This study delves into the mechanisms that spark user curiosity driving active engagement within public Telegram groups. By analyzing approximately 6 million messages from 29,196 users across 409 groups, we identify and quantify the key factors that stimulate users to actively participate (i.e., send messages) in group discussions. These factors include social influence, novelty, complexity, uncertainty, and conflict, all measured through metrics derived from message sequences and user participation over time. After clustering the messages, we apply explainability techniques to assign meaningful labels to the clusters. This approach uncovers macro categories representing distinct curiosity stimulation profiles, each characterized by a unique combination of various stimuli. Social influence from peers and influencers drives engagement for some users, while for others, rare media types or a diverse range of senders and media sparks curiosity. Analyzing patterns, we found that user curiosity stimuli are mostly stable, but, as the time between the initial message increases, curiosity occasionally shifts. A graph-based analysis of influence networks reveals that users motivated by direct social influence tend to occupy more peripheral positions, while those who are not stimulated by any specific factors are often more central, potentially acting as initiators and conversation catalysts. These findings contribute to understanding information dissemination and spread processes on social media networks, potentially contributing to more effective communication strategies.

Mobility is a fundamental aspect of human life, and mobility data offers valuable insights into user behavior. Yet, this data also exposes users to privacy risks given pattern unicity in their trajectories, i.e., the singularity in the displacements made by users. Existing strategies to quantify such user exposure either focus only on the sequences of places visited by each user, as the widely used uniqueness measure, or are tied to specific attack models. We here introduce MoBES, a novel, scalable, customizable and highly interpretable measure of user exposure in mobility data. MoBES leverages multiple existing metrics to build a multi-dimensional space, which in turn is used to capture each user's mobility signature behavior. MoBES quantifies user exposure based on how distinct a user's signature is from her neighbors in the defined metric space. As such, MoBES is designed to be a fundamental expression of user behavior, and not tied to any specific attack model. We evaluate MoBES on a real mobility dataset, showing that it effectively captures user exposure within the behavioral metric space. We also compare MoBES with the uniqueness measure, showing that MoBES is able to uncover users who, even though visiting the same places as others in the crowd, are still at risk of exposure due to the unicity of their mobility behavior.

Web data are often modelled as complex networks in which entities interact and form communities. Nevertheless, web data evolves over time, and network communities change alongside it. This makes Community Detection (CD) in dynamic graphs a relevant problem, calling for evolutionary CD algorithms. The choice and evaluation of such algorithm performance is challenging because of the lack of a comprehensive set of benchmarks and specific metrics. To address these challenges, we propose CoDÆN—Community Detection Algorithms in Evolving Networks—a benchmarking framework for evolutionary CD algorithms in dynamic networks, that we offer as open source to the community. CoDÆN allows us to generate synthetic community-structured graphs with known ground truth and design evolving scenarios combining nine basic graph transformations that modify edges, nodes, and communities. We propose three complementary metrics (i.e., Correctness, Delay, and Stability) to compare evolutionary CD algorithms. Armed with CoDÆN, we consider three evolutionary modularity-based CD approaches, dissecting their performance to gauge the trade-off between the stability of the communities and their correctness. Next, we compare the algorithms in real Web-oriented datasets, confirming such a trade-off. Our findings reveal that algorithms that introduce memory in the graph maximise stability but add delay when abrupt changes occur. Conversely, algorithms that introduce memory by initialising the CD algorithms with the previous solution fail to identify the split and birth of new communities. These observations underscore the value of CoDÆN in facilitating the study and comparison of alternative evolutionary community detection algorithms.

Mobility patterns are inherently linked to human nature (e.g., individual variability, temporal dynamics, behavioral factors, curiosity, social interaction), making mobility prediction a multifaceted and challenging problem that requires sophisticated models and comprehensive data. Machine learning (ML) models excel at predicting the location a person will be at the next time interval, but they often raise privacy concerns. To address these privacy issues while maintaining the benefits of ML models, Federated Learning (FL) offers a distributed framework that enables collaborative training of human mobility prediction models without requiring the sharing of highly sensitive location data. However, in the domain of FL for individual mobility prediction, prior work lacks a thorough understanding of the many factors that may impact the performance of FL-based prediction models. In this work, we provide a comprehensive study of the impact of various aspects related to human behavior, data characteristics, ML algorithmic solutions, and FL architectural structuring. We quantify the impact of such factors on effectiveness (accuracy) and efficiency (execution time, memory, and energy usage) of the prediction, revealing that, ignoring these factors lead to misleading result interpretation, and acknowledging them empowers both effectiveness and efficiency results.

Social media has gained importance as a channel to influence people's behavior and decisions, affecting not only the online world but also real-life (offline) events. This is especially evident in Brazil, where platforms like Telegram have been instrumental in disseminating political content rapidly and widely. However, the potential coordinated use of Telegram for promoting specific political narratives at critical times, such as the 2022 Brazilian elections, remains an area that requires further investigation. This study aims to investigate this phenomenon, focusing on the first and second rounds of voting and the January 8th riots. To this end, we conducted a comprehensive analysis of 620,000 messages from 256 Telegram groups, focusing on the dynamics of message dissemination and user interactions. Using network backbone extraction and text analysis methods, we identified key users who may be orchestrating the distribution of content. Our findings suggest that these individuals play a central role in the network's topology, relaying messages to a broader audience on dominant topics of discussion that reflect Brazil's political landscape during this turbulent period. This study not only highlights the growing influence of messaging apps on political mobilization but also contributes to our understanding of digital communication strategies in modern electoral contexts, emphasizing the need for further research in this field.

Offloading high-demanding applications to the edge provides better quality of experience (QoE) for users with limited hardware devices. However, to maintain a competitive QoE, infrastructure, and service providers must adapt to users’ different mobility patterns, which can be challenging, especially for location-based services (LBS). Another issue that needs to be tackled is the increasing demand for user privacy protection. With less (accurate) information regarding user location, preferences, and usage patterns, forecasting the performance of offloading mechanisms becomes even more challenging. This work discusses the impacts of users’ privacy and mobility when offloading to the edge. Different privacy and mobility scenarios are simulated and discussed to shed light on the trade-offs (e.g., privacy protection at the cost of increased latency) among privacy protection, mobility, and offloading performance.

A large volume of content related to claims of election fraud, often associated with hate speech and extremism, was reported on Twitter during the 2020 US election, with evidence that coordinated efforts took place to promote such content on the platform. In response, Twitter announced the suspension of thousands of user accounts allegedly involved in such actions. Motivated by these events, we here propose a novel network-based approach to uncover evidence of coordination in a set of user interactions. Our approach is designed to address the challenges incurred by the often sheer volume of noisy edges in the network (i.e., edges that are unrelated to coordination) and the effects of data sampling. To that end, it exploits the joint use of two network backbone extraction techniques, namely Disparity Filter and Neighborhood Overlap, to reveal strongly tied groups of users (here referred to as communities) exhibiting repeatedly common behavior, consistent with coordination. We employ our strategy to a large dataset of tweets related to the aforementioned fraud claims, in which users were labeled as suspended, deleted or active, according to their accounts status after the election. Our findings reveal well-structured communities, with strong evidence of coordination to promote (i.e., retweet) the aforementioned fraud claims. Moreover, many of those communities are formed not only by suspended and deleted users, but also by users who, despite exhibiting very similar sharing patterns, remained active in the platform. This observation suggests that a significant number of users who were potentially involved in the coordination efforts went unnoticed by the platform, and possibly remained actively spreading this content on the system.

Attention is a scarce resource disputed by algorithms and people on the Internet. This competition for attention is part of online spaces especially online small groups where there is a limited number of individuals interacting with each other using text and media content that is not controlled by algorithms or human curators. In these groups, as certain participants and piece of content can catch the collective attention, a question that naturally arises is: how to analyze topic attention in online small groups? In this paper, we propose a methodology aimed at answering this question. Our proposal consists of sets of analyses over topical (obtained from topic analysis) transition graphs for characterizing attention allocation, permanence and shifting as well as participant role characterization during discussions in online small groups. We experimented with our methodology using WhatsApp groups as a case study. Among other results, we identified and characterized abrupt and smooth topic transitions as well as patterns of participant activity related to certain topics.

WhatsApp is a free messaging app with more than one billion active monthly users which has become one of the main communication platforms in many countries, including Saudi Arabia, Germany, and Brazil. In addition to allowing the direct exchange of messages among pairs of users, the app also enables group conversations, where multiple people can interact with one another. A number of recent studies have shown that WhatsApp groups play an important role as an information dissemination platform, especially during important social mobilization events. In this paper, we build upon those prior efforts by taking a first look into the use of audio messages in WhatsApp groups, a type of content that is becoming increasingly important in the platform. We present a methodology to analyze audio messages shared in WhatsApp groups, characterizing content properties (e.g, topics and language characteristics), their propagation dynamics and the impact of different types of audios (e.g., speech versus music) on such dynamics.

Big data applications and analytics are employed in many sectors for a variety of goals: improving customers satisfaction, predicting market behavior or improving processes in public health. These applications consist of complex software stacks that are often run on cloud systems. Predicting execution times is important for estimating the cost of cloud services and for effectively managing the underlying resources at runtime. Machine Learning (ML), providing black box solutions to model the relationship between application performance and system configuration without requiring in-detail knowledge of the system, has become a popular way of predicting the performance of big data applications. We investigate the cost-benefits of using supervised ML models for predicting the performance of applications on Spark, one of today's most widely used frameworks for big data analysis. We compare our approach with Ernest (an ML-based technique proposed in the literature by the Spark inventors) on a range of scenarios, application workloads, and cloud system configurations. Our experiments show that Ernest can accurately estimate the performance of very regular applications, but it fails when applications exhibit more irregular patterns and/or when extrapolating on bigger data set sizes. Results show that our models match or exceed Ernest's performance, sometimes enabling us to reduce the prediction error from 126-187% to only 5-19%.

Curiosity is a natural trait of human behavior. When we take into account the time we spend consuming content online, it is expected that at least a fraction of that time was driven by curious behavior. Aiming at understanding how curiosity drives online information consumption, we here propose a model that captures user curiosity relying on several stimulus metrics. Our model relies on the well-established Wundt's curve from psychology and is based on metrics capturing Novelty, Complexity and Uncertainty as key stimuli driving one's curiosity. As a case study, we apply our model on a dataset of online music consumption from LastFM. We found that there are four main types of user behaviors in terms of how the curiosity stimulus metrics drive the user accesses to online music. These are characterized based on the diversity in the songs, artists and musical genres accessed.

Big data applications are among the most suitable applications to be executed on cluster resources because of their high requirements of computational power and data storage.Correctly sizing the resources devoted to their execution does not guarantee they will be executed as expected.Nevertheless, their execution can be affected by perturbations which can change the expected execution time.Identifying when these types of issue occurred by comparing their actual execution time with the expected one is mandatory to identify potentially critical situations and to take the appropriate steps to prevent them.To fulfill this objective, accurate estimates are necessary.In this paper, machine learning techniques coupled with a posteriori knowledge are exploited to build performance estimation models.Experimental results show how the models built with the proposed approach are able to outperform a reference state-of-the-art method (i.e., Ernest method), reducing in some scenarios the error from the 221.09-167.07%to 13.15-30.58%.

Human mobility has been studied from different perspectives. One approach addresses predictability, deriving theoretical limits on the accuracy that any prediction model can achieve in a given dataset. This approach focuses on the inherent nature and fundamental patterns of human behavior captured in the dataset, filtering out factors that depend on the specificities of the prediction method adopted. In this paper, we revisit the state-of-the-art method for estimating the predictability of a person's mobility, which, despite being widely adopted, suffers from low interpretability and disregards external factors that have been suggested to improve predictability estimation, notably the use of contextual information (e.g., weather, day of the week, and time of the day). We also conduct a thorough analysis of how this widely used method works, by looking into two different measures (one proposed by us) which are easier to understand and, as shown, capture reasonably well the effects of the original technique. Additionally, we investigate strategies to incorporate different types of contextual information into predictability estimates, and show that the benefits vary depending on the underlying prediction task. Finally, we propose and evaluate alternative estimates of predictability which, while being much easier to interpret, provide comparable results to the state-of-the-art.

Today, video streaming is a well-established application on the internet. Providers are able of transmitting video globally, in large-scale, although there are still some challenges, such as offering high video quality for all clients, particularly in live transmission, due to limited butter size and intense client arrival. Low quality result in low engagement (i.e. client permanence). To increase engagement, the provider can adopt strategies to anticipate (i.e. predict) quality losses. To make this possible, the provider needs to learn what session characteristics lead to an early session abandonment. If a new session presents such characteristics, the provider can presume that it will terminate prematurely. To implement an accurate prediction scheme we must choose a feature set able to explain the user permanence (i.e. engagement). Traditional prediction approaches use user-centered quality metrics such stall rate and average bitrate. In our data-set, these metrics give a prediction accuracy of less than 70%. To improve this, we propose a new set of metrics that explores the client adaptation flow. We model it through Markov chains and show that this increases the prediction accuracy up to 82%. Besides this, grouping similar sessions and train a predictor for each group can increase the accuracy. Finally, we present a use case for the predictor in order to investigate its feasibility in real systems.

WhatsApp has revolutionized the way people communicate and interact. It is not only cheaper than the traditional Short Message Service (SMS) communication but it also brings a new form of mobile communication: the group chats. Such groups are great forums for collective discussions on a variety of topics. In particular, in events of great social mobilization, such as strikes and electoral campaigns, WhatsApp group chats are very attractive as they facilitate information exchange among interested people. Yet, recent events have raised concerns about the spreading of misinformation in WhatsApp. In this work, we analyze information dissemination within WhatsApp, focusing on publicly accessible political-oriented groups, collecting all shared messages during major social events in Brazil: a national truck drivers' strike and the Brazilian presidential campaign. We analyze the types of content shared within such groups as well as the network structures that emerge from user interactions within and cross-groups. We then deepen our analysis by identifying the presence of misinformation among the shared images using labels provided by journalists and by a proposed automatic procedure based on Google searches. We identify the most important sources of the fake images and analyze how they propagate across WhatsApp groups and from/to other Web platforms.

Political campaigns have frequently used the online social network as an important environment to exhibit the candidate ideas, their activities, and their electoral plans if elected. Some users are more politically engaged than others. As an example, we can observe intense political debates, especially during major campaigns on Twitter. In such context, this paper presents a characterization of politically engaged user groups on Twitter during the 2016 US Presidential Campaign. Using a rich dataset with 23 million tweets, 115 thousand user profiles and their contact network collected from January 2016 to November 2016, we identified four politically engaged user groups: advocates for both main candidates, political bots, and regular users. We present a characterization of how Twitter users behave during a political campaign through the language patterns analysis of tweets, which users receive more popularity during the campaign and how tweets from each candidate may have affected their mood variation, as expressed by the messages they share.

Data heterogeneity and irregularity are key characteristics of big data applications that often overwhelm the existing software and hardware infrastructures. In such context, the exibility and elasticity provided by the cloud computing paradigm over a natural approach to cost-effectively adapting the allocated resources to the application's current needs. Yet, the same characteristics impose extra challenges to predicting the performance of cloud-based big data applications, a central step in proper management and planning. This paper explores two modeling approaches for performance prediction of cloud-based big data applications. We evaluate a queuing-based analytical model and a novel fast ad-hoc simulator in various scenarios based on different applications and infrastructure setups. Our results show that our approaches can predict average application execution times with 26% relative error in the very worst case and about 12% on average. Moreover, our simulator provides performance estimates 70 times faster than state of the art simulation tools.

In modern web, users contact lots of services, identified by the domain name of the server. The temporal sequence and transitions of visited domains form a trajectory of the user on the web. In this work, we analyze 4 weeks of such trajectories, extracted from logs collected in our university network, and mine them via big data and machine learning methodologies to extract the interests of users. Our goal is to create a model of such trajectories and find similarities so to observe peculiarity of users' browsing. Thanks to the model, we propose a methodology to automatically group together the trajectories of single users and/or communities into highly descriptive environments which in turn allow the analyst to identify the topic of interest. We propose an automatic way to highlight differences in terms of popularity and content of environments. Lastly, we analyze the transition among environments, showing how people in smaller communities, e.g., in the same department, have a much more homogeneous behaviour than people at large, e.g., in the university.

Understanding the impact of performance degradation on users' QoE during live Internet streaming is key to maximize the audience and increase content providers' revenues. It is known that some problems have a strong correlation with low QoE--e.g., users experiencing video stalls tend to leave video sessions earlier. It is, however, mostly unknown whether such observations hold for live streaming of large-scale events (e.g., the FIFA World Cup). Such events are particular due to the widespread interest in the streamed content, reaching an impressively high audience worldwide. We study whether and to what extent performance degradation during live streaming of large-scale events affects users' QoE. We leverage a unique dataset collected from a major content provider in South America during the 2014 FIFA Soccer World Cup. We first extract performance metrics from the logs: stream bitrate, bitrate switches, playback stalls, and playback startup latency. We then correlate these performance metrics with session duration, which we use as a QoE indicator. We confirm the strong correlations between the metrics and QoE indicators; in particular, frequent stalls are often accompanied by higher probability of early session termination. Moreover, we quantify how such correlations vary according to broadcast matches and client terminals. Some of our findings challenge intuition--e.g., we find that PC users seem more tolerant to problems than users on mobile terminals. Our results provide better understanding of user QoE and are an important step towards user QoE models in large-scale events.

Personal Cloud Storage (PCS) is a very popular Internet service. It allows users to backup data to the cloud as well as to perform collaborative work while sharing content. Notably, content sharing is a key feature for PCS users. It however comes with extra costs for service providers, as shared files must be synchronized to multiple user devices, generating more downloads from cloud servers. Despite the increasing interest in this type of service, a thorough investigation on the costs and benefits of PCS for service providers and end users has not been conducted yet. To that end, we propose a model to analyze cost-benefit tradeoffs for both parties. We develop utility functions that capture, in an abstract level, the satisfaction of the service provider and users in various scenarios. Then, we apply our model to evaluate alternative policies for content sharing in PCS. We consider two alternative policies for the current PCS sharing architecture, which count on user collaboration to reduce providers' costs. Our results show that such policies are advantageous for providers and users, leading to 39% utility improvements for both parties, while requiring low commitment of resources from participating users.

Tags (keywords freely assigned by users to describe web content) have become highly popular on Web 2.0 applications, because of the strong stimuli and easiness for users to create and describe their own content. This increase in tag popularity has led to a vast literature on tag recommendation methods. These methods aim at assisting users in the tagging process, possibly increasing the quality of the generated tags and, consequently, improving the quality of the information retrieval (IR) services that rely on tags as data sources. Regardless of the numerous and diversified previous studies on tag recommendation, to our knowledge, no previous work has summarized and organized them into a single survey article. In this article, we propose a taxonomy for tag recommendation methods, classifying them according to the target of the recommendations, their objectives, exploited data sources, and underlying techniques. Moreover, we provide a critical overview of these methods, pointing out their advantages and disadvantages. Finally, we describe the main open challenges related to the field, such as tag ambiguity, cold start, and evaluation issues.

Which song will Smith listen to next? Which restaurant will Alice go to tomorrow? Which product will John click next? These applications have in common the prediction of user trajectories that are in a constant state of flux over a hidden network (e.g. website links, geographic location). Moreover, what users are doing now may be unrelated to what they will be doing in an hour from now. Mindful of these challenges we propose TribeFlow, a method designed to cope with the complex challenges of learning personalized predictive models of non-stationary, transient, and time-heterogeneous user trajectories. TribeFlow is a general method that can perform next product recommendation, next song recommendation, next location prediction, and general arbitrary-length user trajectory prediction without domain-specific knowledge. TribeFlow is more accurate and up to 413x faster than top competitors.

Tag recommendation strategies that exploit term co‐occurrence patterns with tags previously assigned to the target object have consistently produced state‐of‐the‐art results. However, such techniques work only for objects with previously assigned tags. Here we focus on tag recommendation for objects with no tags, a variation of the well‐known \textit{cold start} problem. We start by evaluating state‐of‐the‐art co‐occurrence based methods in cold start. Our results show that the effectiveness of these methods suffers in this situation. Moreover, we show that employing various automatic filtering strategies to generate an initial tag set that enables the use of co‐occurrence patterns produces only marginal improvements. We then propose a new approach that exploits both positive and negative user feedback to iteratively select input tags along with a genetic programming strategy to learn the recommendation function. Our experimental results indicate that extending the methods to include user relevance feedback leads to gains in precision of up to 58% over the best baseline in cold start scenarios and gains of up to 43% over the best baseline in objects that contain some initial tags (i.e., no cold start). We also show that our best relevance‐feedback‐driven strategy performs well even in scenarios that lack user cooperation (i.e., users may refuse to provide feedback) and user reliability (i.e., users may provide the wrong feedback).

Automatic text classification (TC) continues to be a relevant research topic and several TC algorithms have been proposed. However, the majority of TC algorithms assume that the underlying data distribution does not change over time. In this work, we are concerned with the challenges imposed by the temporal dynamics observed in textual data sets. We provide evidence of the existence of temporal effects in three textual data sets, reflected by variations observed over time in the class distribution, in the pairwise class similarities, and in the relationships between terms and classes. We then quantify, using a series of full factorial design experiments, the impact of these effects on four well‐known TC algorithms. We show that these temporal effects affect each analyzed data set differently and that they restrict the performance of each considered TC algorithm to different extents. The reported quantitative analyses, which are the original contributions of this article, provide valuable new insights to better understand the behavior of TC algorithms when faced with nonstatic (temporal) data distributions and highlight important requirements for the proposal of more accurate classification models.

The design and evaluation of tag recommendation methods has historically focused on maximizing the relevance of the suggested tags for a given object, such as a movie or a song. However, relevance by itself may not be enough to guarantee recommendation usefulness. Promoting novelty and diversity in tag recommendation not only increases the chances that the user will select “some” of the recommended tags but also promotes complementary information (i.e., tags), which helps to cover multiple aspects or topics related to the target object. Previous work has addressed the tag recommendation problem by exploiting at most two of the following aspects: (1) relevance, (2) explicit topic diversity, and (3) novelty. In contrast, here we tackle these three aspects conjointly, by introducing two new tag recommendation methods that cover all three aspects of the problem at different levels. Our first method, called Random Forest with topic-related attributes , or RF t , extends a relevance-driven tag recommender based on the Random Forest ( RF ) learning-to-rank method by including new tag attributes to capture the extent to which a candidate tag is related to the topics of the target object. This solution captures topic diversity as well as novelty at the attribute level while aiming at maximizing relevance in its objective function. Our second method, called Explicit Tag Recommendation Diversifier with Novelty Promotion , or xTReND , reranks the recommendations provided by any tag recommender to jointly promote relevance, novelty, and topic diversity. We use RF t as a basic recommender applied before the reranking, thus building a solution that addresses the problem at both attribute and objective levels. Furthermore, to enable the use of our solutions on applications in which category information is unavailable, we investigate the suitability of using latent Dirichlet allocation (LDA) to automatically generate topics for objects. We evaluate all tag recommendation approaches using real data from five popular Web 2.0 applications. Our results show that RF t greatly outperforms the relevance-driven RF baseline in diversity while producing gains in relevance as well. We also find that our new xTReND reranker obtains considerable gains in both novelty and relevance when compared to that same baseline while keeping the same relevance levels. Furthermore, compared to our previous reranker method, xTReD , which does not consider novelty, xTReND is also quite effective, improving the novelty of the recommended tags while keeping similar relevance and diversity levels in most datasets and scenarios. Comparing our two new proposals, we find that xTReND considerably outperforms RF t in terms of novelty and diversity with only small losses (under 4%) in relevance. Overall, considering the trade-off among relevance, novelty, and diversity, our results demonstrate the superiority of xTReND over the baselines and the proposed alternative, RF t . Finally, the use of automatically generated latent topics as an alternative to manually labeled categories also provides significant improvements, which greatly enhances the applicability of our solutions to applications where the latter is not available.

In the past years a lot of effort has been spent analyzing online social network data to understand how the world reality is reflected in the "virtual" world. Twitter is by far the network most used in these studies, given its policy of public data availability. However, a big discussion is still on on whether the data available is enough to make user characterization or event outcomes prediction, and what are the pitfalls people do not usually account for. In this direction, we propose a new methodology for drawing representative samples from Twitter data, which is divided into four phases: (i) user filtering, (ii) user demographic characterization, (iii) user sampling, and (iv) event prediction. The methodology is tested into a common scenario in Twitter event outcome prediction: elections. The methodology was tested with municipality elections from six different Brazilian cities, and compared to official election results. Results show it is worth further investigating the topic, but that a very hight number of messages is required to match real data distributions.

In this paper we characterize the impact of failures in Brazil's national research network (RNP) on traffic at a large client university. We analyze reachability disruptions, caused by failures of RNP's interdomain links, that block all international traffic. We also analyze performance disruptions, caused by simultaneous failure of multiple RNP intradomain links, which result in congestion and performance degradation. We study the impact of disruptions on traffic, application mix, and user behavior. Our results show that users adapt their behavior when some applications become unavailable and when network performance degrades. For example, users tend to migrate to Youtube when Facebook becomes unavailable during reachability disruptions; similarly, users migrate to Facebook when congestion during performance disruptions severely degrade Youtube experience. We also correlate the impact of disruptions to network topology and show that performance during a performance disruption depends on the location and importance of failed links.

We here investigate what drives the popularity of information on social media platforms. Focusing on YouTube, we seek to understand the extent to which content by itself determines a video's popularity. Using mechanical turk as experimental platform, we asked users to evaluate pairs of videos, and compared users' relative perception of the videos' content against their relative popularity reported by YouTube. We found that in most evaluations users could not reach consensus on which video had better content as their perceptions tend to be very subjective. Nevertheless, when consensus was reached, the video with preferred content almost always achieved greater popularity on YouTube, highlighting the importance of content in driving information popularity on social media.

Cloud storage systems are currently very popular, generating a large amount of traffic. Indeed, many companies offer this kind of service, including worldwide providers such as Dropbox, Microsoft and Google. These companies, as well as new providers entering the market, could greatly benefit from knowing typical workload patterns that their services have to face in order to develop more cost-effective solutions. However, despite recent analyses of typical usage patterns and possible performance bottlenecks, no previous work investigated the underlying client processes that generate workload to the system. In this context, this paper proposes a hierarchical two-layer model for representing the Dropbox client behavior. We characterize the statistical parameters of the model using passive measurements gathered in 3 different network vantage points. Our contributions can be applied to support the design of realistic synthetic workloads, thus helping in the development and evaluation of new, well-performing personal cloud storage services.

The high popularity of human communication based on mobile phone services exhibits different behaviors in different places and time. Large-scale events, such as the FIFA World Cup, create an auspicious atmosphere for the attendees to use such services to share their experiences with relatives and friends. This work characterizes the workload dynamics of a major mobile phone network on different days when major soccer matches took place in Rio de Janeiro, Brazil. We analyze the usage behavior of three profiles of attendees of two major matches of the FIFA 2014 World Cup, including the final between Germany and Argentina on July 13, 2014. We compare these profiles with similar results computed on a dataset of calls and SMS collected on the day of the final match of 2013 Brazilian Cup, which happened at the same stadium of the World Cup final. We note that users of each analyzed profile, i.e. (i) attendees who made cellphone calls and sent SMS messages, (ii) attendees who only made calls, and (iii) attendees who only sent SMS, have different behaviors in different periods of the event.

Understanding the factors that impact the popularity dynamics of social media can drive the design of effective information services, besides providing valuable insights to content generators and online advertisers. Taking YouTube as case study, we analyze how video popularity evolves since upload, extracting popularity trends that characterize groups of videos. We also analyze the referrers that lead users to videos, correlating them, features of the video and early popularity measures with the popularity trend and total observed popularity the video will experience. Our findings provide fundamental knowledge about popularity dynamics and its implications for services such as advertising and search.

We here investigate the potential of participatory sensor networks derived from location sharing systems, such as Foursquare, to understand the human dynamics of cities. We propose the City Image visualization technique, which builds a transition graph mapping people's movements between location categories, and demonstrate its use to identify similarities and differences of human dynamics across cities by clustering cities according to their citizens' routines. We also analyze centrality metrics of the transition graphs built for different cities, considering transitions between specific venues. We show that these metrics complement the City Image technique, contributing to a deeper understanding of city dynamics.

Participatory sensing systems (PSSs) have the potential to become fundamental tools to support the study, in large scale, of urban social behavior and city dynamics. To that end, this work characterizes the photo sharing system Instagram, considered one of the currently most popular PSS on the Internet. Based on a dataset of approximately 2.3 million shared photos, we characterize user's behavior in the system showing that there are several advantages and opportunities for large scale sensing, such as a global coverage at low cost, but also challenges, such as a very unequal photo sharing frequency, both spatially and temporally. We also observe that the temporal photo sharing pattern is a good indicator about cultural behaviors, and also says a lot about certain classes of places. Moreover, we present an application to identify regions of interest in a city based on data obtained from Instagram, which illustrates the promising potential of PSSs for the study of city dynamics.

The large volume of unwanted email (spam) traffic wastes network resources. We have previously proposed SpaDeS, a method for spammer detection at the source network, which uses only network-layer metrics. We here present an extension of SpaDeS, focusing on its diversity and adaptability to new behavior patterns of spammers. To that end, we propose the use of a new active-learning-based strategy to select new, very informative, training samples, aiming at reducing the loss of effectiveness over time. The new method was applied to a real data set and the results show that, despite some variation in performance, the use of active learning to better select the training set improves the classification of legitimate users by as much as 21%, with just a small performance loss (less than 3%) in spammer classification.

Cities are not identical and evolve over time, and sensing in large scale can be used to capture these differences. Research in Wireless Sensor Networks has provided several tools, techniques and algorithms to solve the problem of sensing in restricted scenarios (e.g., factory). However, sensing large scale areas, such as big cities, brings many challenges and incurs high costs related to system building and management. Thus, sensing those areas becomes more feasible when people collaborate among themselves using their portable devices, and building what has been named participatory sensing systems. This work analyzes an emerging type of network derived from this type of system, the Participatory Sensor Network (PSN), where nodes are autonomous mobile entities and the sensing depends on whether they want to participate in the sensing process. Based on four datasets of participatory sensing systems (27 million of records), we show that this type of network brings many challenges related to structural problems, e.g. instant coverage very limited, and also because of big data issues, which may restrict the use of this emerging type of network. However, it presents also, as shown here, many advantages and open opportunities, mainly related to large scale study of cities dynamics.

Tag recommendation approaches have historically focused on maximizing the relevance of the recommended tags for a given object, such as a movie or a song. Nevertheless, different users may be interested in the same object for different reasons---for instance, the Star Wars movies may appeal to both adventure as well as to fantasy movie fans. In this situation, a sensible strategy is to provide a user with diverse recommendations of how to tag the object. In this paper, we address the problem of recommending relevant and diverse tags as a ranking problem. In particular, we propose a novel tag recommendation approach that explicitly takes into account the possible topics (e.g., categories) underlying an object in order to promote tags with high coverage and low redundancy with respect to these topics. We thoroughly evaluate our proposed approach using data collected from two popular Web 2.0 applications, namely, LastFM and MovieLens. Our experimental results attest the effectiveness of our approach at promoting more relevant and diverse tags in contrast to state-of-the-art relevance-based methods as well as a recently proposed method that takes both relevance and diversity into account.

The advent of the Web 2.0 has given rise to an interesting phenomenon: there is currently much more data than what can be effectively analyzed without relying on sophisticated automatic tools. Some of these tools, which target the organization and extraction of useful knowledge from this huge amount of data, rely on machine learning and data or text mining techniques, specifically automatic document classification algorithms. However, these algorithms are still a computational challenge because of the volume of data that needs to be processed. Some of the strategies available to address this challenge are based on the parallelization of ADC algorithms. In this work, we present GPU-NB, a parallel version of one of the most widely used document classification algorithms, the Naïve Bayes, that uses graphics processing units (GPUs). In our evaluation using 6 different document collections, we show that the GPU-NB can maintain the same classification effectiveness (in most cases) while increasing the efficiency by up to 34x faster than its sequential version using CPU. GPU-NB is also up to 11x faster than a CPU-based parallel implementation of Naive Bayes running with 4 threads. Moreover, assuming an optimistic behavior of the CPU parallelization, GPU-NB should outperform the CPU-based implementation with up to 32 cores, at a small fraction of the cost. We also show that the efficiency of the GPU-NB parallelization is impacted by features of the document collections, particularly the number of classes, although the density of the collection (average number of occurrences of terms per document) has a significant impact as well.

Predicting Web content popularity is an important task for supporting the design and evaluation of a wide range of systems, from targeted advertising to effective search and recommendation services. We here present two simple models for predicting the future popularity of Web content based on historical information given by early popularity measures. Our approach is validated on datasets consisting of videos from the widely used YouTube video-sharing portal. Our experimental results show that, compared to a state-of-the-art baseline model, our proposed models lead to significant decreases in relative squared errors, reaching up to 20% reduction on average, and larger reductions (of up to 71%) for videos that experience a high peak in popularity in their early days followed by a sharp decrease in popularity.

We here propose a new method for expanding entity related queries that automatically filters, weights and ranks candidate expasion terms extracted from Wikipedia articles related to the original query. Our method is based on state-of-the-art tag recommendation methods that exploit heuristic metrics to estimate the descriptive capacity of a given term. Originally proposed for the context of tags, we here apply these recommendation methods to weight and rank terms extracted from multiple fields of Wikipedia articles according to their relevance for the article. We evaluate our method comparing it against three state-of-the-art baselines in three collections. Our results indicate that our method outperforms all baselines in all collections, with relative gains in MAP of up to 14% against the best ones.

In the last few years, the increasing interest in location-based services (LBS) has favored the introduction of geo-referenced information in various Web 2.0 applications, as well as the rise of location-based social networks (LBSN). Foursquare, one of the most popular LBSNs, gives incentives to users who visit (check in) specific places (venues) by means of, for instance, mayorships to frequent visitors. Moreover, users may leave tips at specific venues as well as mark previous tips as done in sign of agreement. Unlike check ins, which are shared only with friends, the lists of mayorships, tips and dones of a user are publicly available to everyone, thus raising concerns about disclosure of the user's movement patterns and interests. We analyze how users explore these publicly available features, and their potential as sources of information leakage. Specifically, we characterize the use of mayorships, tips and dones in Foursquare based on a dataset with around 13 million users. We also analyze whether it is possible to easily infer the home city (state and country) of a user from these publicly available information. Our results indicate that one can easily infer the home city of around 78% of the analyzed users within 50 kilometers.

Peer-to-Peer live video streaming systems are becoming increasingly popular. Nevertheless, in spite of various studies of client behavior aspects and system optimizations, the current knowledge about the dynamic properties of the system, particularly how the P2P overlay network changes over time during a live transmission, is still superficial. In this paper, we provide a characterization of the dynamic properties of a popular P2P live streaming media application, namely Sop Cast. We use complex network metrics to analyze how the structure of the network evolves over time from the perspective of individual nodes (local view) and of the whole network (global view). We find that Sop Cast peers may be clustered into three profiles based on their centrality properties in the network. Moreover, in spite of peers changing their partners over time, they tend to remain with the same centrality profile. Also, the global network structure tends to remain roughly stable over time, except for a decaying clustering coefficient. Our findings can be used to generate more realistic synthetic P2P workloads and to drive future system designs and simulations.

Online Location Based Social Networks (LBSNs), which combine social network features with geographic information sharing, are becoming increasingly popular. One such application is Foursquare, which doubled its user population in less than six months. Among other features, Foursquare allows users to leave tips (i.e., reviews or recommendations) at specific venues as well as to give feedback on previously posted tips by adding them to their to-do lists or marking them as done. In this paper, we analyze how Foursquare users exploit these three features - tips, dones and to-dos - uncovering different behavior profiles. Our study reveals the existence of very active and influential users, some of which are famous businesses and brands, that seem engaged in posting tips at a large variety of venues while also receiving a great amount of user feedback on them. We also provide evidence of spamming, showing the existence of users that post tips whose contents are unrelated to the nature or domain of the venue where the tips were left.

A number of online video sharing systems, out of which YouTube is the most popular, provide features that allow users to post a video as a response to a discussion topic. These features open opportunities for users to introduce polluted content, or simply pollution, into the system. For instance, spammers may post an unrelated video as response to a popular one, aiming at increasing the likelihood of the response being viewed by a larger number of users. Moreover, content promoters may try to gain visibility to a specific video by posting a large number of (potentially unrelated) responses to boost the rank of the responded video, making it appear in the top lists maintained by the system. Content pollution may jeopardize the trust of users on the system, thus compromising its success in promoting social interactions. In spite of that, the available literature is very limited in providing a deep understanding of this problem. In this paper, we address the issue of detecting video spammers and promoters. Towards that end, we first manually build a test collection of real YouTube users, classifying them as spammers, promoters, and legitimate users. Using our test collection, we provide a characterization of content, individual, and social attributes that help distinguish each user class. We then investigate the feasibility of using supervised classification algorithms to automatically detect spammers and promoters, and assess their effectiveness in our test collection. While our classification approach succeeds at separating spammers and promoters from legitimate users, the high cost of manually labeling vast amounts of examples compromises its full potential in realistic scenarios. For this reason, we further propose an active learning approach that automatically chooses a set of examples to label, which is likely to provide the highest amount of information, drastically reducing the amount of required training data while maintaining comparable classification effectiveness.

The Simple Network Management Protocol (SNMP), proposed almost 20 years ago, still remains as the de facto standard for TCP/IP networks managing. However, there are few studies that present the real usage of SNMP in production networks. This work aims at presenting a characterization study that was performed over a sample of the management traffic from the Brazilian Education and Research Network (RNP). The achieved results may help both academy and industry in the development of new technologies for managing networks, in order to better meet the needs of the production networks users and administrators.

MapReduce is an important paradigm to support modern data-intensive applications. In this paper we address the challenge of modeling performance of one implementation of MapReduce called Hadoop Online Prototype (HOP), with a specific target on the intra-job pipeline parallelism. We use a hierarchical model that combines a precedence model and a queuing network model to capture the intra-job synchronization constraints. We first show how to build a precedence graph that represents the dependencies among multiple tasks of the same job. We then apply it jointly with an approximate Mean Value Analysis (aMVA) solution to predict mean job response time and resource utilization. We validate our solution against a queuing network simulator in various scenarios, finding that our performance model presents a close agreement, with maximum relative difference under 15%.

This work addresses the task of recommending relevant tags to a target object by jointly exploiting three dimensions of the problem: (i) term co-occurrence with tags pre-assigned to the target object, (ii) terms extracted from multiple textual features, and (iii) several metrics of tag relevance. In particular, we propose several new heuristic methods, which extend state-of-the-art strategies by including new metrics that try to capture how accurately a candidate term describes the object's content. We also exploit two learning-to-rank (L2R) techniques, namely RankSVM and Genetic Programming, for the task of generating ranking functions that combine multiple metrics to accurately estimate the relevance of a tag to a given object. We evaluate all proposed methods in various scenarios for three popular Web 2.0 applications, namely, LastFM, YouTube and YahooVideo. We found that our new heuristics greatly outperform the methods on which they are based, producing gains in precision of up to 181%, as well as another state-of-the-art technique, with improvements in precision of up to 40% over the best baseline in any scenario. Further improvements can also be achieved with the new L2R strategies, which have the additional advantage of being quite flexible and extensible to exploit other aspects of the tag recommendation problem.

We present GreenMeter, a tool for assessing the quality and recommending tags for Web 2.0 content. Its goal is to improve tag quality and the effectiveness of various information services (e.g., search, content recommendation) that rely on tags as data sources. We demonstrate an implementation of GreenMeter for the popular Last.fm application.

Understanding content popularity growth is of great importance to Internet service providers, content creators and online marketers. In this work, we characterize the growth patterns of video popularity on the currently most popular video sharing application, namely YouTube. Using newly provided data by the application, we analyze how the popularity of individual videos evolves since the video's upload time. Moreover, addressing a key aspect that has been mostly overlooked by previous work, we characterize the types of the referrers that most often attracted users to each video, aiming at shedding some light into the mechanisms (e.g., searching or external linking) that often drive users towards a video, and thus contribute to popularity growth. Our analyses are performed separately for three video datasets, namely, videos that appear in the YouTube top lists, videos removed from the system due to copyright violation, and videos selected according to random queries submitted to YouTube's search engine. Our results show that popularity growth patterns depend on the video dataset. In particular, copyright protected videos tend to get most of their views much earlier in their lifetimes, often exhibiting a popularity growth characterized by a viral epidemic-like propagation process. In contrast, videos in the top lists tend to experience sudden significant bursts of popularity. We also show that not only search but also other YouTube internal mechanisms play important roles to attract users to videos in all three datasets.

Parameter setting of Evolutionary Algorithms is a time consuming task with two main approaches: parameter tuning and parameter control. In this work we describe a new methodology for tuning parameters of Genetic Programming algorithms using factorial designs, one-factor designs and multiple linear regression. Our experiments show that factorial designs can be used to determine which parameters have the largest effect on the algorithm's performance. This way, parameter setting efforts can focus on them, largely reducing the parameter search space. Two classical GP problems were studied, with six parameters for the first problem and seven for the second. The results show the maximum tree depth as the parameter with the largest effect on both problems. A one-factor design was performed to fine-tune tree depth on the first problem and a multiple linear regression to fine-tune tree depth and number of generations on the second.

This work addresses the task of recommending high quality tags by exploiting not only previously assigned tags, but also terms extracted from other textual features (e.g., title and description) associated with the target object.To estimate the quality of a candidate tag recommendation, we use several metrics related to both tag co-occurrence and information quality. We also propose a heuristic function to combine the metrics to produce a final ranking of the recommended tags. We evaluate our heuristic function in various scenarios, for three popular Web 2.0 applications. Our experimental results indicate that our heuristic function significantly outperforms two state-of-the-art tag recommendation algorithms.

Services that offer users the possibility of transmitting their own live streaming video content, using Web 2.0-based platforms, are increasing in popularity. In this context, we propose and evaluate solutions that contribute to improve the scalability of this type of system. This work encompasses two major steps. First, we collected data from a popular online live video sharing service, and provide a characterization of key aspects of user dynamic behavioral patterns. Next, we used our characterization findings to drive the design and evaluation, via simulation, of alternative content distribution architectures. In particular, motivated by some of our findings, we combine the traditional Peer-to-Peer and client-server architectures into a new hybrid strategy, evaluating its cost-effectiveness in comparison with the two more conventional schemes. Simulation results, covering different metrics and scenarios, indicate that the hybrid strategy yields the best tradeoffs between quality of service and bandwidth requirements.

The growth of popularity of Web 2.0 applications greatly increased the amount of social media content available on the Internet. However, the unsupervised, user-oriented nature of this source of information, and thus, its potential lack of quality, have posed a challenge to information retrieval (IR) services. Previous work focuses mostly only on tags, although a consensus about its effectiveness as supporting information for IR services has not yet been reached. Moreover, other textual features of the Web 2.0 are generally overseen by previous research.

Bag-of-Visual-Features (BoVF) representations have achieved a great success when used for object recognition, mainly because of their robustness to several kinds of variations and occlusion. Recently, a number of BoVF approaches has been proposed also for recognition of human actions from videos. One important issue that arises when using BoVF for videos is how to take dynamic information into account, and most proposals rely on 3D extensions of 2D visual descriptors for this. However, we envision alternative approaches based on 2D descriptors applied to the spatio-temporal video planes, instead of to the traditionally explored by previous work. Thus, in this paper, we address the following question: what is the cost-effectiveness of a BoVF approach built from such 2D descriptors when compared to one based on the state-of-the-art 3D Spatio-Temporal Interest Points (STIPs) descriptor? We evaluate the recognition rate and time complexity of alternative 2D descriptors applied to different sets of spatio-temporal planes, and the state-of-the-art STIPs. Experimental results show that, with proper settings, 2D descriptors can yield the same recognition results as those provided by STIP, but at a significantly higher time complexity.

A number of online video social networks, out of which YouTube is the most popular, provides features that allow users to post a video as a response to a discussion topic. These features open opportunities for users to introduce polluted content, or simply pollution, into the system. For instance, spammers may post an unrelated video as response to a popular one aiming at increasing the likelihood of the response being viewed by a larger number of users. Moreover, opportunistic users--promoters--may try to gain visibility to a specific video by posting a large number of (potentially unrelated) responses to boost the rank of the responded video, making it appear in the top lists maintained by the system. Content pollution may jeopardize the trust of users on the system, thus compromising its success in promoting social interactions. In spite of that, the available literature is very limited in providing a deep understanding of this problem.

This article characterizes video-based interactions that emerge from YouTube's video response feature, which allows users to discuss themes and to provide reviews for products or places using much richer media than text. Based on crawled data covering a representative subset of videos and users, we present a characterization from two perspectives: the video response view and the interaction network view. In addition to providing valuable statistical models for various characteristics, our study uncovers typical user behavioral patterns in video-based environments and shows evidence of opportunistic behavior.

Online auction services have reached great popularity and revenue over the last years. A key component for this success is the seller. Few studies proposed analyzing how the seller and the auction configuration affect the negotiation results. In this work we propose a methodology to characterize online auctions by the seller's perspective. This methodology is based on: (1) recognizing the characteristics of the variables related to the auction results and (2) capturing the correlation among these variables to identify seller profiles and selling strategies. We applied our methodology to a real case study, using an eBay dataset, to validate two hypotheses about sellers and their practices. These results are useful to understand the complex mechanisms that guide ending prices, success (or failure), and the attraction of bids in online auctions, which can support decision strategies for buyers and sellers.

Peer-to-peer live streaming media systems are becoming more popular each day. As in file sharing P2P system, they are susceptible to content pollution attack. In this kind of attack, a peer alters the media content decreasing the perceived quality of the streaming. In this paper we evaluate the impact of pollution attack in P2P live streaming and we present two reputation system to avoid content polluted dissemination and isolate malicious peers. Our results show that a few number of polluters is capable to compromise all the application and the 2 proposed reputation systems can quickly identify and isolate polluters and also be resistant to peers collusion.

This paper seeks understanding the user behavior in a social network created essentially by video interactions. We present a characterization of a social network created by the video interactions among users on YouTube, a popular social networking video sharing system. Our results uncover typical user behavioral patterns as well as show evidences of anti-social behavior such as self-promotion and other types of content pollution.

An important issue in the design of a wireless sensor network (WSN) is to devise techniques to make efficient use of its energy, and thus, extend its lifetime. When two WSNs are deployed at the same place and their sensors cooperate with the other networks forwarding their packets, the distance of the transmissions decreases and, therefore, the power consumption as well. The goal of this work is to examine the extent to which different WSNs can cooperate and save their energy. Simulation results reveal that different densities and data collecting rates among WSNs, the routing algorithm and the path loss exponent have major impact in the establishment of cooperation.

Capacity management of a hosting infrastructure has traditionally focused only on performance goals. However, the quality of service provided to the hosted applications, and ultimately the revenues achieved by the provider, depend also on other aspects, such as security and energy constraints. This paper extends our self-adaptive SLA-driven capacity management solution to capture, in an unified framework, key performance and cost tradeoffs that arise when operating under security attacks and energy constraints. A number of scenarios and strategies based on dynamic SLA contracts are designed to help uncover, via simulation experiments, the main tradeoffs, considering both the provider’s interests (i.e., revenues) and the customer’s interests (i.e., legitimate throughput, response time distribution and costs). Finally, we also assess the cost-effectiveness of our framework under highly variable application service times.

This paper addresses the problem of hosting multiple applications on a provider's virtualized multi-tier infrastructure. Building from a previous model, we design a new self-adaptive capacity management framework, which combines a two-level SLA-driven pricing model, an optimization model and an analytical queuing-based performance model to maximize the provider's business objective. Our main contributions are the more accurate multi-queue performance model, which captures application specific bottlenecks and the parallelism inherent to multi-tier platforms, as well as the solution of the extended and much more complex optimization model. Our approach is evaluated via simulation with synthetic as well as realistic workloads, in various scenarios. The results show that our solution is significantly more cost-effective, in terms of the provider's achieved revenues, than the approach it is built upon, which uses a single-resource performance model. It also significantly outperforms a multi-tier static allocation strategy for heavy and unbalanced workloads. Finally, preliminary experiments assess the applicability of our framework to virtualized environments subjected to capacity variations caused by the processing of management and security-related tasks.

Broadband Internet pricing scheme should charge users based on their consumption pattern. It could avoid resource wasting because light users will not have to subsidize heavy users. Furthermore, a fair pricing scheme could save bandwidth during hours of day, which would improve broadband carriers capacity management and planning efforts. This paper presents a fair broadband pricing scheme where users' adaptive applications are configured to adjust their demand in accordance with several parameters, such as user's budget, estimated future needs, and usage price defined and published by the service provider. We simulate our pricing scheme and three others schemes using real data from a broadband Internet Service Provider (ISP) in order to analyze bandwidth saving and utility obtained by users. We show that in the proposed pricing scheme 15% of users have positive payoff, 1% have negative, and for the rest of the users the payoff is zero. We also present the re-distribution of the workload curves along the time-of-day and show that bandwidth saving happens in almost 60% of all hours of the day. We compare our pricing scheme with others reported in the literature and the simulation results show the potential of our proposal.

Network coding is an emerging field of research with sound and mature theory supporting it. Recent works shows that it has many benefits like improved fault tolerance, higher flexibility in selection of file parts to transfer and resiliency to network partitions [4, 3]. Despite those appealing properties there is no wide usage of network coding in real file sharing applications. In this work, we try to bridge the gap between theory of network coding and practice. From the one hand, we deploy one the most successful file sharing client, the BitTorrent client. We use the BitTorrent algorithm for optimizing the neighbor selections for maximizing the upload bandwidth. From the other hand, we propose several simple heuristics that improve significantly the efficiency of the network coding deployed. In a nutshell, we propose computation intensive variant of network coding that can be applied to most of the existing network coding protocols. By changing the random selection of coded parts to a selection based on feedback from the network, we significantly improve the network utilization and the efficiency of the protocol. In this paper we report our work in progress building the BitCod client. Using extensive simulations we demonstrate that our technique can compete with the performance of the state-of-the-art BitTorrent [2] file sharing client. Next, we plan to implement and test a prototype of the BitCod client over the WAN.

Pricing could be used to foster fair use of Internet resources and help to control backbone congestion. This paper presents a mechanism for pricing residential broadband Internet services based on the following information: user subscription, ISP backbone usage historic data, and estimated future consumption of each user. Our proposal combines concepts from flat rate pricing, usage based pricing, and time based pricing models. This is an ongoing work and we are currently working a simulation in order to evaluate the results obtained with the proposed pricing scheme.

This work considers the problem of hosting multiple third-party Internet services in a cost-effective manner so as to maximize a provider's business objective. For this purpose, we present a dynamic capacity management framework based on an optimization model, which links a cost model based on SLA contracts with an analytical queuing-based performance model, in an attempt to adapt the platform to changing capacity needs in real time. In addition, we propose a two-level SLA specification for different operation modes, namely, normal and surge, which allows for per-use service accounting with respect to requirements of throughput and tail distribution response time. The cost model proposed is based on penalties, incurred by the provider due to SLA violation, and rewards, received when the service level expectations are exceeded. Finally, we evaluate approximations for predicting the performance of the hosted services under two different scheduling disciplines, namely FCFS and processor sharing. Through simulation, we assess the effectiveness of the proposed approach as well as the level of accuracy resulting from the performance model approximations.

Recently, a number of scalable stream sharing protocols have been proposed with the promise of great reductions in the server and network bandwidth required for delivering popular media content. Although the scalability of these protocols has been evaluated mostly for sequential user accesses, a high degree of interactivity has been observed in the accesses to several real media servers. Moreover, some studies have indicated that user interactivity can severely penalize the scalability of stream sharing protocols.This paper investigates alternative mechanisms for scalable streaming to interactive users. We first identify a set of workload aspects that are determinant to the scalability of classes of streaming protocols. Using real workloads and a new interactive media workload generator, we build a rich set of realistic synthetic workloads. We evaluate Bandwidth Skimming and Patching, two state-of-the-art streaming protocols, covering, with our workloads, a larger region of the design space than previous work. Finally, we propose and evaluate five optimizations to Bandwidth Skimming, the most scalable of the two protocols. Our best optimization reduces the average server bandwidth required for interactive workloads in up to 54%, for unlimited client buffers, and 29%, if buffers are constrained to 25% of media size.

Mobile networks are becoming increasingly popular in enterprise environments as a means for distributing information to a large community of highly dynamic users. In comparison to traditional wired networks, mobile networks are distinguished by a potentially much higher variability in users demand due to user mobility. Most previous content distribution techniques assume a static user demand distribution and, thus, may not perform well in mobile networks. This paper proposes and analyzes a mobile dynamic content distribution network model, which takes demand variations into account to decide whether to replicate a content and whether to remove previously created replicas in order to minimize total network traffic. We develop two solutions to our model: an offline optimal solution, which provides an ideal lower bound on the total traffic, and a practical heuristic online algorithm, which uses demand forecasting to make replication decisions. We provide a thorough evaluation of our solutions, comparing them against ACDN, the only previous dynamic content placement algorithm targeting bandwidth minimization that we are aware of. Our results show that our online algorithm significantly outperforms ACDN, reducing total network traffic by up to 85% in a number of experiments covering a large system design space.

The rapid increase in the volume of unsolicited commercial e-mails, also known as spam, is beginning to take its toll in system administrators, business corporations and end-users. Widely varying estimates of the cost associated with spam are available in the literature. However, a quantitative analysis of the determinant characteristics of spam traffic is still an open problem. This work fills this gap and presents what we believe to be the first extensive characterization of a spam traffic.

Mobile networks are becoming increasingly popular as means for distributing inform tion to large number of users. In comparison to wired networks, mobile networks distinguished by potentially much higher variability in demand due to user mobility. Most previous content techniques ssume static client demand distribution and, thus, may not perform well in mobile networks.This paper proposes and analyzes Mobile Dynamic Content Distribution Network model, which takes demand variations into account to decide whether to replicate content and whether to remove previously created replicas in order to minimize total network traffic. We develop two solutions to our model: an offline optimal, which provides an ideal lower-bound on the total traffic, and practical heuristic online algorithm, which uses demand forecasting to make replication decisions. We provide thorough evaluation of our solutions, comparing them against ACDN, the only previous dynamic content placement algorithm targeting bandwidth minimization that we are aware of. Our results show that our online algorithm significantly outperforms the ACDN one, reducing total network traffic by up to 85% in a number of experiments covering large system design space.

This paper provides an extensive analysis of pre-stored streaming media workloads, focusing on the client interactive behavior. We analyze four workloads that fall into three different domains, namely, education, entertainment video and entertainment audio. Our main goals are: (a) to identify qualitative similarities and differences in the typical client behavior for the three workload classes and (b) to provide data for generating realistic synthetic workloads.

Recent scalable multicast streaming protocols for on-demand delivery of media content offer the promise of greatly reduced server and network bandwidth. However, a key unresolved issue is how to design scalable content distribution systems that place replica servers closer to various client populations and route client requests and response streams so as to minimize the total server and network delivery cost. This issue is significantly more complex than the design of distribution systems for traditional Web files or unicast on-demand streaming, for two reasons. First, closest server and shortest path routing does not minimize network bandwidth usage; instead, the optimal routing of client requests and server multicasts is complex and interdependent. Second, the server bandwidth usage increases with the number of replicas. Nevertheless, this paper shows that the complex replica placement and routing optimization problem, in its essential form, can be expressed fairly simply, and can be solved for example client populations and realistic network topologies. The solutions show that the optimal scalable system can differ significantly from the optimal system for conventional delivery. Furthermore, simple canonical networks are analyzed to develop insights into effective heuristics for near-optimal placement and routing. The proposed new heuristics can be used for designing large and heterogeneous systems that are of practical interest. For a number of example networks, the best heuristics produce systems with total delivery cost that is within 16% of optimality.

This paper develops simple cost models for provisioning content distribution networks that use the simple and highly scalable bandwidth skimming protocol for streaming. New insight is obtained into: (1) how cost-effective proxy servers are in multicast streaming systems; (2) the most effective streaming protocol; and (3) the optimal proxy content, as a function of the system configuration and workload. A key result is that proxy servers are only cost effective if: (a) the origin server does not have a multicast capability; or (b) the file request rate is low, and thus multicast is not highly effective; or (c) the cost of a proxy server stream is a very small fraction (i.e., approximately 1/P) of the cost of an origin server stream, where P is the number of proxy servers and the cost of either type of stream includes both the server and network resource costs. For cases where proxy servers are cost effective, results in the paper provide the optimal proxy content and the most effective streaming protocol, as a function of a wide range of system configuration and workload parameters. In contrast to previous work, full file caching outperforms prefix caching over a significant region of this system design space, due to more efficient multicast streaming protocols as well as a more complete exploration of the practical system configuration space.

This paper presents an extensive analysis of the client workloads for educational media servers at two major U.S. universities. The goals of the analysis include providing data for generating synthetic workloads, gaining insight into the design of streaming content distribution networks, and quantifying how much server bandwidth can be saved in interactive educational environments by using recently developed multicast streaming methods for stored content.

The sharing of caches among Web proxies is an important technique to reduce Web traffic and alleviate network bottlenecks. Nevertheless it is not widely deployed due to the overhead of existing protocols. In this paper we demonstrate the benefits of cache sharing, measure the overhead of the existing protocols, and propose a new protocol called "summary cache". In this new protocol, each proxy keeps a summary of the cache directory of each participating proxy, and checks these summaries for potential hits before sending any queries. Two factors contribute to our protocol's low overhead: the summaries are updated only periodically, and the directory representations are very economical, as low as 8 bits per entry. Using trace-driven simulations and a prototype implementation, we show that, compared to existing protocols such as the Internet cache protocol (ICP), summary cache reduces the number of intercache protocol messages by a factor of 25 to 60, reduces the bandwidth consumption by over 50%, eliminates 30% to 95% of the protocol CPU overhead, all while maintaining almost the same cache hit ratios as ICP. Hence summary cache scales to a large number of proxies. (This paper is a revision of Fan et al. 1998; we add more data and analysis in this version.).

The sharing of caches among Web proxies is an important technique to reduce Web traffic and alleviate network bottlenecks. Nevertheless it is not widely deployed due to the overhead of existing protocols. In this paper we propose a new protocol called "Summary Cache"; each proxy keeps a summary of the URLs of cached documents of each participating proxy and checks these summaries for potential hits before sending any queries. Two factors contribute to the low overhead: the summaries are updated only periodically, and the summary representations are economical --- as low as 8 bits per entry. Using trace-driven simulations and a prototype implementation, we show that compared to the existing Internet Cache Protocol (ICP), Summary Cache reduces the number of inter-cache messages by a factor of 25 to 60, reduces the bandwidth consumption by over 50%, and eliminates between 30% to 95% of the CPU overhead, while at the same time maintaining almost the same hit ratio as ICP. Hence Summary Cache enables cache sharing among a large number of proxies.