We address the question of compressing, into a fixed amount of storage space, a periodically sampled sequence of scalar values. The length of the sequence is, a-priori, unknown. The compression logic targets the minimization of the L∞ reconstruction error. Because of the limited storage, all relevant compression processing is performed in–place. The proposed scheme involves a synthesis of an early (“batch”) phase, and a subsequent “incremental” (online) phase. The technique is suitable for wireless sensor networks with limited storage and with unpredictable, and possibly rare, opportunities to communicate their data. We demonstrate that our scheme outperforms legacy sub-sampling techniques and that its reconstruction error gracefully degrades as the length of the data sequence increases. We show how a couple kilobytes of storage is sufficient to represent long sequences of sampled data, making it suitable for microcontroller-based wireless sensor nodes.

In this work, we characterize the potential information leakage from IoT platforms during their setup phase. Setup involves an IoT device, its ''app'', and a cloud-based service. We assume that the on-device firmware is inaccessible, e.g., read-protected. We focus on the combination of information that can be extracted from analyzing the app and the local communication between the app and the IoT device. An attacker can trivially obtain the app, analyze its operation, and potentially eavesdrop on the wireless communication occurring during the setup phase. We develop a semi-automated general methodology involving off-the-shelf tools to examine information disclosure during the setup phase. We tested our methodology on twenty commodity-grade IoT devices. The outcome reveals a wide range of device-dependent choices for encryption at various layers and the potential for exposure of, among other things, device-identifying information and local networking (WiFi) credentials. Our methodology contributes towards a means to assess and ''certify'' IoT devices.

With the growing adoption of electric vehicles (EVs) and advent of bidirectional chargers, EV aggregators, such as charging stations, will become a major player in electricity markets, providing voltage regulation (VR) or other services. We present a novel and practical VR scheme that takes advantage of the charging flexibility of EVs in charging stations that are connected to buses in a distribution grid. This VR scheme relies on real-time measurements, as well as estimates of the distribution system state and regulation capacity of each charging station. We then propose a novel false data injection attack against the VR capacity estimation process that exploits the uncertainty in EV mobility and network conditions. We show the attack vector with the largest expected adverse impact is the solution of a stochastic optimization problem, subject to a constraint that ensures it bypasses bad data detection. We determine this attack vector by solving a sequence of convex quadratically constrained linear programs. The case studies examined in a cosimulation platform, based on two standard test feeders, reveal the vulnerability of the VR capacity estimation process.

We consider the behavior of mobile users that, upon depletion of their device's energy due to communication, can detour from their regular path, in order to reach a location where the device can be recharged. We develop abstractions of this recharging behavior and analytically derive a charging-aware mobility model. The device is viewed as a mobile node. Based on the Palm inversion formula, we derive the (integral-form) stationary probability density function of the node's location, subject to attraction exerted by one charger in a bounded convex area. The analytical and numerical results demonstrate the distortion effect the charger has on the spatial node distribution. We observe, and explain, a counter-intuitive effect whereby the node density exhibits peaks that are asymmetric, coupled with a relative decrease of the density around the charger area. In addition, we provide a method to approximate the probability density function when multiple chargers are deployed. We examine the accuracy of this approximation and comment on its limitations.

We consider the problem of lossy compression of time series data collected by wireless sensor nodes, such that it produces a limited volume of compressed data for a given amount of raw data. The lossy compression is performed in a manner that minimizes the resulting L∞reconstruction error. Producing a bounded volume of compressed data is desirable in circumstances where we wish to know, or limit, the rate at which the compressed data are communicated, e.g., for periodic communication scheduling of fixed payload transmissions. The work is also geared to understanding the impact of, and accommodating for, storage limitations of Wireless Sensor Network (WSN) nodes. The proposed scheme belongs to the class of piecewise linear approximations (PLAs) and its performance is compared to other PLA schemes proposed for WSNs. The evaluation is carried out using existing public data sets.

Across different Wi-Fi devices, there exist differences in the probing behavior during active scanning. We conjecture that the behavior is sufficiently distinct to identify individual device types. We propose a feature engineering strategy to training machine learning algorithms for determination of the device type. We propose a concurrent capture across multiple Wi-Fi channels, thus allowing the features to include attributes for the transitions happening between channels during active scanning. Small-scale proof-of-concept results provide encouraging results about the method's potential.

The growing adoption of Distributed Energy Re-sources (DER) in low-voltage distribution grids calls for new feedback control algorithms that rely on quasi-real-time data collected by remote sensors. The design and evaluation of such algorithms necessitates a prudent and comprehensive approach since these algorithms require a tight integration of power and communication systems. A simple link failure or a sophisticated cyberattack launched against the grid's monitoring, communication, and control infrastructure could rapidly grow out of control, making the grid unstable. We investigate the design and implementation of a high-fidelity smart grid simulation platform which integrates a network simulator and a power flow simulator using the Mosaik co-simulation framework. The platform allows for evaluating the performance of new control algorithms and understanding dynamics of modern distribution grids. Example case studies are presented to validate the proposed platform.

Today's mobile battery-powered communication devices require that users access chargers via wired and, recently, wireless recharging facilities. For a device departing from a location with a given energy “budget”, a plausible strategy is to seek a charger location once the energy is exhausted. We present a model of mobile nodes that captures the paths followed by the nodes with depleted energy seeking, possibly via a detour, to reach a charger. The derived location-dependent mobile node density distribution is used to express the location-dependent congestion of a wireless network whose capacity is used by the mobile nodes. The boundaries, and the relative placement of the charger, create intriguing discontinuities in the probability density function of the nodes across space. We find that chargers are not always “hotspots” in terms of node density, and that the energy budget of the nodes at the beginning of their trip impacts that density and determines the hotspot. Moreover, upon energy depletion, to detour from one's path or not, results in a distinctly different relative impact of the charger placement on the ability of nodes to sustain communication, depending, again on the energy budget the nodes have upon departure from a waypoint.

We present a model to predict the short-term utilization of an IEEE 802.11 channel. We approximate the time-varying utilization process via a Markovian state transition model and subsequently create a lumped representation of the transition matrix. Each lumped state can then be treated as a class. The lumped matrix provides a simpler to understand description of the channel utilization behavior and naturally includes the persistence in one lumped state which resembles the characteristic behavior of naive predictors (where predicted state equals the current state). We demonstrate that treating the lumped states as classes allows good prediction models to be built using Logistic Regression and Neural Network models. Our results are based on IEEE 802.11 wireless utilization data collected as reported in the channel utilization (CU) field of the QBSS Load Element in Beacon frames. The presented approach can be implemented as an edge computing task, whereby edge nodes calculate the lumped states and train models, informing nearby client devices of the model parameters, allowing them to produce predictions on their own.

We consider the problem of loss of continuous data feeds from sensor networks, due to transient failures. Because the failures are recoverable, part of the missing data may be, eventually, acquired. Even then, the limited resources of the nodes can result in an incomplete reconstruction of the missing data. In this paper we study a set of proposed data imputation methods, and their variations, on a real data set. We determine the tradeoffs involved in the proposed techniques. A common characteristic of the studied techniques is that they depend on the recent behavior of the data stream and do not make specific assumptions about the long-term stochastic behavior of the data. We consider also the case where simple, sub-sampling based, handling of accumulated missing data is implemented by the nodes.

Received Signal Strength Indication (RSSI) is a notoriously noisy metric, yet attractive for localization purposes. We study the effect of reducing the RSSI value to a single bit, essentially turning the RSSI measurement into a proximity indicator. We consider systems where a device, that needs to be localized, transmits and the transmission is received by a number of receivers placed at fixed and known locations. We consider two modes of 1-bit quantization: one (global) where the quantization is defined uniformly across all signal receivers, and one (local) where the RSSI is quantized separately for each receiver. We compare the effects of 1-bit quantization across three profiling k-NN-based localization algorithms, comparing them with each other in addition to comparing against results when no quantization is performed. We furthermore consider the case of global 2-bit quantization. Our study is based on profiling data collected before and after modifications were performed to the profiled environment, allowing us to also study the impact of those changes. Our results show comparable performance between no quantization and local 1-bit quantization, and similar results for global 2-bit quantization.

We study tradeoffs between aggregation convergecast and energy harvesting in wireless sensor networks. Existing aggregation convergecast algorithms do not capture the volatile nature of energy reserves of energy harvesting nodes. We therefore propose, and evaluate, new scheduling schemes to address this gap. We also introduce metrics to capture the impact of the inevitable energy depletion on the quantity of aggregated data received at a sink node. Specifically, we consider node behaviors where the inability to perform prompt communication due to energy depletion results in a reduction of the sampling rate (including for aggregated data) and, if it persists, loss of data. The performance evaluation is based on heat flow data collected from an apartment building in the Canadian North. The collected heat flow data are used to approximate the energy harvesting output of thermoelectric harvesters.

We consider the problem of increasing the data collection frequency of aggregation convergecast. Previous studies attempt to increase the data collection frequency by shortening the completion of a single data collection cycle. We aim at increasing the frequency at which data collection updates are collected by the use of pipelining and, consequently, increasing the overall data collection frequency and throughput. To achieve this, we overlap the propagation schedule of multiple data snapshots within the same overall schedule cycle, thus increasing parallelism through pipelining. Consequently, the effective data collection time of an individual snapshot may span over multiple, successive, schedule cycles. To this end, we modify the aggregation convergecast model, decoupling schedule length, and data collection delay, by relaxing its precedence constraints. Our solution for this new problem involves the unconventional approach of constructing the schedule before finalizing the exact form of the data aggregation tree, which, in turn, requires that the schedule construction phase guarantees that every node can reach the sink. We compare our results using snapshot pipelining against a previously proposed algorithm that also uses a form of pipelining, as well as against an algorithm that though lacking pipelining, exhibits the ability to produce very short schedules. The results confirm the potential to achieve a substantial throughput increase, at the cost of some increase in latency.

Recognizing the occupants movement and locations within a home is a basic functionality, underlying a variety of smart-home services, including energy management, ambient environment control, and assistive-living services for seniors and people with disabilities. The outdoor-localization variant of the problem is effectively addressed with the use of GPS; however, GPS does not work well inside buildings, which makes the indoor positioning problem a very active research topic. In this paper, we report on a study of the indoor-localization problem, relying on easy-to-deploy, inexpensive, BLE-enabled stickers and beacons and WiFi access points.

Received Signal Strength Indication (RSSI) measurements are known to be noisy and to exhibit high degree of variability, yet massive sets of RSSI data are streamed and collected with the purpose of using them for localization. In this paper we determine that the full resolution of RSSI measurements is unnecessary and propose, based on an evaluation of indoor localization experiments, a technique of modest computational effort, using Genetic Algorithms, to determine how to best quantize RSSI measurements. Our approach treats the localization algorithm as a “black box” and, hence, its applicability is broad. However, for the purpose of exposition and to generate absolute performance metrics, a particular profiling-based localization algorithm is used. Results from experiments involving real collected RSSI values indicate that it is possible to reduce the RSSI data volume by approximately 72% with no noticeable reduction in localization accuracy.

Motivated by the computational, bandwidth and energy restrictions of wireless sensor network nodes and their need to, collectively, determine the presence of exogenous interference that could impair their communication, we consider schemes that could support the task of interference classification as a first step towards interference mitigation strategies. In particular, we examine the effectiveness of the Discrete Wavelet Transform (DWT) to communicate to other nodes the state of the channel, as sampled by a node, in a compressed, denoised form. We examine the suitability of different wavelet filters and thresholding methods in order to: (a) preserve key features of the interference, (b) denoise the noisy interference samples, and (c) reduce the amount of information that needs to be communicated to describe the interference.

We consider multi-hop wireless networks composed of nodes with transceivers capable of multi-packet transmission and reception (MPT/MPR). Legacy MAC protocols based on CSMA/CA are overly restrictive in the interest of avoiding collisions, and are unable to exploit the MPR capability of receivers. We demonstrate how a combination of mechanisms, based on well-known techniques, such as Additive Increase Multiplicative Decrease (AIMD), and the back--pressure (BP) principle, can be used to effectively control medium access in multi-hop MPT/MPR networks. The AIMD component is used to regulate the size of "bundles" of simultaneously transmitted packets, while back--pressure provides the basis for prioritizing, locally, which flows' packets should be transmitted in a bundle. We study the performance of the proposed protocol, AB-MAC, under three different models of node coordination in static wireless multi-hop MPT/MPR networks. We find that, under various scenarios and for the same capacity resources, AB-MAC's throughput performance surpasses that of IEEE 802.11b.

The Smart-Condo is a hardware/software platform that aims to support and assist an individual in performing a variety of everyday tasks within his/her living space. The key to achieving this goal is being able to recognize the individual's general activities in real-time, without impeding these activities or compromising privacy. Since location and movement constitute meaningful evidence for many everyday tasks (e.g., presence in the bathroom correlates with personal hygiene activities), we are motivated to develop an efficient, accurate, and noninvasive occupant-localization method. To this end, we propose a methodology for planning the deployment of an array of privacy-respecting binary motion sensors. In particular, given the geometric constraints of the deployment space, we generate a model of indoor mobility patterns typical for a single occupant. We then use this model as the basis for a specific optimization problem: maximizing a measure of how well the frequently-visited areas of the living space are covered by a number of sensors, subject to a cardinality constraint on this number. We argue this optimization objective is a good surrogate for maximizing localization accuracy, and prove that it bears exploitable properties that make it receptive to a simple optimization routine. As a result, we obtain sensor configurations with localization accuracy superior to that achievable with the same number of sensors placed manually or randomly in the same environment.

Due to recent progress in signal processing techniques, the traditional collision channel model is inadequate for communication networks where the transceivers are endowed with the ability to transmit or receive multiple packets simultaneously (the MPT/MPR capability). This paper studies the schedule construction in MPT/MPR systems with the objective to understand how the flows traversing the network can leverage the MPT/MPR capabilities in multi-hop wireless networks. Towards our goal, we present a heuristic algorithm MDSatur to produce a schedule and propose the wireless water-filling (WF) algorithm which extends the traditional water-filling algorithm to multi-hop wireless scenarios to compute max-min allocations when the MPT and MPR capabilities are equal. By combining MDSatur and WF, we also design the LEX scheme to approximate the lexicographically optimal allocation of the system when the network has nonidentical MPT and MPR capabilities.

We consider the problem of scheduling wireless transmissions in a sensor network to perform aggregation convergecast. In contrast to studies that attempt to shorten the schedule for completing a single data collection cycle, we aim to increase the frequency at which updates are collected (higher throughput). To achieve higher throughput, we use a form of pipelined concurrent collection of multiple data snapshots through the network. To attain high performance pipelining, we “expand” the time in which precedence constraints need to be satisfied such that they span over multiple schedule cycles. Our approach involves the unconventional approach of constructing the schedule before finalizing the exact form of the precedence constraints, i.e., before determining the data aggregation tree, which in turn requires that the schedule construction phase guarantees that every node can reach the sink. We compare our results using pipelining against a previously proposed algorithm that also uses pipelining, as well as against an algorithm that, although lacking pipelining, exhibits the ability to produce very short schedules. The results confirm the potential to achieve a substantial throughput increase at the cost of increased latency.

Medication adherence is an important challenge for many patients with chronic conditions, most of them elderly. Technology has an important role to play in this area potentially, with electronic devices equipped with reminder capability and medication intake recording. In this paper, we present a remotely programmable pillbox. This pillbox is equipped with a web application which gives the health professional or caregiver a tool to check and program the pill-box. Also, a mobile application is implemented to establish a connection with the web-application to show pills' daily schedule and pill taking notifications.

The low-powered transmissions in a wireless sensor network (WSN) are highly susceptible to interference from external sources. Our work is a step towards enabling WSN devices to better understand the interference in their environment so that they can adapt to it and communicate more efficiently. We extend our previous work in which we collected received signal strength traces using mote-class synchronized receivers at sample rates that are, to the best of our knowledge, higher than previously described in the literature. These traces contain distinct interference patterns, each with a different potential for being exploited by cognitive radio strategies. In order to exploit a pattern, however, a node must first recognize it. Given the energy and space constraints of a node, we explore succinct decision tree classifiers for the two most disruptive patterns. We expand on a basic feature set to incorporate attributes based on the dip statistic and the Lomb periodogram, both of which address specific, empirically observed behaviour, and we show their positive impact on both the decision tree structure and the overall classification performance. Moreover, we present an approximation of the periodogram that makes its construction feasible for mote-class devices, and we describe the simplification's impact on classification performance.

Broadcasting is an elementary problem in wireless networks. Energy - efficient broadcasting is important, e.g., to coordinate the distributed computing operations by sending periodic messages in a network of Automatic Identification System installed on energy constrained maritime lighthouses. To that end logical tree topologies that are based on Connected Dominating Sets have been proposed vigorously in the literature. In this paper we present Biased Shortest Path Tree (BISPT), a new logical tree topology for efficient broadcasting in wireless networks. In simulations we find that BISPT outperforms state-of-the-art solutions.

We consider the transmission scheduling problem for aggregation convergecast in wireless sensor networks. All transmissions required for a completed data collection to the sink node are to be scheduled in a single TDMA frame. Concurrent transmissions are possible within each TDMA slot as long as no collisions occur at the intended recipients of each transmission. We observe that previously proposed heuristics to solving the aggregation convergecast scheduling problem rely on a decomposition into two phases. The first phase is the construction of an aggregation tree, typically a Shortest Path Tree (SPT), and the second phase is scheduling transmissions. We express the constraints of the aggregation convergecast problem in a manner compatible with a constraint programming solver, and extract solutions for small sized instances. The, admittedly sparse, evidence suggests that the resultant aggregation trees is almost never SPTs and point to alternative considerations that would be useful when constructing aggregation convergecast scheduling heuristics.

In this paper, we consider the exact top-k query problem in wireless sensor networks, i.e., where one seeks to find the k highest reported values as well as the complete set of nodes that reported them. Our primary contribution in this context is EXTOK, a provably correct and topology-independent new filtering-based algorithm for processing exact top-k queries. As a secondary contribution we confirm a previous result of ours by showing that the efficiency of top-k query processing algorithms, including EXTOK, can be further improved by simply choosing a proper underlying logical tree topology. We examine EXTOK's performance with respect to a number of parameters and different logical tree topologies while using both synthetic and real data sets. Our simulation reveal that EXTOK consistently outperforms the current state-of-the-art algorithm by a very significant margin and regardless of the underlying logical tree topology.

We present a beacon-based clustering algorithm aimed at prolonging the cluster lifetime in VANETs. We use a new aggregate local mobility criterion to decide upon cluster re-organisation. The scheme incorporates a contention method to avoid triggering frequent re-organisations when two clusterheads encounter each other for a short period of time. Simulation results show a significant improvement of cluster lifetime and reduced node state/role changes compared to previous popular clustering algorithms.

We propose a dynamic scheduler for multichannel 802.15.3 WPANs which exploits both channel reuse and interference control. A maximal-weighted-matching based scheduling scheme is designed to arrange simultaneous transmissions in each single channel and a multichannel assignment scheme is proposed to obtain the ultimate schedule in each superframe. For successful signal decoding, accumulated interference among simultaneous transmissions is taken into account to ensure that signal reception occurs with sufficient SINR. To improve the efficiency of channel reuse, the scheduler fragments each transmission request to multiple requests with the equal duration. An analysis is given of the optimal way to implement such process according to online traffic load and overhead. Simulations are used to evaluate the performance and to compare against existing schemes.

We propose an energy-efficient multi-hop scheduling scheme for single-channel 802.15.3 networks which produces interference-avoiding schedules that are throughput and energy efficient. The scheduler replaces high-interference peer-to-peer transmission with low-interference multi-hop transmissions by interference-aware multi-hop routing, which reduces both interference and energy consumption of transmission. To maximize channel reuse, a multi-hop maximal-weighted-matching grouping scheme is applied to arrange concurrent transmissions. For successful signal decoding, the accumulated interference among concurrent transmissions is taken into account to ensure that signal reception occurs with sufficient SINR.

We consider the interplay between data aggregation and workload balancing in a Wireless Sensor Network (WSN). In particular, we formulate the problem of deciding where to forward traffic, in order to benefit the most from potential aggregation, while also splitting the resulting traffic (once aggregated) on the way to the Sink for the sake of load balancing. Load balancing across multiple paths results in a corresponding balanced energy consumption and therefore prolongs the lifetime of WSNs. The optimization formulation leads to an IP/LP model where the objective is that of maximizing the lifetime of the WSN. Heuristics are proposed for assigning an efficient data aggregation node for each sensor source. Simulation results show that our schemes significantly improve performance both in terms of longer system lifetime and less total data received (indicating more aggressive aggregation) at the Sink.

We propose a scheduling scheme for single-channel 802.15.3 networks which produces interference-avoiding schedules for both isochronous and asynchronous traffic. The scheduler explores spatial channel reuse by means of maximal weighted matching and request grouping. For fairness, a binary exponential weight adjustment algorithm is applied to dynamically adjust the priority of requests. To enhance scheduling efficiency, requests are split and channel time is allocated accordingly. An efficient schedule update mechanism is also introduced to handle on-the-fly traffic request modifications.

We consider the problem of extending the lifetime of wireless sensor networks (WSNs) where sensors report their data to a base station via multi-hop transmission. In-network data aggregation is introduced to reduce the traffic. The lifetime is optimized using a LP (linear programming) framework built on a multicommodity network flow model to which we introduce the concept of flow loss multiplier to express the impact of data aggregation over correlated data. To balance energy consumption across the network, data aggregation is performed only at the first hop (FH) transmission. Heuristics are proposed to obtain significant FH aggregation, accordingly prolonging system lifetime. Simulation results show that FH aggregation possesses characteristics that make it suitable under certain data collection applications/scenarios

We consider the problem of transmission scheduling in wireless networks. A good transmission scheduling scheme should not only be conflict-free, in the sense of resolving collisions and the hidden terminal problem, but should also provide fairness and transmission efficiency to each node. We propose a class of scheduling schemes that rely on the logical ordering of the schedule update operations. The ordering is imposed by the precedence relationship of vertices in a directed acyclic graph (DAG). Two examples of establishing the logical DAG are given: (a) one relies on a distributed spanning tree construction; (b) one is based on location information. A separate mechanism ensures that nodes do not behave in a greedy fashion and monopolize the schedule slots. We prove the correctness of DAG-based schemes and provide simulation results that demonstrate their fairness and efficiency attributes.

We consider the problem of anonymous file exchange in wireless ad-hoc networks. We propose a new protocol, dubbed RASH (roam and share), to address the problem. RASH is a peer-to-peer (P2P) protocol which fulfills three objectives: (a) protection of user privacy (understood as the confidentiality of their interests), (b) support for a high degree of file authentication and integrity (to reduce or eliminate false downloads and spoofs), and (c) enforcement of reciprocity (to remove freeloaders from the system). The environment of our scheme is purely ad-hoc, in that it does not assume pre-established trust, fixed-infrastructure, or the existence of specialized nodes

Cooperation based on reputation--based trust schemes in mobile ad hoc networks has been proposed with the intention of securing networks against possible selfish behavior. Selfishness is defined as refusal of (non--malicious) nodes to participate in network activities such as packet forwarding. We analyze the effectiveness of one such cooperation enforcement mechanism, namely reputation--based cooperation, using a simulated ad hoc network environment. The energy consumption of selfish nodes is also analyzed, to identify whether selfish behavior is actually beneficial to a node, and thus, whether selfish behavior is something that might actually be observed in a real ad hoc network. We find that reputation--based cooperation enforcement is only effective in non--mobile ad hoc network environment, however, we also find evidence that selfish behavior may not have benefits when the cost of energy consumption of a node is considered.

The objective of increasing capacity in mobile ad-hoc networks (MANETs) by exploiting the mobility of nodes has produced interesting but rather unrealistic results. The throughput improvement is not followed by reasonable delay performance, and in realistic settings there is a need to compensate for the potentially low mobility of all, or parts of, the network. We explore the throughput/delay tradeoff by attempting to utilize all possible service (transmission) opportunities that may exist. To this end, a measure of redundancy is introduced and its impact characterized. The definition of redundancy admits both the replication of data items as well as the traversal of paths. We study three schemes, where the third is a new proposal for a hybrid scheme that balances off a limited form of routing with relayed packet propagation. We find that the hybrid is resilient to the impact of mobility, providing good performance even at low mobility settings.

We study classification schemes for TCP flows based on (a) lifetime (short vs. long lived flows) and (b) round-trip-time (RTT) attributes, as well as their combinations. The double objective served by the classification schemes is to satisfy the need for reduced response time, which is the primary concern of short lived flows, while at the same time ensuring fairness among long lived flows. The presented results indicate that the combined classification performs, under all tested case, comparably or better than RED and, as such, it is a reasonable alternative to the cumbersome task of properly parameterizing RED. In fact, the presented combined classification schemes achieve results equal or better than simple RED even when used in conjunction with DropTail queue management.

We present an active queue management policy and we study its ability to steer, in a distributed fashion, an entire TCP/IP network into global max-min fairness of its long-lived flows. A previously proposed technique, initially applicable to circuit-switched networks and feedback-based flow control networks is adapted to the realities of packet-switching and TCP flow/congestion control, together with a per-hop policy, called FairShare. The proposed solution is evaluated using several example topologies, including linear, loop, and irregular topologies as well as dynamic load fluctuations.

We study a collection of K-shortest path routing schemes and investigate their performance under a diverse set of network topologies and traffic conditions. We subsequently demonstrate that K-shortest path routing offers a lower blocking probability and more balanced link utilisation than other routing methods. With the proposed approach, it is possible to reduce the frequency of link state exchange, and the incurred bandwidth overhead, without sacrificing the overall performance of the network.

We study the construction of the source-initiated (one-to-all) wireless broadcast tree to minimize the total required power for a given source node, a group of intended destination nodes and a given propagation constant, ie, the power attenuation constant /spl lambda/. The minimum energy broadcasting (MEB) problem has received much attention recently due to the two main challenges of mobile communication: the limited bandwidth of wireless networks and the limited power supply of mobile units. In a limited-bandwidth environment, push-based techniques, ie, broadcast schemes, appear to be a very effective way to allow mobile units to share the broadcast data on air. In a limited-energy environment, energy- efficient communication architectures and techniques are essential. We first give an insight analysis on the MEB problem and prove the NP-hardness of this problem. We then present an efficient heuristic called iterative maximum-branch minimization (IMBM) to approximate the construction of the minimum-energy broadcast tree, which fully utilizes the wireless broadcast advantage and demonstrates better performance compared with the related approaches. Due to the power-efficient way of the construction of the broadcast tree, the lifetime of the networks can be maximized.

No abstract available.

A protocol for the reliable multicast of data to mobile hosts is described. A logical ring is maintained between all the base stations that handle the multicast traffic of the same multicast group. A token passing protocol enforces a consistent view between all base stations with respect to the frames that are considered delivered to all mobiles. The interaction of the reliable multicasting and the handoff events of the mobiles is controlled by a special handoff protocol. A description of the management protocol is also given. We use simulation results to illustrate the feasibility of the approach and reveal the performance interplay between the base station buffer size and the token rotation frequency.

Several proposals for video-on-demand (VoD) systems assume batching of the user requests and subsequent multicasting of the selected videos. We propose that the waiting time, which is essential for the sake of batching, can be overlapped with a more elaborate call admission scheme. The call admission scheme provides sufficient bandwidth at any point in time, in order to achieve guaranteed delivery of the video stream without any buffering or additional delays. The call admission is based on the construction of deterministic time-dependent envelopes. Essentially, statistical multiplexing is replaced by deterministic multiplexing. Bandwidth efficiency is maintained through the particular envelope construction which minimizes the amount of over-allocated bandwidth. The construction of the envelopes can be performed using either a computationally intensive exact algorithm or by a fast heuristic. Both the envelope construction and the call admission procedure are presented and simulated in detail. The results indicate the feasibility, benefits and tradeoffs of the proposed scheme.

In order to address the scalability problems of video-on-demand systems, several periodic broadcast schemes have been proposed that partition a video into segments and repetitively broadcast each segment on a separate channel. A new scheme is presented for the bandwidth-efficient periodic broadcast of video. The proposed scheme determines the segment sizes and their corresponding channel bandwidths as a result of a non-linear optimization problem which minimizes the total required bandwidth for the broadcast. The new scheme outperforms the existing schemes in terms of bandwidth demands while it also decouples the playout latency from the number of available channels. Further analysis reveals that its asymptotic bandwidth requirements exactly match the asymptotic bandwidth requirements reported for poly-harmonic broadcasting.

We present a technique to enable the efficient simulation of large scale ATM networks, while preserving the accuracy of the end-to-end delay statistics. Our approach uses on-the-fly aggregation by observing traffic at monitoring points, then substituting an aggregate model, when appropriate. We focus in particular on the end-to-end delay distribution, given the importance of this distribution for continuous media. We find that our methods are able to achieve speedup of one order of magnitude, while maintaining accuracy within 5% of the unaggregated simulation. These results are observed in both a series of multiplexers and a 100-switch wide-area ATM topology.

A serious fairness problem occurs whenever homogeneous, variable bit-rate, video traffic streams are multiplexed at a finite buffer statistical multiplexer. Namely, the cell loss ratio performance varies widely across individual traffic streams. The cause of the problem is the particular random relation of the periodic frame transmission epochs of the video sources that feed the multiplexer. In this paper a protocol is presented which controls the frame transmission epochs, in order to achieve fair distribution of losses among the admitted connections. The protocol operates with the participation of the sources by imposing an additional per-source delay on the traffic entering the network. These additional delays are calculated whenever a connection is accepted or terminated. The overall additional delay imposed by the protocol is guaranteed to be minimal through the use of an appropriate optimization algorithm. The performance of the proposed protocol, operating in conjunction with the optimization algorithm, is simulated and evaluated. A basic framework is given for its implementation on an ATM network. Finally, the issues of buffer overhead and the impact of the protocol on the delay experienced at the multiplexer are also discussed.

The multiplexing of several lightly loaded links onto a more heavily loaded output link is a problem of considerable importance to the design and traffic engineering of many types of packet-oriented telecommunications equipment, including that used in Asynchronous Transfer Mode (ATM) networks. Network configurations generally require the cascaded operation of such multiplexers and switches. Important objectives to achieve small cell loss ratios while maintaining efficient utilization of the transmission links. The small cell loss ratio objective results in extremely long simulation runs. To address this problem, we propose a new technique that relies on a compact description for the arriving/departing traffic at the multiplexers and a time-parallel scheme without fix-up phases for effective parallelization. The technique does not make assumptions about the analytical nature of the arrival process, thereby allowing trace-driven simulations to be performed as well. We demonstrate the method for a number of configurations and traffic scenarios, and observe that it yields one to two orders of magnitude speedup on a 32 processor Kendall Square Research KSR-1 multiprocessor compared to an efficient cell-level simulation executing on a Sparc-10 workstation.