Abstract Machine learning models are inherently bound to the distribution of the training data, often exploiting non-causal shortcuts. As a result, achieving robustness to spurious correlations remains a challenge. While existing approaches rely on data manipulation or re-weighting strategies to achieve robustness, they typically require dense group labels, multiple training domains, or specialized pre-processing. We propose Invariance Pair Guidance (IPG), a method to mitigate reliance on spurious correlations using a sparse set of counterfactual pairs. Unlike other methods demanding extensive supervision, IPG utilizes a novel dual-update mechanism to dynamically correct the optimization trajectory. We generate input pairs that isolate the spurious attribute to define the invariance, a characteristic that should not affect the outcome of the model. Based on these pairs, we define a corrective gradient that complements the traditional gradient descent approach. The correction adapts via a predefined invariance condition. Experiments on ColoredMNIST, Waterbirds-100, and CelebA datasets demonstrate the effectiveness of our approach and its robustness to group shifts, supported by a theoretical convergence analysis. IPG offers a data-efficient and theoretically grounded path to robustness.

Cyber-Physical Systems (CPS) implement critical infrastructures, in which physical objects interact with services of the cyber domain consequently building a heterogeneous System-of-Systems (SoS). Although this marriage extends the functionality of traditionally closed systems, it also introduces a variety of challenges—especially for engineers. One critical aspect relates to establishing and sustaining a sufficient level of security, as exploited vulnerabilities may cause severe effects, either toward involved humans or sensitive information. Consequently, engineers must be aware of potential security-related threats and vulnerabilities. To this end, threat models are usually used to identify such weaknesses within a specification. However, existing solutions may not be able to comprehensively identify all threats in a CPS-like environment, as they often do not consider all relevant dependencies between interacting systems on an SoS level. To address this gap, we have elaborated a methodology—called Semantic Threat Model (STM), which can identify and evaluate potential threats toward a given CPS specification. In detail, the framework focuses on the semantic relationships and side effects between security objects, e.g., attacks, and the actual specification of the CPS. In contrast to existing solutions, STM takes an SoS point of view, while analyzing semantic data to gain a comprehensive view on security. The quantitative output of the method can then be used to identify the most severe attacks or to point out necessary security enhancements. We highlight the usage and benefits of the STM in the form of a case study in the domain of intelligent transportation systems.

Message Queuing Telemetry Transport (MQTT) is an established data transmission protocol in virtually all Smart-* domains. The protocol is based on the publish/subscribe pattern. Messages are published to so-called topics, to which subscribers subscribe, in order to receive the content of interest. As the name of the topic can be chosen freely by the publishing entity without relying on a shared semantic model, the subscribers can insufficiently reason on the semantic of the topic name. However, this can be desirable when topic trees closely resemble structures found in the real world. One approach to formulate semantics in a machine-readable format are ontologies. This paper demonstrates how ontologies allow to couple topic trees with explicit semantics. To this end we present a novel editor that supports users in creating topic trees that follow explicit semantics.

Bandwidth is a valuable and scarce resource in mobile networks. Therefore, bandwidth reservation may become necessary to support time-sensitive and safety-critical networked vehicular applications such as autonomous driving. Such applications require individual and deterministic approaches for reservations. This is challenging as vehicles usually have insufficient information to reason about future driving paths as well as future network resources availability and costs. In particular, the optimal time for a vehicle to place a cost-efficient reservation request is crucial. If a reservation is conducted too early, the uncertainty in path prediction may become high resulting in frequent cancellations with high costs. If a reservation is requested too late, resources may no longer be available. In this paper, we study the optimal timing for a given vehicle to place a bandwidth reservation request for an upcoming trip. Our proposal is based on predicting bandwidth costs using well-selected temporal machine learning techniques while achieving high accuracy levels. The proposed reservation scheme relies on a corpus of real-world traffic data. The experimental results prove that the model can effectively learn to find an optimized timing for bandwidth reservation. In addition, our model may allow vehicles to save considerably costs compared to the baseline of an immediate reservation scheme.

We consider the wireless network utility maximization problem for various vehicular applications with different priorities. Our approach jointly optimizes bandwidth reservation and immediate slicing allocation based on priorities. The basestation-vehicle association happens in a two-tier heterogeneous wireless network. We first determine a sub-optimal network bandwidth reservation ratio and slicing allocation ratio based on sub-optimal basestation association. To this end, we formulate the problem as a biconcave maximization problem which we solve using an alternative concave search algorithm. Intensive simulations demonstrate that the proposed priority-based reservation and slicing allocation scheme outperforms the conventional slicing allocation schemes in terms of efficiency in network resource utilization.

The importance of Cyber-Physical Systems (CPS) gains more and more weight in our daily business and private life. Although CPS build the backbone for major trends, like Industry 4.0 and connected vehicles, they also propose many new challenges. One major challenge can be found in achieving a high level of security within such highly connected environments, in which an unpredictable number of heterogeneous systems with often-distinctive characteristics interact with each other. In order to develop high-level security solutions, system designers must eventually know the current level of security of their specification. To this end, security metrics and scoring frameworks are essential, as they quantitatively express security of a given design or system. However, existing solutions may not be able to handle the proposed challenges of CPS, as they mainly focus on one particular system and one specific attack. Therefore, we aim to elaborate a security scoring mechanism, which can efficiently be used in CPS, while considering all essential information. We break down each system within the CPS into its core functional blocks and analyze a variety of attacks in terms of exploitability, scalability of attacks, as well as potential harm to targeted assets. With this approach, we get an overall assessment of security for the whole CPS, as it integrates the security-state of all interacting systems. This allows handling the presented complexity in CPS in a more efficient way, than existing solutions.

The coupling of safety-relevant embedded- and cyber-space components to build Cyber-Physical Systems (CPS) extends the functionality and quality in many business domains, while also creating new ones. Prime examples like Internet of Things and Industry 4.0 enable new technologies and extend the service capabilities of physical entities by building a universe of connected devices. In addition to higher complexity, the coupling of these heterogeneous systems results in many new challenges, which should be addressed by engineers and administrators. Here, security represents a major challenge, which may be well addressed in cyber-space engineering, but less in embedded system or CPS design. Although model-based engineering provides significant benefits for system architects, like reducing complexity and automated analysis, as well as being considered as standard methodology in embedded systems design, the aspect of security may not have had a major role in traditional engineering concepts. Especially the characteristics of CPS, as well as the coupling of safety-relevant (physical) components with high-scalable entities of the cyber-space domain have an enormous impact on the overall level of security, based on the introduced side effects and uncertainties. Therefore, we aim to define a model-based security-engineering framework, which is tailored to the needs of CPS engineers. Hereby, we focus on the actual modeling process, the evaluation of security, as well as quantitatively expressing security of a deployed CPS. Overall and in contrast to other approaches, we shift the engineering concepts on a semantic level, which allows to address the proposed challenges in CPS in the most efficient way.

Crowdsourcing offers great opportunities to recognise user context and prescribe relevant services for both offline and real-time activities. In this work, we present a zoning model that leverages spatio-temporal dimensions and then employs different contexts to recommend necessary customised services. The context model takes into consideration three context sets: fully restricted, fully unrestricted and semi-restricted with respect to both spatial and temporal dimensions. As a proof of concept, we apply this zoning model in a scenario where a very large crowd get together to perform spatio-temporal activities. The user context of the heterogeneous crowd is captured using the carried smartphones, i.e. via crowdsourcing. Depending on the context sets and zone, the system can recommend a set of services to each user. The system has been deployed since 2014 to support the spatio-temporal activities of a very large crowd. We present our implementation details and the user feedback, which is very encouraging.

Wireless Sensor Networks (WSNs), being power and communication capacity constrained, often employ data compression schemes to reduce the data volumes. At the same time, various factors, such as low node reliability and communication faults, can compromise the core objective of accurate collection and delivery of the sensor data. However, contemporary compression schemes often fail to consider operational faults, and the resulting data errors arising from low node reliability (node crashes), and communication faults (data and links corruption) result in erroneous data being collected. This paper proposes a novel model-based fault-tolerance technique applicable to varied compression schemes. Being fully distributed, our scheme corrects data errors at the point of origin (faulty sensor nodes), and thus avoids costly transmissions of corrupt data, decreasing message cost, and enhances compression effectiveness by correcting the erroneous samples.

Wireless Sensor Networks (WSNs) are often deployed to sample the desired environmental attributes and deliver the acquired samples to a central station, termed as the sink, for processing as needed by the application. Many applications stipulate high granularity and data accuracy that results in high data volumes. However, sensor nodes are battery powered, and sending the requested large amounts of data rapidly depletes their energy. Fortunately, environmental attributes (e.g., temperature, pressure) often exhibit spatial and temporal correlations. Moreover, a large class of applications such as scientific analysis and simulations tolerate high latency for sensor data collection. Hence, we exploit the spatiotemporal correlation of sensor readings while benefiting from possible data delivery latency tolerance to minimize the amount of data to be transported to the sink. Accordingly, we develop a fully distributed adaptive hybrid compression scheme that exploits both spatial and temporal data redundancies and fuses both temporal and spatial compression for maximal data compression with accuracy guarantees. We present two main contributions: (i) an adaptive modeling technique that allows frugal and maximized temporal compression on resource-constraint sensor nodes by exploiting the data collection latency, and (ii) a novel model-based hierarchical clustering technique that allows for maximized spatial compression resulting into a hybrid compression scheme. Compared to the existing spatiotemporal compression approaches, our approach is fully decentralized and the proposed clustering scheme is based on sensor data models rather than instantaneous sensor data values, which allows merging nearby nodes with similar models into large clusters over a longer period of time rather than specific time instances. The analysis for computation and message overheads, the analysis for theoretical compressibility, and simulations using real-world data demonstrate that our proposed scheme can provide significant communication/energy savings without sacrificing the accuracy of collected data.

We propose a context aware framework that offers a set of cloud-based services to support a very large Hajj and Umrah crowd by capturing their contexts using smartphones. The proposed framework captures the individual's context, provides a set of adapted services, and allows being in touch with a subset of one's community of interest. We leverage the spatiotemporal sensory data captured by our framework to define users' contexts for optimized services. Our proposed framework is also envisioned to assist the Hajj and Umrah authorities to (1) improve Hajj & Umrah documentation, (2) improve Hajj organization through better understanding of pilgrims' (individual and crowd) spatial and temporal behavior and needs, and (3) protect pilgrims' environment through environmental monitoring. In particular, the developed methods, techniques, and algorithms will support the pilgrimage quality of experience. We have tested our system through end-user subjects and due to apply for the upcoming Hajj events. We present our implementation details and the general impression of end users about our system.

The reliable delivery of data is important in designing Wireless Body Area Networks (WBANs) employed for critical applications such as e-health. In order to communicate the data reliably from the sensors to the base station, the data transmission technique (star or multi-hop) and the transmission power plays a very important role. As transmission power is increased, transmission distance is increased and the data can be sent reliably to far nodes. However, in WBANs, there is always a limit to increase the transmission power. Keeping the power level at some low threshold and increasing the distance between a sensor and the base station results in reduced received power which ultimately degrades the data transmission. Thus, for star data transmission technique, the point to ponder is the maximum separation between a sensor and the base station to transmit the data reliably. The reliability in WBANs can be analyzed through different parameters such as received power, received signal strength indicator, link quality indicator, packet error rate, packet reception rate, etc. This paper aims at performing a reliability analysis for WBAN through the mentioned parameters to suggest an optimal sensor/base station separation using star data transmission technique. This analysis is performed employing the default routing protocol in TinyOS called the Collection Tree Protocol. Our study considers different sensor placements on different parts of the body as well as different angular offsets between sensor and base station.

ABSTRACT Because the Internet of Things (IoT) pervasively extends to all facets of life, the “things” are increasingly extending to include the interconnection of the Internet to critical infrastructures (CIs) such as telecommunication, power grid, transportation, e‐commerce systems, and others. The objective of this paper is twofold: (i) addressing IoT from a CI protection (CIP) and connectivity viewpoint, and (ii) highlighting the need for security quantification to improve the quality of protection (QoP) of CIs. Using a financial infrastructure as an example, a CIP and trust quantification perspective is built up. To this end, we are developing a novel security metrics‐based approach to assess and thereon enhance the CIP. We focus on the communication level of the CI where IoT is playing an increasingly important role with respect to sensing and communication across CI elements. Determining the security and dependability level of the communication over the CI constitutes a basic precondition for assessing the QoP of the whole CI, which is needed for any efforts to improve this QoP. Because metrics play a central role for such quantification, this paper develops their QoP use from an IoT perspective, and a reference implementation along with experimental results is presented. Copyright © 2012 John Wiley & Sons, Ltd.

Multiple databases and repositories exist for collecting known vulnerabilities for different systems and on different levels. However, it is not unusual that extensive time elapses, in some cases more than a year, in order to collect all information needed to perform/publish vulnerability scoring calculations for security management groups to assess and prioritize vulnerabilities for remediation. Scoring a vulnerability also requires broad knowledge about its characteristics, which is not always provided. As an alternative, this paper targets the quantitative understanding of security vulnerabilities in the context of insufficient vulnerability information. We propose a novel approach for the predictive assessment of security vulnerabilities, taking into consideration the relevant scenarios, e.g., zero day vulnerabilities, for which there is limited or no information to perform a typical vulnerability scoring. We propose a new analytical model, the Vulnerability Assessment Model (VAM), which is inspired by the Linear Discriminant Analysis and uses publicly available vulnerability databases such as the National Vulnerability Database (NVD) as a training data set. To demonstrate the applicability of our approach, we have developed a publicly available web application, the VAM Calculator. The experimental results obtained using real-world vulnerability data from the three most widely used Internet browsers show that by reducing the amount of required vulnerability information by around 50%, we can maintain the misclassification rate at approximately 5%.

In emergency health cases such as mass casualty incidents, the death ratio is still high due to lack of an automatic and intelligent system which timely observes and reports patient criticality. Indeed, the existing criticality assessment approaches are manual such as the established Simple Triage and Rapid Treatment (START). Accordingly, it is difficult for care givers to provide optimal healthcare, in particular, if the number of casualties outnumbers the responders. A challenge is how to automatically tag a possibly large number of victims with various types of disorders immediately after an incident and before the arrival of the paramedics. Such an automated tagging would provide for more optimized emergency response. We propose an automatic self-tagging methodology using body sensor networks that deliver relevant vital signs, i.e., respiratory rate, heart rate and mental status. We present three approaches to recognize and grade the criticality level of patients. The proposed approaches are generic and can be easily adapted to different scenario such as patients in intensive care units, patients in surgery and elderlies being monitored in their home. Being fully automated, our methodology is able to provide real-time tagging with higher accuracy and fine-granularity than the simplistic manual current systems. We demonstrate the viability of our self-tagging approaches by statistically demonstrating their accuracy compared to that of experts manual tagging.

Mobile environments increasingly require distributed atomic transactions to support the growing diversity of financial, gaming, social networking and many other applications. The underlying mobile infrastructure is correspondingly evolving with increasingly diverse wired and wireless elements and also with increasing exposure to a variety of operational perturbations at the mobile elements and communication levels. Consequently, the challenge is not only in providing efficient nonblocking mobile commit (as a fundamental basis behind consistent mobile transactions) but to also provide efficient perturbation-resilient atomic commit in the heterogeneous mobile space. The contribution of this paper is in developing a perturbation-resilient mobile commit protocol that efficiently provides for and preserves strict atomicity for transactional applications. The protocol does not necessarily require access to the powerful communication/computation elements of the wired infrastructure during transaction execution. However, in case access to a wired network becomes possible, it then adapts to utilize this to 1) increase the resilience to network perturbations achieving higher commit rates, and 2) reduce the wireless message overhead and the blocking of transaction participants leading to higher transactions throughput. In contrast, existing solutions are often tailored either for 1) infrastructure-based mobile environments, or 2) infrastructure-less ad hoc networks. To our knowledge, there is no existing commit protocol that can adapt across diverse infrastructure communication modes. The proposed perturbation-resilient generalized mobile transaction commit (GMTC) protocol represents the first atomic commit protocol for hybrid mobile environments which 1) takes advantage of accessing infrastructures, by choosing reliable infrastructure nodes for coordination of transactions and for replication of commit data of mobile participants to tolerate network disconnections, and 2) tolerates network partitioning and delivers best-effort resultsâin terms of transaction commit rate, message complexity, and commit/abort decision time (latency)âif the access to wired infrastructure is unavailable. The protocol performance simulations (covering transaction commit rate, message complexity, and commit/abort decision time) demonstrate the effectiveness of the developed protocol in generalized mobile environments.

Compressive Sampling (CS) is a powerful sampling technique that allows accurately reconstructing a compressible signal from a few random linear measurements. CS theory has applications in sensory systems where acquiring individual samples is either expensive or infeasible. A Wireless Sensor Network (WSN) is a distributed sensory system comprised of resource-limited sensor nodes. Transferring all the recorded samples in a WSN can easily result in data traffic that can exceed the network capacity. There are ongoing attempts to devise efficient and accurate compression schemes for WSNs and CS has proved to be a key sampling method compared to many other existing techniques. In this paper, specifically targeting the dominant WSN deployments of multi-hop WSNs, we develop a novel CS-based concept of sampling window as an efficient spatio-temporal signal acquisition/compression technique. We show that much higher energy-efficient signal acquisition is possible, if composite temporal and spatial correlations are considered. Our model is also capable of abnormal event detection which is a crucial feature in WSNs. It guarantees balanced energy consumption by the sensor nodes in a multi-hop topology to prevent overloaded nodes and network partitioning.

Peer-to-Peer (P2P) protocols are susceptible to Localized Eclipse Attacks (LEA), i.e., attacks where a victim peer's environment is masked by malicious peers which are then able to instigate progressively insidious security attacks. To obtain effective placement of malicious peers, LEAs significantly benefit from overlay topology-awareness. Hence, we propose heuristics for Chord, Pastry and Kademlia to assess the protocols' LEA susceptibility based on their topology characteristics and overlay routing mechanisms. As a result, our method can be used for P2P protocol parameter tuning in order to substantially mitigate LEAs. We present evaluations highlighting LEA's impact on contemporary P2P protocols. Our proposed heuristics are abstract in nature, making them applicable plus customizable for many other structured P2P protocols. We validate our model's accuracy through a simulation case study.

Map-based visualization of live spatial sensor data collected from a Wireless Sensor Networks (WSN) is a promising application. This application requires samples from each sensor node. The specific WSN properties such as energy imply that the data collection and/or in-network processing should carefully address the trade-off between the energy/network overhead and the map accuracy. Accordingly, several map construction approaches have been proposed. Their efficiency/accuracy balance has been proved in the corresponding papers for a set of carefully selected parameters. To date there exists no quantitative comparative study that considers a wide range of protocol and network parameters in order to provide a comprehensive evaluation of existing strategies. In this paper, we present the first study to evaluate the comparative effectiveness of the available map construction approaches.

Wireless Sensor Networks (WSN) are often deployed to sample the desired environmental attributes and deliver the acquired samples to the sink for processing, analysis or simulations as per the application needs. Many applications stipulate high granularity and data accuracy that results in high data volumes. Sensor nodes are battery powered and sending the requested large amount of data rapidly depletes their energy. Fortunately, the environmental attributes (e.g., temperature, pressure) often exhibit spatial and temporal correlations. Moreover, a large class of applications such as scientific measurement and forensics tolerate high latencies for sensor data collection. Accordingly, we develop a fully distributed adaptive technique for spatial and temporal in-network data compression with accuracy guarantees. We exploit the spatio-temporal correlation of sensor readings while benefiting from possible data delivery latency tolerance to further minimize the amount of data to be transported to the sink. Using real data, we demonstrate that our proposed scheme can provide significant communication/energy savings without sacrificing the accuracy of collected data. In our simulations, we achieved data compression of up to 95% on the raw data requiring around 5% of the original data to be transported to the sink.

The physical number of sensor nodes constitutes a major cost factor for Wireless Sensor Networks (WSN) deployments. Hence, a natural goal is to minimize the number of sensor nodes to be deployed, while still maintaining the desired properties of the WSN. However, sparse networks even while connected, usually suffer from topology irregularities that negatively impact the network lifetime and responsiveness, i.e., sensor data delivery reliability and latency. In addition, sensor node failures easily complicate/enforce/aggravate these irregularities. Valuable efforts have been conducted to discover topology specific anomalies such as coverage holes or critical/bottleneck nodes. Unfortunately, these efforts suffer from at least one of the following drawbacks: (a) They are centralized and consequently inefficient in large-scale networks, (b) they are tailored to one class of anomalies, or (c) do not propose how to remedy the identified anomaly. In this paper, we focus on sparse WSN which usually show varied topology irregularities and propose an in-network and localized strategy that efficiently (i) discovers generic topology irregularities, and (ii) identifies locations for minimal number of new augmented sensor deployments to remedy topology irregularities and sustain the desired operational requirements. We show the effectiveness and efficiency of the solution through a set of extensive simulations.

Distributed mobile transactions utilize commit protocols to achieve atomicity and consistent decisions. This is challenging, as mobile environments are typically characterized by frequent perturbations such as network disconnections and node failures. On one hand environmental constraints on mobile participants and wireless links may increase the resource blocking time of fixed participants. On the other hand frequent node and link failures complicate the design of atomic commit protocols by increasing both the transaction abort rate and resource blocking time. Hence, the deployment of classical commit protocols (such as two-phase commit) does not reasonably extend to distributed infrastructure-based mobile environments driving the need for perturbation-resilient commit protocols. In this article, we comprehensively consider and classify the perturbations of the wireless infrastructure-based mobile environment according to their impact on the outcome of commit protocols and on the resource blocking times. For each identified perturbation class a commit solution is provided. Consolidating these subsolutions, we develop a family of fault-tolerant atomic commit protocols that are tunable to meet the desired perturbation needs and provide minimized resource blocking times and optimized transaction commit rates. The framework is also evaluated using simulations and an actual testbed deployment.

Wireless Sensor Networks (WSN) display nonuniform energy usage distribution. This is mainly induced by the sink centric traffic or by non-uniform distribution of sensing activities and manifests as energy holes throughout the WSN. Holes can threaten the availability of the WSN by network partitioning and sensing voids. They are hard to predict, and consequently, proper function of the network requires systematic maintenance. Unfortunately, existing approaches do not systematically profile holes and focus only on very specific type of holes. In this work we present new distributed energy profiling algorithms for generalized types of energy holes. The algorithms search for boundary nodes and use them as a reference to calculate the energy needs of nodes within the hole. These, when aggregated, create angular and radial energy profiles. Extensive simulations show that the algorithms, when used for WSN maintenance, significantly help to extend the lifetime of the network.

The support of distributed atomic transactions in Mobile Ad-hoc Networks (MANET) is a key requirement for many mobile application scenarios. Atomicity is a fundamental property that ensures that all nodes decide a consistent outcome. As MANETs are characterized by frequent perturbations due to network partitioning and the fragility of nodes, providing atomicity is challenging. Existing protocols that ensure strict atomicity in MANETs are either bound to specific mobility pattern or based on building blocks such as consensus or group membership, not allowing arbitrary partitions or requiring exact knowledge about the members of a partition. These assumptions limit the deployment of these protocols to very restricted MANET scenarios, and may lead to poor commit rate, high message overhead or blocking related to intolerably long Commit/Abort decision times. In this paper, we present the first Partition-Tolerant Atomic Commit protocol (ParTAC) for MANETs which does not rely on consensus or group partition membership. As a consequence, ParTAC supports a significantly wider range of mobility patterns and partitioning scenarios than existing protocols. To reduce Commit/Abort decision times and prevent the protocol from blocking, ParTAC follows a best-effort strategy by defining a lifetime for every transaction after which the transaction is aborted. Further, we introduce a new coordination strategy based on a flexible preselection of multiple coordinators among the participating nodes. Thus, the failure of a single coordinator can be tolerated in the presence of network partitioning. Moreover, transactions can be aborted by any coordinator based on lifetime expiration. ParTAC is evaluated using simulations to demonstrate the performance of the protocol in terms of commit rate, message efficiency and Commit/Abort decision time.

Data-based agreement is increasingly used to implement traceable coordination across mobile entities such as ad-hoc networked (autonomous) vehicles. In our work, we focus on data-based agreement using database transactions where mobile entities agree on a set of coordinated tasks that need to be performed by them in an atomic way. Atomicity means that all transaction participants agree on a set of tasks which will be performed by them or no one of them is performing any task. The data about the agreed tasks and their corresponding stakeholders are kept in local databases as a proof for the obtained agreement. This proof might be needed by users and regularities/authorities involved depending on the application scenario. In this demo, we demonstrate our effort to provide for partition-aware atomic commit protocols for transactional data-based agreement.

Several proposals describing transport layer protocols for sensor networks appear in the literature. As each proposal is typically evaluated in the context of carefully selected parameters and scenarios, the benefits can be subjective. Also, given the limited details available of different proposals, it is difficult for developers of sensor network applications to select from the range of alternative transport protocols. This paper develops a common basis for evaluation of varied proposals. We first classify and review the existing protocols and evaluate them by measuring their performance in terms of responsiveness and efficiency in a conformal simulation environment and for a wide range of operational conditions. Common sources of poor performance are identified. Based on this experience, a set of design principles for the designers of applications and future transport protocols is presented.

Data transport is a core function for wireless sensor networks (WSNs) with different applications having varied requirements on the reliability and timeliness of data delivery. While node redundancy, inherent in WSNs, increases the fault tolerance, no guarantees on reliability levels can be assured. Furthermore, the frequent failures within WSNs impact the observed reliability over time and make it more challenging to achieve the desired reliability. Unfortunately, a framework for modeling reliability of data transport protocols in WSNs is currently missing. The existence of such a framework would simplify evaluation, comparison and also adaptation of these protocols. We formulate the problem of data transport in a WSN as a set of operations carried out on raw data. The operations aim at filtering the raw data to streamline its reliable transport towards the sink. Based on this formulation we systematically define a reliability framework. This paper argues for the usefulness of the reliability framework by classifying existing transport protocols and comparing their reliability

Transactions are required not only for wired networks but also for the emerging wireless environments where mobile and fixed hosts participate side by side in the execution of the transaction. This heterogenous environment is characterized by constraints in mobile host capabilities, network connectivity and also an increasing number of possible failure modes. Classical atomic commit protocols used in wired networks are therefore not directly suitable for this heterogenous environment. Furthermore, the few commit protocols designed for mobile transactions either consider mobile hosts only as initiators though not as active participants, or show a high resource blocking time. We present the Fault-Tolerant Pre-Phase Transaction Commit (FT-PPTC) protocol for mobile environments. FT-PPTC decouples the commit of mobile participants from that of fixed participants. Consequently, the commit set can be reduced to a set of entities in the fixed network. Thus, the commit can easily be supported by any traditional atomic commit protocol, such as the established 2PC protocol. We integrate fault-tolerance as a key feature of FT-PPTC. Performance evaluations confirm the efficiency, scalability and low resource blocking time of our approach

Mobility plays a major role in mobile ad hoc networks (MANETs) since it stresses networking tasks such as routing on one hand but aids to increase the network capacity and to overcome network partitioning on the other hand. To benefit from node mobility, a new class of MANET protocols and applications are designed to be delay-tolerant and mobility-aided. For delay-tolerant mobility-aided networking mobility on a large time-scale is a key feature. So far, in MANETs, the mobility is investigated on a short time-scale. That is why we present novel mobility metrics that quantify large time-scale mobility. Our approach is based on the pair-wise contacts between mobile nodes. We present a detailed statistical study of our novel metrics using the widely used random waypoint mobility model as an example. For the random waypoint model we introduce an analytical model, which allows protocol developers to analytically compute some of the designed metrics. In order to provide an easy access to these metrics in a network simulator, we provide a framework for ns-2.

Choosing appropriate information dissemination strategies is crucial in mobile ad hoc networks (MANET) due to the frequent topology changes. Flooding-based approaches like diffusion have a strong similarity with epidemic spreading of diseases. Applying epidemiological models to information diffusion allows the evaluation of such strategies depending on the MANET characteristics, e.g. the node density. In order to choose appropriate strategies at run time, the model should be easily evaluated.In this paper, an epidemic model is developed for a simple information diffusion algorithm based on simulation results. We analytically investigate the impact of node density on information diffusion. The analytical model allows the evaluation at runtime, even on devices with restricted resources, and thus enables mobile nodes to dynamically adapt their diffusion strategies depending on the local node density.