires the presence of reproducible network conditions and a heterogeneous dataset of media materials. Setting up such experiments represents a complex challenge in itself. This challenge becomes even more complex when we consider the new QUIC transport protocol, which has many tunable features, yet is difficult to analyze due to its inherent encrypted nature. In this paper, we introduce Vegvisir, which aims to solve these aforementioned challenges by providing an open-source automated testing framework for orchestrating media streaming experiments over HTTP/3. We describe how users can steer the behavior of Vegvisir through its configuration system. We provide a high-level overview of its inner workings and its broad applicability by describing two use cases: one covering sizeable experiments spanning multiple days and another covering HAS evaluation scenarios.

With HTTP Adaptive Streaming (HAS), client-side Adaptive Bitrate (ABR) algorithms drive the (quality-variant) scheduling and downloading of media segments. These ABR algorithms are implemented in the application layer and can therefore base their logic only on relatively coarse and/or inaccurate application-layer metrics. The recently standardized QUIC transport protocol has many userspace implementations, which paves the way for cross-layer optimizations by exposing transport-layer metrics to application-layer algorithms. In this paper, we investigate whether the availability of fine-grained transport-level throughput metrics can positively impact the operation of ABR algorithms and hence the Quality of Experience (QoE) of HAS users in Video on Demand (VoD) settings. Our results show that QUIC-level throughput data can indeed aid ABR algorithms to more accurately predict playout buffer under-runs, which in turn allows the ABR logic to take reactive measures in a timely fashion such that playback stalls can be avoided under challenging network conditions. Overall, our work presents a step towards improving ABR operation via cross-layer data exchange.

Abstract Several data protection regulations permit individuals to request all personal information that an organization holds about them by utilizing Subject Access Requests (SARs). Prior work has observed the identification process of such requests, demonstrating weak policies that are vulnerable to potential data breaches. In this paper, we analyze and compare prior work in terms of methodologies, requested identification credentials and threat models in the context of privacy and cybersecurity. Furthermore, we have devised a longitudinal study in which we examine the impact of responsible disclosures by re-evaluating the SAR authentication processes of 40 organizations after they had two years to improve their policies. Here, we demonstrate that 53% of the previously vulnerable organizations have not corrected their policy and an additional 27% of previously non-vulnerable organizations have potentially weakened their policies instead of improving them, thus leaking sensitive personal information to potential adversaries. To better understand state-of-the-art SAR policies, we interviewed several Data Protection Officers and explored the reasoning behind their processes from a viewpoint in the industry and gained insights about potential criminal abuse of weak SAR policies. Finally, we propose several technical modifications to SAR policies that reduce privacy and security risks of data controllers.

This paper demystifies listeners’ wishes with respect to broadcast radio innovation (with a specific focus on radio-mediated music consumption). Our study encompasses an ideation workshop with radio experts, an exploratory survey and a mixed methods empirical evaluation. The empirical evaluation uses two concrete concepts (i.e., letting listeners on-the-fly replace radio content with preferred content and fostering participatory radio production by involving listeners as radio content curators) as a lens to zoom in on the questionable desirability of radio innovation. It is learned that a significant consumer group exists who will stay loyal to broadcast radio even if it does not evolve substantially, whereas others need disruptive incentives to start listening to radio (again). From our results we distill design recommendations to educate the radio production community about how best to approach radio innovation.

QUIC is marketed to hold many advantages over TCP. However, preliminary experimentation has shown that simply running contemporary HTTP Adaptive Streaming (HAS) implementations over QUIC does not improve but actually hurts streaming performance compared to a traditional TCP deployment. We argue that this behavior can be attributed to the amount of TCP specialization that HAS Adaptive BitRate (ABR) algorithms have received over the years. In contrast to TCP (which can be regarded as a "black box"), QUIC actually encompasses all the necessary tools to empower streaming performance optimization (e.g., definition of custom congestion control algorithms, access to transport-layer metrics in the application layer). This however comes at the expense of added complexity which in turn could lead to misinterpretations of the root causes of suboptimal streaming performance. To facilitate research on HTTP adaptive bitrate streaming over QUIC, in this paper we propose a solution towards jointly visualizing transport- and application-layer metrics to allow for a better understanding of HAS streaming performance over various types of transport protocols (i.e., TCP versus QUIC). We see the work presented in this paper as an important stepping stone towards cross-layer optimization of HAS ABR performance to achieve a better overall Quality of Experience (QoE) for streaming users.

Determining which operations are being executed by a black-box device is an important challenge to tackle in reverse engineering. Furthermore, in order to perform a successful side-channel analysis (SCA) of said operations, their precise timing must be determined. In this paper, we tackle these two challenges in context of an electromagnetic (EM) analysis of a NodeMCU Amica IoT device. More specifically, we propose a convolutional neural network (CNN) architecture that is designed to classify operations performed by the NodeMCU out of a set of 8 possible operations, namely OpenSSL AES, native AES, TinyAES, OpenSSL DES, SHA1-PRF, HMAC-SHA1, SHA1, and SHA1Transform. In addition, we use the same architecture to predict the start and end times of the operation, thereby removing the need for firmware modifications or manual triggers in SCA. Our approach is evaluated using a 66 GB dataset containing 69,632 complex traces of EM leakage, captured with a USRP B210 software defined radio. The best variant of our methodology achieves a classification accuracy of 96.47%, and is able to predict the start and end times of the operation within 34 |is of the ground truth on average. We compare our methodology to classical template matching, and provide our open-source implementation and datasets to the community so that the achieved results can be reproduced.

Zero-rating is a technique where internet service providers (ISPs) allow consumers to utilize a specific website without charging their internet data plan. Implementing zero-rating requires an accurate website identification method that is also efficient and reliable to be applied on live network traffic. In this paper, we examine existing website identification methods with the objective of applying zero-rating. Furthermore, we demonstrate the ineffectiveness of these methods against modern encryption protocols such as Encrypted SNI and DNS over HTTPS and therefore show that ISPs are not able to maintain the current zero-rating approaches in the forthcoming future. To address this concern, we present “Open-Knock,” a novel approach that is capable of accurately identifying a zero-rated website, thwarts free-riding attacks, and is sustainable on the increasingly encrypted web. In addition, our approach does not require plaintext protocols or preprocessed fingerprints upfront. Finally, our experimental analysis unveils that we are able to convert each IP address to the correct domain name for each website in the Tranco top 6000 websites list with an accuracy of 50.5% and therefore outperform the current state-of-the-art approaches.

In webpage fingerprinting (WPF), an adversary attempts to identify webpages in encrypted network traffic. Identifying social media webpages however is a challenging task, due to the similarity and dynamic nature of such pages. Existing webpage fingerprinting attacks often have unrealistic assumptions regarding the capability of government agencies or knowledge of the criminal's environment, which renders these attacks ineffective when applied to social media platforms. In this paper, we unravel the current concerns in state of the art WPF attacks in a social network context for forensic analysis. To resolve the issues presented, we propose an enhanced version of the WPF attack 'IUPTIS' and introduce an intelligent observer that significantly improves upon previous works. Furthermore, our improvements are compared to related WPF attacks by conducting extensive experiments on two social platforms: Twitter and Instagram. Our examination shows that the improved IUPTIS attack defeats previous works in terms of realistic obstacles such as HTTP/2, caching and performance costs, thus making it feasible to identify social media webpages with minimal resources.

Over-the-top (OTT) streaming services like YouTube and Netflix induce massive amounts of video data, hereby putting substantial pressure on network infrastructure. This paper describes a demonstration of the object-based video (OBV) methodology that allows for the quality-variant MPEG-DASH streaming of respectively the background and foreground object(s) of a video scene. The OBV methodology is inspired by research into human visual attention and foveated compression, in that it allows to adaptively and dynamically assign bitrate to those portions of the visual scene that have the highest utility in terms of perceptual quality. Using a content corpus of interview-like video footage, the described demonstration proves the OBV methodology's potential to downsize video bitrate requirements while incurring at most marginal perceptual impact (i.e., in terms of subjective video quality). Thanks to its standards-compliant Web implementation, the OBV methodology is directly and broadly deployable without requiring capital expenditure.

Over-the-top (OTT) streaming services like YouTube and Netflix induce massive amounts of video traffic. To combat the resulting network load, this article empirically explores the use of the object-based video (OBV) methodology that allows for the quality-variant HTTP Adaptive Streaming of respectively the background and foreground object(s) of a video scene. In particular, we study two alternative video object representation methods where the first meticulously follows the object contour, while the second uses axis-aligned bounding box enclosures. We subjectively compare both techniques to traditional, frame-based video compression in the context of live action content featuring talking persons. The resulting mixed methods data shows that (i) OBV-informed users tolerate substantial background quality degradations, and (ii) at an average bitrate reduction of 14 percent, perceptual differences between respectively contour-based OBV and traditional encoding are small or even non-existing for the non-movie content in our corpus. Although our evaluation focuses on interview-like footage, our qualitative data hints that the presented results might be extrapolatable to other video genres. As such, our findings inform content owners and network operators about video bitrate saving opportunities with marginal perceptual impact.

Not even five years after the standardization of HTTP/2, work is already well underway on HTTP/3. This latest version is necessary to make optimal use of that other new and exiting protocol: QUIC. However, some of QUIC’s unique characteristics make it challenging to keep HTTP/3’s functionalities on par with those of HTTP/2. Especially the efforts on adapting the prioritization system, which governs how multiple resources can be multiplexed on a single QUIC connection, have led to some difficult to answer questions. This paper aims to help answer some of those questions by being the first to provide experimental evaluations and result comparisons for 11 different possible HTTP/3 prioritization approaches in a variety of simulation settings. We present some non-trivial insights, discuss advantages and disadvantages of various approaches, and provide results-backed actionable advice to the standardization working group. We also help foster further experimentation by contributing our complete HTTP/3 implementation, results dataset and custom visualizations to the community.

LoRa is a recently introduced modulation scheme specifically designed for Low-Power Wide-Area Networks. \nIn this paper, we provide the first detailed and complete description of the LoRa PHY layer, and present a \nnovel methodology for detecting and decoding LoRa frames using Software Defined Radios. Our proposed \ndecoding approach can efficiently decode multiple channels simultaneously in software due to an invariance \ntowards the signal frequency. Hence, our approach also removes the need for correcting frequency offset errors \nimparted by the transmitter or receiver. We have evaluated our decoding technique in a lab setup using three \nSoftware Defined Radios (USRP B210, HackRF, and RTL-SDR) and three commercial off-the-shelf hardware \nLoRa transceivers (Microchip RN2483, HopeRF RFM96, and Semtech SX1272). We show that our decoder \nis fully compatible with all configurations of the RN2483 and SX1272, achieving an overall packet error rate \nof 0 for a signal-to-noise ratio of 20 dB. The source code of the decoder and datasets used in the evaluation \nare made available publicly

Static light fields are an effective technology to precisely visualize complex inanimate objects or scenes, synthetic and real-world alike, in Augmented, Mixed and Virtual Reality contexts. Such light fields are commonly sampled as a collection of 2D images. This sampling methodology inevitably gives rise to large data volumes, which in turn hampers real-time light field streaming over best effort networks, particularly the Internet. This paper advocates the packaging of the source images of a static light field as a segmented video sequence so that the light field can then be interactively network streamed in a quality-variant fashion using MPEG-DASH, the standardized HTTP Adaptive Streaming scheme adopted by leading video streaming services like YouTube and Netflix. We explain how we appropriate MPEG-DASH for the purpose of adaptive static light field streaming and present experimental results that prove the feasibility of our approach, not only from a networking but also a rendering perspective. In particular, real-time rendering performance is achieved by leveraging video decoding hardware included in contemporary consumer-grade GPUs. Important trade-offs are investigated and reported on that impact performance, both network-wise (e.g., applied sequencing order and segmentation scheme for the source images of the static light field) and rendering-wise (e.g., disk-versus-GPU caching of source images). By adopting a standardized transmission scheme and by exclusively relying on commodity graphics hardware, the net result of our work is an interoperable and broadly deployable network streaming solution for static light fields.

Webpage fingerprinting allows an adversary to infer the webpages visited by an end user over an encrypted channel by means of network traffic analysis. If such techniques are applied to websites that contain user profiles (e.g. booking platforms), they can be used for personal identification and pose a clear privacy threat. In this paper, a novel HTTPS webpage fingerprinting method - IUPTIS - is presented, which accomplishes precisely this, through identification and analysis of unique image sequences. It improves upon previous work by being able to fingerprint webpages containing dynamic rather than just static content, making it applicable to e.g. social network pages as well. At the same time, it is not hindered by the presence of caching and does not require knowledge of the specific browser being used. Several accuracy-increasing parameters are integrated that can be tuned according to the specifics of the adversary model and targeted online platform. To quantify the real-world applicability of the IUPTIS method, experiments have been conducted on two popular online platforms. Favorable results were achieved, with a F1 scores between of 82% and 98%, depending on the parameters used. This makes the method practically viable as a means for personal identification.

Web performance is a hot topic, as many studies have shown a strong correlation between slow webpages and loss of revenue due to user dissatisfaction. Front and center in Page Load Time (PLT) optimization is the order in which resources are downloaded and processed. The new HTTP/2 specification includes dedicated resource prioritization provisions, to be used in tandem with resource multiplexing over a single, well-filled TCP connection. However, little is yet known about its application by browsers and its impact on page load performance. This article details an extensive survey of modern User Agent implementations, with the conclusion that the major vendors all approach HTTP/2 prioritization in widely different ways, from naive (Safari, IE, Edge) to complex (Chrome, Firefox). We investigate the performance effect of these discrepancies with a full-factorial experimental evaluation involving eight prioritization algorithms, two off-the-shelf User Agents, 40 realistic webpages, and five heterogeneous (emulated) network conditions. We find that in general the complex approaches yield the best results, while naive schemes can lead to over 25% slower median visual load times. Also, prioritization is found to matter most for heavy-weight pages. Finally, it is ascertained that achieving PLT optimizations via generic server-side HTTP/2 re-prioritization schemes is a non-trivial task and that their performance is influenced by the implementation intricacies of individual browsers.

Web page performance is becoming increasingly important for end users but also more difficult to provide by web developers, in part because of the limitations of the legacy HTTP/1 protocol. The new HTTP/2 protocol was designed with performance in mind, but existing work comparing its improvements to HTTP/1 often shows contradictory results. It is unclear for developers how to profit from HTTP/2 and whether current HTTP/1 best practices such as resource concatenation, resource embedding, and hostname sharding should still be used. In this work we introduce the Speeder framework, which uses established tools and software to easily and reproducibly test various setup permutations. We compare and discuss results over many parameters (e.g., network conditions, browsers, metrics), both from synthetic and realistic test cases. We find that in most non-extreme cases HTTP/2 is on a par with HTTP/1 and that most HTTP/1 best practices are applicable to HTTP/2. We show that situations in which HTTP/2 currently underperforms are mainly caused by inefficiencies in implementations, not due to shortcomings in the protocol itself.

Physical-layer fingerprinting investigates how features extracted from radio signals can be used to uniquely identify devices. This paper proposes and analyses a novel methodology to fingerprint LoRa devices, which is inspired by recent advances in supervised machine learning and zero-shot image classification. Contrary to previous works, our methodology does not rely on localized and low-dimensional features, such as those extracted from the signal transient or preamble, but uses the entire signal. We have performed our experiments using 22 LoRa devices with 3 different chipsets. Our results show that identical chipsets can be distinguished with 59% to 99% accuracy per symbol, whereas chipsets from different vendors can be fingerprinted with 99% to 100% accuracy per symbol. The fingerprinting can be performed using only inexpensive commercial off-the-shelf software defined radios, and a low sample rate of 1 Msps. Finally, we release all datasets and code pertaining to these experiments to the public domain.

In recent works, numerous physical-layer security systems have been proposed as alternatives to classic cryptography. Such systems aim to use the intrinsic properties of radio signals and the wireless medium to provide confidentiality and authentication to wireless devices. However, fundamental vulnerabilities are often discovered in these systems shortly after their inception. We therefore challenge the assumptions made by existing physical-layer security systems, and postulate that weaker assumptions are needed in order to adapt for practical scenarios. We also argue that if no computational advantage over an adversary can be ensured, secure communication cannot be realistically achieved.

We present two novel noncooperative MAC layer fingerprinting and tracking techniques for Wi-Fi (802.11) enabled mobile devices. Our first technique demonstrates how a per-bit entropy analysis of a single captured frame allows an adversary to construct a fingerprint of the transmitter that is 80.0 to 67.6 percent unique for 50 to 100 observed devices and 33.0 to 15.1 percent unique for 1,000 to 10,000 observed devices. We show how existing mitigation strategies such as MAC address randomization can be circumvented using only this fingerprint and temporal information. Our second technique leverages peer-to-peer 802.11u Generic Advertisement Service (GAS) requests and 802.11e Block Acknowledgement (BA) requests to instigate transmissions on demand from devices that support these protocols. We validate these techniques using two datasets, one of which was recorded at a music festival containing 28,048 unique devices and the other at our research lab containing 138 unique devices. Finally, we discuss a number of countermeasures that can be put in place by mobile device vendors in order to prevent noncooperative tracking through the discussed techniques.

The bulk of contemporary online video traffic is encoded in a traditional manner, hereby neglecting most, if not all, of the semantics of the underlying visual scene. One essential piece of semantic information in the context of video streaming is awareness of the objects that jointly constitute the scene. A canonical example of a benefit associated with such object awareness is the ability to subdivide a video fragment in respectively a background and one or more foreground objects. This paper reports on a pragmatically designed video streaming approach that exploits object-related knowledge in order to improve the real-time adaptability of video streaming sessions (manifested in the form of increased granularity in terms of streaming quality control). The proposed approach is completely compliant with present-day video codecs and HTTP Adaptive Streaming schemes, most notably H.264 and MPEG-DASH. Findings from subjecting the proposed video streaming technique to a comparative subjective evaluation suggest that scenarios exist where the presented approach holds the capacity to improve on traditional streaming in terms of user-perceived video quality.

Wicability is an open platform created for researchers, that aims to provide insights into the spatial and temporal impact of both novel and past 802.11 security vulnerabilities. This is achieved through the automated collection and analysis of large datasets containing 802.11 Information Elements (IEs) transmitted by access points and stations. The results of this analysis are anonymized and provided free of charge to researchers through a web interface.

Video consumption has since the emergence of the medium largely been a passive affair. This paper proposes augmented Omni-Directional Video (ODV) as a novel format to engage viewers and to open up new ways of interacting with video content. Augmented ODV blends two important contemporary technologies: Augmented Video Viewing and 360 degree video. The former allows for the addition of interactive features to Web-based video playback, while the latter unlocks spatial video navigation opportunities. The augmented ODV principle is exposed to developers through a collection of standards-compliant Web interfaces, this way allowing for a wide range of usage scenarios and heterogeneous consumer profiles to be targeted in a broad spectrum of application domains. The paper covers the extensive set of technologies powering the concept, hereby addressing a variety of topics ranging from content authoring and editing to visualization and rendering. Two use case prototypes showcase the expressiveness and practical feasibility of the augmented ODV methodology, while performance evaluation results establish that the augmented ODV implementation is computationally sufficiently lightweight to warrant execution on low-cost PC hardware and even tablet devices.

The ability to inject packets into a network is known to be an important tool for attackers: it allows them to exploit or probe for potential vulnerabilities residing on the connected hosts. In this paper, we present a novel practical methodology for injecting arbitrary frames into wireless networks, by using the Packet-In-Packet (PIP) technique to exploit the frame aggregation mechanism introduced in the 802.11n standard. We show how an attacker can apply this methodology over a WAN -- without physical proximity to the wireless network and without requiring a wireless interface card. The practical feasibility of our injection method is then demonstrated through a number of proof-of-concept attacks. More specifically, in these proof-of-concepts we illustrate how a host scan can be performed on the network, and how beacon frames can be injected from a remote location. We then both analytically and experimentally estimate the success rate of these attacks in a realistic test setup. Finally, we present several defensive measures that network administrators can put in place in order to prevent exploitation of our frame injection methodology.

Traditional video productions fail to cater to the interactivity standards that the current generation of digitally native customers have become accustomed to. This paper therefore advertises the "activation" of the video consumption process. In particular, it proposes to enhance HTML5 video playback with interactive features in order to transform video viewing into a dynamic pastime. The objective is to enable the authoring of more captivating and rewarding video experiences for end-users. The proposed paradigm extends video consumption, much like Augmented Reality (AR) does with the physical world. Given this conceptual analogy, we have adopted the term Augmented Video Viewing (AVV) to denominate our approach. The current AVV implementation embraces two independent yet cooperative video augmentation concepts: enriching video playback with interactive overlay elements, and erecting real-time interaction bridges between video and digital games. A total of four AVV test cases are presented to provide a hint of how our vision can be realized and of the attainable creative results.

Smartphones leak personal information about their owner when they use it to connect to the Internet. Despite recent coverage of these issues in popular media, raising awareness remains problematic since it remains largely invisible to the users. We designed a system, SASQUATCH, consisting of a netwo

Twitter, Social Networking Sites (SNSs); social media; TweetPos; geographic trends; investigative tool

Many of today's enterprise-scale wireless networks are protected by the WPA2-Enterprise Protected Extensible Authentication Protocol (PEAP). In this paper it is demonstrated how an attacker can steal a user's credentials and gain unauthorized access to such networks, by utilizing a class of vulnerable devices as MSCHAPv2 challenge response oracles. More specifically this paper explains how on these devices, Lightweight EAP (LEAP) MSCHAPv1 credentials can be captured and converted to PEAP MSCHAPv2 credentials by using a rogue Access Point. This man-in-the-middle vulnerability was found to be present in all current versions of Apple's iOS and OS X operating systems, and may impact other devices as well. A proof-of-concept implementation is available that shows how Authentication Server certificate validation and certificate pinning mechanisms may be bypassed. Mitigation strategies for the attack and protective actions which can be undertaken by end-users are also described in this paper.

Large crowds at music festivals or other mass events create challenging environments for traditional infrastructure based wireless networks. Mobile devices carried by the attendees produce large amounts of network traffic that can result in network outage or serious delays. Opportunistic networks may offer solutions to enable communication between attendees and/or organizers through direct communication between devices, without requiring a fixed infrastructure. In previous work, researchers have developed numerous opportunistic routing protocols designed to enable communication in such ad hoc networks. In this paper, several of these protocols are evaluated for the specific use case of a music festival by running simulations that make use of realistic mobility data collected during an international music festival. This analysis allows for identification of candidate routing protocols that exhibit properties that make them suitable for the envisaged scenario. The goal is to narrow down the set of candidate protocols and to eventually fine tune them to optimize their working. Based on the simulations, we propose two candidate routing protocols that are most suited for use at mass events.

To simulate crowds at mass events, realistic movement data of people is required. Despite their limited capacity for approximating real human mobility, synthetic movement models are traditionally used for this purpose. More realistic simulations can be achieved by using real-life movement data, gathered by observing people in the desired context. This paper presents a method for tracking people at mass events without the need for active cooperation by the subjects. The mechanism works by scanning at multiple locations for packets sent out by the Wi-Fi interface on visitors' smartphones, and correlating the data captured at these different locations. The proposed method can be implemented at very low cost on Raspberry Pi computers. This implementation was trialed in two different contexts: a popular music festival and a university campus. The method allows for tracking thousands of people simultaneously, and achieves a higher coverage rate than similar methods for involuntary crowd tracking. Moreover, the coverage rate is expected to increase even further as more people will start using smartphones. The proposed method has many applications in different domains. It also entails privacy implications that must be considered when deploying a similar system.

Real-time social interaction possibilities are increasingly disappearing from the media consumption process. A decisive contributing factor to this tendency is growing user location disparity. This paper proposes synchronous MediaSharing (sMS), a distributed, purely web-based framework consisting of APIs plus a back-end that enables geographically separated persons to socially consume multimedia content in a synchronized fashion. The sMS service currently resides in a proof-of-concept stage, awaiting qualitative evaluation by means of user experience research methods, yet its underlying network infrastructure as well as the majority of its principal functional components have already been designed, implemented and evaluated. A chief innovation of the system is that its exclusive reliance on open web standards warrants cross-platform support and unlocks seamless content synchronization across the physical and virtual worlds. We outline our vision for the sMS service, motivate the need for such a system, discuss its current implementation, present tentative practical results that confirm the feasibility and validity of our design, and overview the sMS feature roadmap. The sMS functionality has apparent utility in education, training and professional settings, but also in the field of entertainment and the recreational market.

In this paper, a scalable solution is presented for distributing panoramic video sequences to multiple viewers at high resolution and quality levels. In contrast to traditional broadcast scenarios, panoramic video enables the content consumer to manipulate the camera view direction and viewport size. By segmenting the panoramic input video into a set of separate sequences, transporting them over standard delivery channels and recombining them at end user side, bandwidth utilization is optimized and the quality of the video that is visualized is increased. The proposed solution, called the segmentation approach, is thoroughly explained and evaluated versus a single-stream solution with regards to several metrics, including bandwidth utilization, encoding speed, objective quality levels and seeking performance.

In this paper, we present an empirical evaluation of four spatial subdivision and three load balancing schemes that may be used for large-scale single-shard virtual environments, such as 3D networked Games and Virtual Communities. ALVIC-NG is an architecture that targets these environments and supports dynamic resource allocation. The scalability of an application based on ALVIC-NG and networked virtual environments in general is (among other factors) dependent on two metrics: the efficiency of the spatial subdivision scheme and the ability for the load balancing strategies to maintain a clustered assignment of regions to server instances. Both are discussed and evaluated in this paper through two modes of simulation. An offline version is used to decide on the most appropriate spatial subdivision scheme by creating a single-instance application that focuses on this specific metric. The online version implements the proposed strategies in ALVIC-NG and is used to both confirm the findings of the offline system and to determine the efficiency of the load balancing strategies by making use of several server instances and so-called bot servers to re-create the conditions of real-life deployments.

Abstract In this paper we describe the way in which emotion‐related data can efficiently be exchanged between participants in a large‐scale networked virtual environment. This type of metadata is extracted from real‐time captured video streams using off‐the‐shelf webcams and applied onto a two‐dimentional (2D) stylised avatar; thereby improving the immersion the user experiences while navigating and communicating in the virtual world. As emotion‐related data—once processed through the system—can be considered a specific type of state information, a generic networked virtual environment architecture can be used to distribute the information between participants. We have opted to extend the in‐house developed architecture for large‐scale virtual interactive communities (ALVIC‐NG) architecture to be able to process the information flows. We will show that the inclusion of this new type of information does not have a detrimental effect on the scalability of the system. Copyright © 2009 John Wiley & Sons, Ltd.

In this paper we present our efforts and ideas in designing a framework for networked virtual environment applications that incorporates real-time video communication between avatars. The system is primarily designed around scalability for large scale networked virtual environments. In order to achieve this, our approach is to optimize client responsibilities and make maximal use of direct client-client communication streams. Relying on a number of servers to transcode a high-quality video stream from each client into an arbitrary number of output streams is a solution that scales very poorly with a growing number of users. The main feature of importance for this paper is the ability to efficiently adapt bandwidth usage at client side through multiple multicast groups. Server tasks are reduced to the bare minimum, thereby ensuring a highly scalable end-result, depending only on the processing power of the connected clients. We do show however, that by introducing a simple server setup, the system can be adapted to be deployed on xDSL based access networks