Search

Gab is an online social network often associated with the alt-right political movement and users barred from other networks. It presents an interesting opportunity for research because near-complete data is available from day one of the network's creation. In this paper, we investigate the evolution of the user interaction graph, that is the graph where a link represents a user interacting with another user at a given time. We view this graph both at different times and at different timescales. The latter is achieved by using sliding windows on the graph which gives a novel perspective on social network data. The Gab network is relatively slowly growing over the period of months but subject to large bursts of arrivals over hours and days. We identify plausible events that are of interest to the Gab community associated with the most obvious such bursts. The network is characterised by interactions between `strangers' rather than by reinforcing links between `friends'. Gab usage follows the diurnal cycle of the predominantly US and Europe based users. At off-peak hours the Gab interaction network fragments into sub-networks with absolutely no interaction between them. A small group of users are highly influential across larger timescales, but a substantial number of users gain influence for short periods of time. Temporal analysis at different timescales gives new insights above and beyond what could be found on static graphs., 10 pages, 9 figures

Getting good statistical models of traffic on network links is a well-known, often-studied problem. A lot of attention has been given to correlation patterns and flow duration. The distribution of the amount of traffic per unit time is an equally important but less studied problem. We study a large number of traffic traces from many different networks including academic, commercial and residential networks using state-of-the-art statistical techniques. We show that traffic obeys the log-normal distribution which is a better fit than the Gaussian distribution commonly claimed in the literature. We also investigate an alternative heavy-tailed distribution (the Weibull) and show that its performance is better than Gaussian but worse than log-normal. We examine anomalous traces which exhibit a poor fit for all distributions tried and show that this is often due to traffic outages or links that hit maximum capacity. We demonstrate that the data we look at is stationary if we consider samples of 15- minute long or even 1-hour long. This gives confidence that we can use the distributions for estimation and modelling purposes. We demonstrate the utility of our findings in two contexts: predicting that the proportion of time traffic will exceed a given level (for service level agreement or link capacity estimation) and predicting 95th percentile pricing. We also show that the log-normal distribution is a better predictor than Gaussian or Weibull distributions in both contexts.

Temporal graphs capture the development of relationships within data throughout time. This model fits naturally within a streaming architecture, where new events can be inserted directly into the graph upon arrival from a data source and be compared to related entities or historical state. However, the vast majority of graph processing systems only consider traditional graph analysis on static data, with some outliers supporting batched updating and temporal analysis across graph snapshots. In this work we define a temporal graph model which can be updated via event streams and discuss the challenges of distribution and graph maintenance. To solve these challenges, we introduce Raphtory, a distributed temporal graph management system which maintains the full graph history in memory, leveraging this to insert streamed events directly into the model without batching or centralised ordering. Raphtory additionally provides an API to perform both approximative analysis on the most up-to-date version of the graph, as well as temporal analysis throughout its full history; executed in parallel with ingestion.

Sensitive inferences and user re-identification are major threats to privacy when raw sensor data from wearable or portable devices are shared with cloud-assisted applications. To mitigate these threats, we propose mechanisms to transform sensor data before sharing them with applications running on users' devices. These transformations aim at eliminating patterns that can be used for user re-identification or for inferring potentially sensitive activities, while introducing a minor utility loss for the target application (or task). We show that, on gesture and activity recognition tasks, we can prevent inference of potentially sensitive activities while keeping the reduction in recognition accuracy of non-sensitive activities to less than 5 percentage points. We also show that we can reduce the accuracy of user re-identification and of the potential inference of gender to the level of a random guess, while keeping the accuracy of activity recognition comparable to that obtained on the original data., Accepted to appear in Pervasive and Mobile computing (PMC) Journal, Elsevier

TCP remains the dominant transport protocol for Internet traffic. It is usually considered to have its sending rate covered by a sliding window congestion control mechanism. However, in addition to this normal congestion control, a number of other mechanisms limit TCP throughput. This paper analyses the extent to which network, host and application settings define flow throughput over time and across autonomous systems. Our study draws on data from a longitudinal study spanning five years of passive traces collected from a single transit link. Mechanisms for this include limiting by application, interference with the TCP window control mechanism and artificial limitations on maximum window sizes by the operating system. This paper uses a large data set to assess the impact of each mechanism. We conclude that more than half of all heavy-hitter inbound traffic remains throttled by constraints beyond network capacity. For this data set, TCP congestion control is no longer the dominant mechanism that moderates throughput.

Cloud computing has revolutionized service delivery by providing on-demand invocation and elasticity. To reap these benefits, computation has been displaced from client devices and into data centers. This partial centralization is undesirable for applications that have stringent locality requirements, e.g., low latency. This problem could be addressed with large numbers of smaller cloud resources closer to users. However, as cloud computing diffuses from within data centers and into the network, there will be a need for cloud resource allocation algorithms that operate on resource-constrained computational units that serve localized subsets of customers. In this paper, we present a mechanism for service provisioning in distributed clouds where applications compete for resources. The mechanism operates by enabling execution zones to assign resources based on Vickrey auctions and provides high-quality probabilistic models that applications can use to predict the outcomes of such auctions. This allows applications to use knowledge of the locality distribution of their clients to accurately select the number of bids to be sent to each execution zone and their value. The proposed mechanism is highly scalable, efficient, and validated by extensive simulations.

Getting good statistical models of traffic on network links is a well-known, often-studied problem. A lot of attention has been given to correlation patterns and flow duration. The distribution of the amount of traffic per unit time is an equally important but less studied problem. We study a large number of traffic traces from many different networks including academic, commercial and residential networks using state-of-the-art statistical techniques. We show that the log-normal distribution is a better fit than the Gaussian distribution commonly claimed in the literature. We also investigate a second heavy-tailed distribution (the Weibull) and show that its performance is better than Gaussian but worse than log-normal. We examine anomalous traces which are a poor fit for all distributions tried and show that this is often due to traffic outages or links that hit maximum capacity. We demonstrate the utility of the log-normal distribution in two contexts: predicting the proportion of time traffic will exceed a given level (for service level agreement or link capacity estimation) and predicting 95th percentile pricing. We also show the log-normal distribution is a better predictor than Gaussian or Weibull distributions., Accepted for INFOCOM 2019

Discriminating between competing explanatory models as to which is more likely responsible for the growth of a network is a problem of fundamental importance for network science. The rules governing this growth are attributed to mechanisms such as preferential attachment and triangle closure, with a wealth of explanatory models based on these. These models are deliberately simple, commonly with the network growing according to a constant mechanism for its lifetime, to allow for analytical results. We use a likelihood-based framework on artificial data where the network model changes at a known point in time and demonstrate that we can recover the change point from analysis of the network. We then use real datasets and demonstrate how our framework can show the changing importance of network growth mechanisms over time.

There is growing concern about how personal data are used when users grant applications direct access to the sensors of their mobile devices. In fact, high resolution temporal data generated by motion sensors reflect directly the activities of a user and indirectly physical and demographic attributes. In this paper, we propose a feature learning architecture for mobile devices that provides flexible and negotiable privacy-preserving sensor data transmission by appropriately transforming raw sensor data. The objective is to move from the current binary setting of granting or not permission to an application, toward a model that allows users to grant each application permission over a limited range of inferences according to the provided services. The internal structure of each component of the proposed architecture can be flexibly changed and the trade-off between privacy and utility can be negotiated between the constraints of the user and the underlying application. We validated the proposed architecture in an activity recognition application using two real-world datasets, with the objective of recognizing an activity without disclosing gender as an example of private information. Results show that the proposed framework maintains the usefulness of the transformed data for activity recognition, with an average loss of only around three percentage points, while reducing the possibility of gender classification to around 50\%, the target random guess, from more than 90\% when using raw sensor data. We also present and distribute MotionSense, a new dataset for activity and attribute recognition collected from motion sensors., 6 pages

Motion sensors such as accelerometers and gyroscopes measure the instant acceleration and rotation of a device, in three dimensions. Raw data streams from motion sensors embedded in portable and wearable devices may reveal private information about users without their awareness. For example, motion data might disclose the weight or gender of a user, or enable their re-identification. To address this problem, we propose an on-device transformation of sensor data to be shared for specific applications, such as monitoring selected daily activities, without revealing information that enables user identification. We formulate the anonymization problem using an information-theoretic approach and propose a new multi-objective loss function for training deep autoencoders. This loss function helps minimizing user-identity information as well as data distortion to preserve the application-specific utility. The training process regulates the encoder to disregard user-identifiable patterns and tunes the decoder to shape the output independently of users in the training set. The trained autoencoder can be deployed on a mobile or wearable device to anonymize sensor data even for users who are not included in the training dataset. Data from 24 users transformed by the proposed anonymizing autoencoder lead to a promising trade-off between utility and privacy, with an accuracy for activity recognition above 92% and an accuracy for user identification below 7%., Comment: 10 pages

Temporal graphs capture the relationships within data as they develop throughout time. Intuition, therefore, suggests that this model would fit naturally within a streaming architecture, where new points of comparison can be inserted directly into the graph as they arrive from the data source. However, the current state of the art has yet to join these two concepts, supporting either temporal analysis on static data or streaming into one-dimensional dynamic graphs. To solve this problem we introduce Raphtory, a temporal graph streaming platform, which maintains a full graph history whilst efficiently inserting new alterations.

Multi-user video conferencing is a ubiquitous technology. Increasingly end-hosts in a conference are assisted by cloud-based servers that improve the quality of experience for end users. This paper evaluates the impact of strategies for placement of such servers on user experience and deployment cost. We consider scenarios based upon the Amazon EC2 infrastructure as well as future scenarios in which cloud instances can be located at a larger number of possible sites across the planet. We compare a number of possible strategies for choosing which cloud locations should host services and how traffic should route through them. Our study is driven by real data to create demand scenarios with realistic geographical user distributions and diurnal behaviour. We conclude that on the EC2 infrastructure a well chosen static selection of servers performs well but as more cloud locations are available a dynamic choice of servers becomes important.

An increasing number of sensors on mobile, Internet of things (IoT), and wearable devices generate time-series measurements of physical activities. Though access to the sensory data is critical to the success of many beneficial applications such as health monitoring or activity recognition, a wide range of potentially sensitive information about the individuals can also be discovered through access to sensory data and this cannot easily be protected using traditional privacy approaches. In this paper, we propose a privacy-preserving sensing framework for managing access to time-series data in order to provide utility while protecting individuals' privacy. We introduce Replacement AutoEncoder, a novel algorithm which learns how to transform discriminative features of data that correspond to sensitive inferences, into some features that have been more observed in non-sensitive inferences, to protect users' privacy. This efficiency is achieved by defining a user-customized objective function for deep autoencoders. Our replacement method will not only eliminate the possibility of recognizing sensitive inferences, it also eliminates the possibility of detecting the occurrence of them. That is the main weakness of other approaches such as filtering or randomization. We evaluate the efficacy of the algorithm with an activity recognition task in a multi-sensing environment using extensive experiments on three benchmark datasets. We show that it can retain the recognition accuracy of state-of-the-art techniques while simultaneously preserving the privacy of sensitive information. Finally, we utilize the GANs for detecting the occurrence of replacement, after releasing data, and show that this can be done only if the adversarial network is trained on the users' original data., Comment: 12 pages, 11 figures

Researchers have examined crowd behavior in the past by employing a variety of methods including ethnographic studies, computer vision techniques and manual annotation based data analysis. However, because of the resources to collect, process and analyze data, it remains difficult to obtain large data sets for study. In an attempt to alleviate this problem, researchers have recently used mobile sensing, however this technique is currently only able to detect either stationary or moving crowds with questionable accuracy. In this work we present a system for detecting stationary interactions inside crowds using the Received Signal Strength Indicator of Bluetooth Smart (BLE) sensor, combined with the Motion Activity of each device. By utilizing Apple's iBeacon™ implementation of Bluetooth Smart, we are able to detect the proximity of users carrying a smartphone in their pocket. We then use an algorithm based on graph theory to predict interactions inside the crowd and verify our findings using video footage as ground truth. Our approach is particularly beneficial to the design and implementation of crowd behavior analytics, design of influence strategies, and algorithms for crowd reconfiguration.

This paper proposes a new decentralised multi-channel tree building protocol with a centralised controller for the Internet of Things. The protocol alleviates the effect of interference which results in improved network efficiency and stability, and link reliability. The proposal takes into account all available channels to utilise the spectrum and aims to use the spectrum efficiently by transmitting on several channels. The protocol detects which channels suffer interference and changes away from those channels. The algorithm for channel selection is a two-hop colouring protocol that reduces the chances of nearby nodes to transmit on the same channel. All nodes are battery operated except for the low power border router (LPBR). This enables a centralised channel switching process at the LPBR. The protocol is built based on the routing protocol for low power and lossy networks (RPL). In its initial phase, the protocol uses RPL's standard topology formation to create an initial working topology and then seeks to improve this topology by switching channels. The implementation and evaluation of the protocol is performed using the Contiki framework. The experimental results demonstrate an increased resilience to interference and significantly higher throughput making better use of the total available spectrum and link stability.

This paper deals with the problem of assessing probabilistic models that represent the evolution of a target graph. Such models have long been a topic of interest for a number of networks, especially communications networks. The solution developed in this paper gives a rigorous way to calculate the likelihood of the observed graph evolution having arisen from a wide variety of hypothesized models encompassing many already present in the literature. The framework is shown to recover parameters from artificial data and is tested on real data sets from Facebook and from emails from the company Enron.

This paper considers fundamental measurements which drive TCP flows: throughput, RTT and loss. It is clear that throughput is, in some sense, a function of both RTT and loss. In their seminal paper Padyhe et al [1] begin with a mathematical model of the TCP sliding window evolution process and come up with an equation showing that TCP throughput is (roughly) proportional to 1/RTT√p where p is the probability of packet loss. Their equation is shown to be consistent with data gathered on several links. This paper takes the opposite approach and analyses a large number of packet traces from well-known sources in order to create a data-driven estimate of the functions which relate TCP, loss and RTT. Regression analysis is used to fit models to connect the quantities. The fitted models show different behaviour from that expected in [1].

The production of a large-scale monitoring system for a high-speed network leads to a number of challenges. These challenges are not purely techinical but also socio-political and legal. The number of stakeholders in a such a monitoring activity is large including the network operators, the users, the equipment manufacturers and of course the monitoring researchers. The MASTS project (Measurement at All Scales in Time and Space) was created to instrument the high-speed JANET Lightpath network, and has been extended to incorporate other paths supported by JANET(UK). Challenges the project has faced have included: simple access to the network; legal issues involved in the storage and dissemination of the captured information, which may be personal; the volume of data captured and the rate at which this data appears at store. To this end the MASTS system will have established four monitoring points each capturing packets on a high speed link. Traffic header data will be continuously collected, anonymised, indexed, stored and made available to the research community. A legal framework for the capture and storage of network measurement data has been developed which allows the anonymised IP traces to be used for research pur poses., 21 pages, 4 figures

The increasing demand for efficient content distribution using the Internet has fuelled the deployment of varied techniques such as peer-to-peer overlays, content distribution networks and distributed caching systems. These have had considerable impact on ISP infrastructure demand, motivating the development of protocols that enable mutually beneficial cooperative outcomes. In this paper we propose a parameterised cooperation utility that can be used to study the tradeoff between the benefit that an overlay obtains from the ISPs that carry its traffic and the costs that it imposes on them. With this utility, we find a closed-form expression for the optimal resource allocation given a particular cooperation tradeoff, subject to both minimal benefit and maximal cost constraints. Since the model is implementation-independent and has very modest computational demands, it is ideal for large scale simulation. We explore the properties of the proposed model through simulation in both a simple illustrative scenario and a more complete one based on network datasets. The results obtained from the model are shown to be consistent with those of measurement-based studies of overlay-ISP collaboration.

Networking Named Content (NNC) was recently proposed as a new networking paradigm to realise Content Centric Networks (CCNs). The new paradigm changes much about the current Internet, from security and content naming and resolution, to caching at routers, and new flow models. In this paper, we study the caching part of the proposed networking paradigm in isolation from the rest of the suggested features. In CCNs, every router caches packets of content and reuses those that are still in the cache, when subsequently requested. It is this caching feature of CCNs that we model and evaluate in this paper. Our modelling proceeds both analytically and by simulation. Initially, we develop a mathematical model for a single router, based on continuous time Markov-chains, which assesses the proportion of time a given piece of content is cached. This model is extended to multiple routers with some simple approximations. The mathematical model is complemented by simulations which look at the caching dynamics, at the packet-level, in isolation from the rest of the flow.

There has long been a need for a robust and reliable system which distributes traffic across multiple paths. In particular such a system must rarely reorder packets, must not require per-flow state, must cope with different paths having different bandwidths and must be self-tuning in a variety of network contexts. PREFLEX, proposed herein, uses estimates of loss rate to balance congestion. This paper describes a method of automatically adjusting how PREFLEX will split traffic in order to balance loss across multiple paths in a variety of network conditions. Equations are derived for the automatic tuning of the time scale and traffic split at a decision point. The algorithms described allow the load balancer to self-tune to network conditions. The calculations are simple and do not place a large burden on a router which would implement the algorithm. The algorithm is evaluated by simulation using ns-3 and is shown to perform well under a variety of circumstances. The resulting adaptive, end-to-end traffic balancing architecture provides the necessary framework to meet the increasing demands of users while simultaneously offering edge networks more fine-grained control at far shorter timescales.