Publications
246 results found
Cao J, McCann J, Nahrstedt K, et al., 2017, Message from General and Program Co-Chairs
Breza M, McCann J, 2017, Polite Broadcast Gossip for IOT Configuration Management, SmartComp 2017
Jackson G, Kartakis S, McCann J, 2017, Accurate models of energy harvesting for smart environments, IEEE International Conference on Smart Computing (SMARTCOMP 2017), Publisher: IEEE
Over the last decade, the energy optimization ofresource constrained sensor nodes constitutes a major researchtopic in smart environments. However, state of the art energyoptimization algorithms make strong and unrealistic assumptionsof energy models, both in simulations and during the operation ofsmart systems. For instance, simplistic energy models for energyharvesting leads to inaccurate representation and prediction ofthe true dynamics of energy. Consequently, systems for smartenvironments are unable to meet expected performance criteria.In this paper, we propose innovative models to overcome thedrawbacks of simplistic energy representations in smart environments.We provide the insights of how to generate preciselightweight energy models. Using the physical properties of solarand flow energy harvesting as case studies, the trade-off betweenenergy harvesting inference and real-time measurement of energygeneration is explored. To evaluate our proposed energy modelsagainst the simplistic versions, we use real measured data fromour environmental micro-climate monitoring deployment in anurban park and a 103% improvement is seen. Additionally,to define the trade-offs between inferred and measured energygeneration, experiments are conducted utilizing solar and smartwater testbeds.
Ren X, Yu CM, Yu W, et al., 2017, High-dimensional crowdsourced data distribution estimation with local privacy, 2016 IEEE International Conference on Computer and Information Technology (CIT), Publisher: IEEE, Pages: 226-233
High-dimensional crowdsourced data collected from a large number of users may produc3 rich knowledge for our society but also bring unprecedented privacy threats to participants. Recently differential privacy has been proposed as an effective means to mitigate privacy concerns. However, existing work on differential privacy suffers from the 'curse of high-dimensionality' (data with multiple attributes) and high scalability (data with large scale records). Moreover, traditional methods of differential privacy were achieved via aggregation results, which cannot guarantee local privacy for distributed users in crowdsourced systems. To deal with these issues, in this paper we propose a novel scheme that can efficiently estimate multivariate joint distribution for high-dimensional data with local privacy. On the client side, we employ randomized response techniques to locally transform data from distributed users into privacy-preserving bit strings, which can prevent potential inside privacy attacks in crowdsourced systems. On the server side, the crowdsourced bit strings are aggregated for multivariate distribution estimation. Specifically, we first propose a multivariate version of the expectation maximization (EM) based algorithm to estimate the joint distribution of high dimensional data. To speed up the performance, unlike the EM-based method that needs to scan each user's bit string, we propose to use Lasso regression to obtain the distribution estimation from the aggregation information only once, which can significantly reduce the computation time for multivariate distribution estimation. Extensive experiments on real-world datasets demonstrate the efficiency of our multivariate distribution estimation scheme over existing estimation schemes.
Haghighi M, Qin Z, Carboni D, et al., 2017, Game theoretic and auction-based algorithms towards opportunistic communications in LPWA LoRa networks, IEEE World Forum on Internet of Things, Publisher: IEEE, Pages: 735-740
Low Power Wide Area (LPWA) networks have been the enabling technology for large-scale sensor and actuator networks. Low cost, energy-efficiency and longevity of such networks make them perfect candidates for smart city applications. LoRa is a new LPWA standard based on spread spectrum technology, which is suitable for sensor nodes enabling long battery life and bi-directional communication but with low data rates. In this paper, we will demonstrate a use-case inspired model in which, end-nodes with multiple radio transceivers (LoRa/WiFi/BLE) have the option to interconnect via multiple networks to improve communications resilience under the diverse conditions of a smart city of a billion devices. To facilitate this, each node has the ability to switch radio communications opportunistically and adaptively, and this is based on the application requirements and dynamic radio parameters.
Shi F, Adeel, Theodoridis T, et al., 2017, OppNet: enabling citizen-centric urban IoT data collection through opportunistic connectivity service, Internet of Things (WF-IoT), 2016 IEEE 3rd World Forum on, Publisher: IEEE
Urban IoT data collection is challenging due to the limitations of the fixed sensing infrastructures. Instead of transmitting data directly through expensive cellular networks, citizen-centric data collection scheme through opportunistic network takes advantage of human mobility as well as cheap WiFi and D2D communication. In this paper, we present OppNet, which implements a context aware data forwarding algorithm and fills the gap between theoretical modelling of opportunistic networking and real deployment of citizen-centric data collection system. According to the results from a 3-day real-life experiment, OppNet shows consistent performance in terms of number of hops and time delay. Moreover, the underlying social structure can be clearly identified by analysing social contact data collected through OppNet.
Yu W, McCann J, 2017, Random walk with restart over dynamic graphs, 2016 IEEE 16th International Conference on Data Mining (ICDM), Publisher: IEEE, Pages: 589-598, ISSN: 2374-8486
Random Walk with Restart (RWR) is an appealing measure of proximity between nodes based on graph structures. Since real graphs are often large and subject to minor changes, it is prohibitively expensive to recompute proximities from scratch. Previous methods use LU decomposition and degree reordering heuristics, entailing O(|V |3) time and O(|V |2) memory to compute all (|V |2) pairs of node proximities in a static graph. In this paper, a dynamic scheme to assess RWR proximities is proposed: (1) For unit update, we characterize the changes to all-pairs proximities as the outer product of two vectors. We notice that the multiplication of an RWR matrix and its transition matrix, unlike traditional matrix multiplications, is commutative. This can greatly reduce the computation of all-pairs proximities from O(|V |3) to O(|Δ|) time for each update without loss of accuracy, where |Δ| (≪ |V |2) is the number of affected proximities. (2) To avoid O(|V |2) memory for all pairs of outputs, we also devise efficient partitioning techniques for our dynamic model, which can compute all pairs of proximities segment-wisely within O(l|V |) memory and O(⌈ |V | l ⌉) I/O costs, where 1 ≤ l ≤ |V | is a user-controlled trade-off between memory and I/O costs. (3) For bulk updates, we also devise aggregation and hashing methods, which can discard many unnecessary updates further and handle chunks of unit updates simultaneously. Our experimental results on various datasets demonstrate that our methods can be 1-2 orders of magnitude faster than other competitors while securing scalability and exactness.
Wu D, Arkhipov DI, Kim M, et al., 2017, ADDSEN: adaptive data processing and dissemination for drone swarms in urban sensing, IEEE Transactions on Computers, Vol: 66, Pages: 183-198, ISSN: 0018-9340
We present ADDSEN middleware as a holistic solution for Adaptive Data processing and dissemination for Drone swarms in urban SENsing. To efficiently process sensed data in the middleware, we have proposed a cyber-physical sensing framework using partially ordered knowledge sharing for distributed knowledge management in drone swarms. A reinforcement learning dissemination strategy is implemented in the framework. ADDSEN uses online learning techniques to adaptively balance the broadcast rate and knowledge loss rate periodically. The learned broadcast rate is adapted by executing state transitions during the process of online learning. A strategy function guides state transitions, incorporating a set of variables to reflect changes in link status. In addition, we design a cooperative dissemination method for the task of balancing storage and energy allocation in drone swarms. We implemented ADDSEN in our cyber-physical sensing framework, and evaluation results show that it can achieve both maximal adaptive data processing and dissemination performance, presenting better results than other commonly used dissemination protocols such as periodic, uniform and neighbor protocols in both single-swarm and multi-swarm cases.
Johnson M, McCann J, Santer M, et al., 2017, On orbit validation of solar sailing control laws with thin-film spacecraft, The Fourth International Symposium on Solar Sailing, Publisher: Japan Space Forum
Many innovative approaches to solar sail mission and trajectory design have been proposed over the years, but very few ever have the opportunity to be validated on orbit with real spacecraft. Thin-Film Spacecraft/Lander/Rovers (TF-SLRs) are a new class of very low cost, low mass space vehicle which are ideal for inexpensively and quickly testing in flight new approaches to solar sailing. This paper describes using TF-SLR based micro solar sails to implement a generic solar sail test bed on orbit. TF-SLRs are high area-to-mass ratio (A/m) spacecraft developed for very low cost consumer and scientific deep space missions. Typically based on a 5 μm or thinner metalised substrate, they include an integrated avionics and payload system-on-chip (SoC) die bonded to the substrate with passive components and solar cells printed or deposited by Metal Organic Chemical Vapour Deposition (MOCVD). The avionics include UHF/S-band transceivers, processors, storage, sensors and attitude control provided by integrated magnetorquers and reflectivity control devices. Resulting spacecraft have a typical thickness of less than 50 μm, are 80 mm in diameter, and have a mass of less than 100 mg resulting in sail loads of less than 20 g/m2. TF-SLRs are currently designed for direct dispensing in swarms from free flying 0.5U Interplanetary CubeSats or dispensers attached to launch vehicles. Larger 160 mm, 320 mm and 640 mm diameter TF-SLRs utilizing a CubeSat compatible TWIST deployment mechanism that maintains the high A/m ratio are also under development. We are developing a mission to demonstrate the utility of these devices as a test bed for experimenting with a variety of mission designs and control laws. Batches of up to one hundred TF-SLRs will be released on earth escape trajectories, with each batch executing a heterogeneous or homogenous mixture of control laws and experiments. Up to four releases at different points in orbit are currently envisaged with experiments currently
Yu W, Lin X, Zhang W, et al., 2017, Dynamical SimRank search on time-varying networks, Departmental Technical Report: 17/9, Publisher: Department of Computing, Imperial College London, 17/9
SimRank is an appealing pair-wise similarity measure based on graph structure. It iteratively follows the intuition that two nodes are assessed as similar if they are pointed to by similar nodes. Many real graphs are large, and links are constantly subject to minor changes. In this article, we study the efficient dynamical computation of all-pairs SimRanks on time-varying graphs. Existing methods for the dynamical SimRank computation (e.g., L-TSF and READS) mainly focus on top-k search with respect to a given query. For all-pairs dynamical SimRank search, Li et al.'s approach was proposed for this problem. It first factorizes the graph via a singular value decomposition (SVD), and then incrementally maintains such a factorization in response to link updates at the expense of exactness. As a result, all pairs of SimRanks are updated approximately, yielding O(r^4*n^2) time and O(r^2*n^2) memory in a graph with n nodes, where r is the target rank of the low-rank SVD. Our solution to the dynamical computation of SimRank comprises of five ingredients: (1) We first consider edge update that does not accompany new node insertions. We show that the SimRank update ∆S in response to every link update is expressible as a rank-one Sylvester matrix equation. This provides an incremental method requiring O(Kn^2) time and O(n^2) memory in the worst case to update n2pairs of similarities for K iterations. (2) To speed up the computation further, we propose a lossless pruning strategy that captures the "affected areas" of ∆S to eliminate unnecessary retrieval. This reduces the time of the incremental SimRank to O(K(m+|AFF|)), where m is the number of edges in the old graph, and |AFF| (<=n^2) is the size of "affected areas" in ∆S, and in practice, |AFF|<.
Martins PMN, McCann JA, 2017, NETWORK-WIDE PROGRAMMING CHALLENGES IN CYBER-PHYSICAL SYSTEMS, CYBER-PHYSICAL SYSTEMS: FOUNDATIONS, PRINCIPLES AND APPLICATIONS, Editors: Song, Rawat, Jeschke, Brecher, Publisher: ACADEMIC PRESS LTD-ELSEVIER SCIENCE LTD, Pages: 103-113, ISBN: 978-0-12-803801-7
- Author Web Link
- Cite
- Citations: 2
Gutierrez V, Theodoridis E, Mylonas G, et al., 2016, Co-Creating the Cities of the Future, Sensors, Vol: 16, ISSN: 1424-8220
In recent years, the evolution of urban environments, jointly with the progress of theInformation and Communication sector, have enabled the rapid adoption of new solutions thatcontribute to the growth in popularity of Smart Cities. Currently, the majority of the world populationlives in cities encouraging different stakeholders within these innovative ecosystems to seek newsolutions guaranteeing the sustainability and efficiency of such complex environments. In this work,it is discussed how the experimentation with IoT technologies and other data sources form the citiescan be utilized to co-create in the OrganiCity project, where key actors like citizens, researchers andother stakeholders shape smart city services and applications in a collaborative fashion. Furthermore,a novel architecture is proposed that enables this organic growth of the future cities, facilitating theexperimentation that tailors the adoption of new technologies and services for a better quality of life,as well as agile and dynamic mechanisms for managing cities. In this work, the different componentsand enablers of the OrganiCity platform are presented and discussed in detail and include, amongothers, a portal to manage the experiment life cycle, an Urban Data Observatory to explore data assets,and an annotations component to indicate quality of data, with a particular focus on the city-scaleopportunistic data collection service operating as an alternative to traditional communications.
Kolcun R, Boyle D, McCann J, 2016, Efficient In-Network Processing for a Hardware-Heterogeneous IoT, IoT2016 - 6th International Conference on the Internet of Things, Publisher: IEEE
As the number of small, battery-operated, wireless-enabled devices deployed in various applications of Internet of Things (IoT), Wireless Sensor Networks (WSN), and Cyber-physical Systems (CPS) is rapidly increasing, so is the number of data streams that must be processed. In cases where data do not need to be archived, centrally processed, or federated, in-network data processing is becoming more common. For this purpose, various platforms like D RAGON , Innet, and CJF were proposed. However, these platforms assume that all nodes in the network are the same, i.e. the network is homogeneous. As Moore’s law still applies, nodes are becoming smaller, more powerful, and more energy efficient each year; which will continue for the foreseeable future. Therefore, we can expect that as sensor networks are extended and updated, hardwareheterogeneity will soon be common in networks - the same trend as can be seen in cloud computing infrastructures. This heterogeneity introduces new challenges in terms of choosing an in-network data processing node, as not only its location, but also its capabilities, must be considered. This paper introduces a new methodology to tackle this challenge, comprising three new algorithms - Request, Traverse, and Mixed - for efficiently locating an in-network data processing node, while taking into account not only position within the network but also hardware capabilities. The roposed algorithms are evaluated against a naïve approach and achieve up to 90% reduction in network traffic during long-term data processing, while spending a similar amount time in the discovery phase.
Kartakis S, Choudhary B, Gluhak A, et al., 2016, Demystifying Low-Power Wide-Area Communications for City IoT Applications, ACM WiNTECH 2016 Workshop, MobiCom, Publisher: ACM, Pages: 2-8
Low Power Wide Area (LPWA) communication technologieshave the potential to provide a step change in the enablementof cost-effective and energy efficient Internet ofThings (IoT) applications. With an increase in the numberof offerings available the real performance of these emergingtechnologies remain unclear. That is, each technologycomes with its own advantages and limitations; yet there isa lack of comparative studies that examine their trade-offsbased on empirical evidence. This poses a major challengeto IoT solution architects and developers in selecting an appropriatetechnology for an envisioned IoT application in agiven deployment context.In this paper, we look beyond data sheets and white papersof LPWA communication technologies and provide insightsinto the performance of three emerging LPWA solutionsbased on real world experiments with different traf-fic loads and in different urban deployment contexts. Underthe context of this study, specialized hardware was createdto incorporate the different technologies and provide scientificquantitative and qualitative information related to datarates, success rates, transmission mode energy and powerconsumption, and communication ranges. The results of experimentationhighlight the practicalities of placing LPWAtechnologies in real spaces and provide guidelines to IoT solutiondevelopers in terms of LPWA technology selection.Overall aim is to facilitate the design of new LPWA technologiesand adaptive communication strategies that informfuture IoT platforms.
Zhao C, Yang S, Yang X, et al., 2016, Rapid, user-transparent, and trustworthy device pairing for D2D-enabled mobile crowdsourcing, IEEE Transactions on Mobile Computing, Vol: 16, Pages: 2008-2022, ISSN: 1536-1233
Mobile Crowdsourcing is a promising service paradigm utilizing ubiquitous mobile devices to facilitate large-scale crowdsourcing tasks (e.g. urban sensing and collaborative computing). Many applications in this domain require Device-to-Device (D2D) communications between participating devices for interactive operations such as task collaborations and file transmissions. Considering the private participating devices and their opportunistic encountering behaviors, it is highly desired to establish secure and trustworthy D2D connections in a fast and autonomous way, which is vital for implementing practical Mobile Crowdsourcing Systems (MCSs). In this paper, we develop an efficient scheme, Trustworthy Device Pairing (TDP), which achieves user-transparent secure D2D connections and reliable peer device selections for trustworthy D2D communications. Through rigorous analysis, we demonstrate the effectiveness and security intensity of TDP in theory. The performance of TDP is evaluated based on both real-world prototype experiments and extensive trace-driven simulations. Evaluation results verify our theoretical analysis and show that TDP significantly outperforms existing approaches in terms of pairing speed, stability, and security.
Yang S, Adeel U, Tahir Y, et al., 2016, Practical opportunistic data collection in wireless sensor networks with mobile sinks, IEEE Transactions on Mobile Computing, Vol: 16, Pages: 1420-1433, ISSN: 1558-0660
Wireless Sensor Networks with Mobile Sinks (WSN-MSs) are considered a viable alternative to the heavy cost ofdeployment of traditional wireless sensing infrastructures at scale. However, current state-of-the-art approaches perform poorly inpractice due to their requirement of mobility prediction and specific assumptions on network topology. In this paper, we focus on lowdelayand high-throughput opportunistic data collection in WSN-MSs with general network topologies and arbitrary numbers of mobilesinks. We first propose a novel routing metric, Contact-Aware ETX (CA-ETX), to estimate the packet transmission delay caused byboth packet retransmissions and intermittent connectivity. By implementing CA-ETX in the defacto TinyOS routing standard CTP andthe IETF IPv6 routing protocol RPL, we demonstrate that CA-ETX can work seamlessly with ETX. This means that current ETXbasedrouting protocols for static WSNs can be easily extended to WSN-MSs with minimal modification by using CA-ETX. Further,by combing CA-ETX with the dynamic backpressure routing, we present a throughput-optimal scheme Opportunistic BackpressureCollection (OBC). Both CA-ETX and OBC are lightweight, easy to implement, and require no mobility prediction. Through test-bedexperiments and extensive simulations, we show that the proposed schemes significantly outperform current approaches in terms ofpacket transmission delay, communication overhead, storage overheads, reliability, and scalability.
Kolcun R, Boyle DE, McCann JA, 2016, Efficient distributed query processing, IEEE Transactions on Automation Science and Engineering, Vol: 13, Pages: 1230-1246, ISSN: 1042-296X
A variety of wireless networks, including applications of Wireless Sensor Networks, Internet of Things, and Cyber-physical Systems, increasingly pervade our homes, retail, transportation systems, and manufacturing processes. Traditional approaches communicate data from all sensors to a central system, and users (humans or machines) query this central point for results, typically via the web. As the number of deployed sensors, and thus generated data streams, is increasing exponentially, this traditional approach may no longer be sustainable or desirable in some application contexts. Therefore, new approaches are required to allow users to directly interact with the network, for example, requesting data directly from sensor nodes. This is difficult, as it requires every node to be capable of point-to-point routing, in addition to identifying a subset of nodes that can fulfil a user's query. This paper presents Dragon, a platform that allows any node in the network to identify all nodes that satisfy user queries, i.e., request data from nodes, and relay the result to the user. The Dragon platform achieves this in a fully distributed way. No central orchestration is required, network overheads are low, and latency is improved over existing comparable methods. Dragon is evaluated on networks of various topologies and different network densities. It is compared with the state-of-the-art algorithms based on summary trees, like Innet and SENS-Join. Dragon is shown to outperform these approaches up to 88% in terms of network traffic required, also a proxy for energy efficiency, and 84% in terms of processing delay.
Carboni D, Gluhak A, McCann JA, et al., 2016, Contextualising Water Use in Residential Settings: A Survey of Non-Intrusive Techniques and Approaches, Sensors, Vol: 16, ISSN: 1424-8239
Water monitoring in households is important to ensure the sustainability of fresh water reserves on our planet. It provides stakeholders with the statistics required to formulate optimal strategies in residential water management. However, this should not be prohibitive and appliance-level water monitoring cannot practically be achieved by deploying sensors on every faucet or water-consuming device of interest due to the higher hardware costs and complexity, not to mention the risk of accidental leakages that can derive from the extra plumbing needed. Machine learning and data mining techniques are promising techniques to analyse monitored data to obtain non-intrusive water usage disaggregation. This is because they can discern water usage from the aggregated data acquired from a single point of observation. This paper provides an overview of water usage disaggregation systems and related techniques adopted for water event classification. The state-of-the art of algorithms and testbeds used for fixture recognition are reviewed and a discussion on the prominent challenges and future research are also included.
Kartakis S, Yu W, Akhavan M, et al., 2016, Adaptive edge analytics for distributed networked control of water systems, 2016 IEEE First International Conference on Internet-of-Things Design and Implementation, Publisher: IEEE, Pages: 72-82
Over the last decade, there has been a trend where water utility companies aim to make water distribution networks more intelligent in order to improve their quality of service, reduce water waste, minimize maintenance costs etc., by incorporating IoT technologies. Current state of the art solutions use expensive power hungry deployments to monitor and transmit water network states periodically in order to detect anomalous behaviors such as water leakage and bursts. However, more than 97% of water network assets are remote away from power and are often in geographically remote underpopulated areas, facts that make current approaches unsuitable for next generation more dynamic adaptive water networks. Battery-driven wireless sensor/actuator based solutions are theoretically the perfect choice to support next generation water distribution. In this paper, we present an end-to-end water leak localization system, which exploits edge processing and enables the use of battery-driven sensor nodes. Our system combines a lightweight edge anomaly detection algorithm based on compression rates and an efficient localization algorithm based on graph theory. The edge anomaly detection and localization elements of the systems produce a timely and accurate localization result and reduce the communication by 99% compared to the traditional periodic communication. We evaluated our schemes by deploying non-intrusive sensors measuring vibrational data on a real-world water test rig that have had controlled leakage and burst scenarios implemented.
Kartakis S, Milojevic Jevric M, Tzagkarakis G, et al., 2016, Energy-based Adaptive Compression in Water Network Control Systems, CySWater2016, CPSWeek, Publisher: CPS
Contemporary water distribution networks exploitInternet of Things (IoT) technologies to monitor and controlthe behavior of water network assets. Smart meters/sensorand actuator nodes have been used to transfer informationfrom the water network to data centers for further analysis.Due to the underground position of water assets, many watercompanies tend to deploy battery driven nodes which lastbeyond the 10-year mark. This prohibits the use of high-samplerate sensing therefore limiting the knowledge we can obtainfrom the recorder data. To alleviate this problem, efficientdata compression enables high-rate sampling, whilst reducingsignificantly the required storage and bandwidth resourceswithout sacrificing the meaningful information content. Thispaper introduces a novel algorithm which combines the accuracyof standard lossless compression with the efficiencyof a compressive sensing framework. Our method balancesthe tradeoffs of each technique and optimally selects the bestcompression mode by minimizing reconstruction errors, giventhe sensor node battery state. To evaluate our algorithm, realhigh-sample rate water pressure data of over 170 days and 25sensor nodes of our real world large scale testbed was used.The experimental results reveal that our algorithm can reducecommunication around 66% and extend battery life by 46%compared to traditional periodic communication techniques.
Wu D, Liu Q, Li Y, et al., 2016, Adaptive Lookup of Open WiFi Using Crowdsensing, IEEE/ACM Transactions on Networking, Vol: 24, Pages: 3634-3647, ISSN: 1063-6692
Open WiFi access points (APs) are demonstratingthat they can provide opportunistic data services to movingvehicles. We present CrowdWiFi, a novel system to lookup roadside WiFi APs located outdoors or inside buildings.CrowdWiFi consists of two components: online compressivesensing (CS) and offline crowdsourcing. Online CS presents anefficient framework for the coarse-grained estimation of nearbyAPs along the driving route, where received signal strength (RSS)values are recorded at runtime, and the number and location ofthe APs are recovered immediately based on limited RSS readingsand adaptive CS operations. Offline crowdsourcing assigns theonline CS tasks to crowd-vehicles and aggregates answers on abipartite graphical model. Crowd-server also iteratively infersthe reliability of each crowd-vehicle from the aggregated sensingresults, and then refines the estimation of the APs using weightedcentroid processing. Extensive simulation results and real testbedexperiments confirm that CrowdWiFi can successfully reducethe computation cost and energy consumption of roadside WiFilookup, while maintaining satisfactory localization accuracy.
Wu D, Lambrinos L, Przepiorka T, et al., 2016, Facilitating Mobile Access to Social Media Content on Urban Underground Metro Systems, IEEE Conference on Computer Communications Workshops (INFOCOM WKSHPS), Publisher: IEEE, ISSN: 2159-4228
Tahir, Yang, Koliousis, et al., 2015, UDRF: Multi-resource Fairness for Complex Jobs with Placement Constraints, IEEE GLOBECOM 2015, Publisher: IEEE, Pages: 1-7
In this paper, we study the problem of multi- resource fairness in systems running complex jobs that consist of multiple interconnected tasks. A job is considered finished when all its corresponding tasks have been executed in the system. Tasks can have different resource requirements. Because of special demands on particular hardware or software, tasks may have placement constraints limiting the type of machines they can run on. We develop User-Dependence Dominant Resource Fairness (UDRF), a generalized version of max-min fairness that combines graph theory and the notion of dominant re- source shares to ensure multi-resource fairness between complex workflows. UDRF satisfies several desirable properties including strategy proofness, which ensures that users do not benefit from misreporting their true resource demands. We propose an offline algorithm that computes optimal UDRF allocation. But optimality comes at a cost, especially for systems where schedulers need to make thousands of online scheduling decisions per second. Therefore, we develop a lightweight online algorithm that closely approximates UDRF. Besides that, we propose a simple mechanism to decentralize the UDRF scheduling process across multiple schedulers. Large-scale simulations driven by Google cluster-usage traces show that UDRF achieves better resource utilization and throughput compared to the current state-of-the-art in fair resource allocation.
Tahir Y, Yang S, Koliousis A, et al., 2015, UDRF: Multi-Resource Fairness for Complex Jobs with Placement Constraints, 2015 IEEE Global Communications Conference (GLOBECOM), Publisher: IEEE
In this paper, we study the problem of multi-resource fairness in systems with multiple users. Each user requires to run one or more complex jobs that consist of multiple interconnected tasks. A job is considered finished when all its corresponding tasks have been executed in the system. Tasks can have different resource requirements. Because of special demands on particular hardware or software, tasks can have placement constraints limiting the type of machines they can run on. We develop User-Dependence Dominant Resource Fairness (UDRF), a generalized version of max-min fairness that combines graph theory and the notion of dominant resource shares to ensure multi- resource fairness between users with complex jobs. UDRF satisfies several desirable properties including strategy proofness, which ensures that users do not benefit from misreporting their true resource demands. We propose an offline algorithm that computes optimal UDRF allocation while the scheduling process can be to be decentralize across multiple schedulers. But optimality comes at a cost, especially for systems where schedulers need to make thousands of online scheduling decisions per second. Therefore, we develop a lightweight online algorithm that closely approximates UDRF. Large-scale simulations driven by Google cluster- usage traces show that UDRF achieves better resource utilization and throughput compared to the current state-of-the-art in multi-resource fair allocation.
Yang S, Tahir Y, Chen P, et al., 2015, Distributed Optimization in Energy Harvesting Sensor Networks with Dynamic In-network Data Processing, IEEE INFOCOM 2016, Publisher: Institute of Electrical and Electronics Engineers (IEEE), ISSN: 0743-166X
Energy Harvesting Wireless Sensor Networks (EH- WSNs) have been attracting increasing interest in recent years. Most current EH-WSN approaches focus on sensing and net- working algorithm design, and therefore only consider the energy consumed by sensors and wireless transceivers for sensing and data transmissions respectively. In this paper, we incorporate CPU-intensive edge operations that constitute in-network data processing (e.g. data aggregation/fusion/compression) with sens- ing and networking; to jointly optimize their performance, while ensuring sustainable network operation (i.e. no sensor node runs out of energy). Based on realistic energy and network models, we formulate a stochastic optimization problem, and propose a lightweight on-line algorithm, namely Recycling Wasted Energy (RWE), to solve it. Through rigorous theoretical analysis, we prove that RWE achieves asymptotical optimality, bounded data queue size, and sustainable network operation. We implement RWE on a popular IoT operating system, Contiki OS, and eval- uate its performance using both real-world experiments based on the FIT IoT-LAB testbed, and extensive trace-driven simulations using Cooja. The evaluation results verify our theoretical analysis, and demonstrate that RWE can recycle more than 90% wasted energy caused by battery overflow, and achieve around 300% network utility gain in practical EH-WSNs.
Yu W, Mccann, 2015, Effectively Positioning Water Loss Event in Smart Water Networks, 2nd International Electronic Conference on Sensors and Applications, Publisher: MDPI, ISSN: 1424-8220
Kolcun R, Boyle D, McCann JA, 2015, Optimal processing node discovery algorithm for distributed computing in IoT, 5th International Conference on the Internet of Things, Publisher: IEEE, Pages: 72-79
The number of Internet-connected sensing and control devices is growing. Some anticipate them to number in excess of 212 billion by 2020. Inherently, these devices generate continuous data streams, many of which need to be stored and processed. Traditional approaches, whereby all data are shipped to the cloud, may not continue to be effective as cloud infrastructure may not be able to handle myriads of data streams and their associated storage and processing needs. Using cloud infrastructure alone for data processing significantly increases latency, and contributes to unnecessary energy inefficiencies, including potentially unnecessary data transmission in constrained wireless networks, and on cloud computing facilities increasingly known to be significant consumers of energy. In this paper we present a distributed platform for wireless sensor networks which allows computation to be shifted from the cloud into the network. This reduces the traffic in the sensor network, intermediate networks, and cloud infrastructure. The platform is fully distributed, allowing every node in a homogeneous network to accept continuous queries from a user, find all nodes satisfying the user's query, find an optimal node (Fermat-Weber point) in the network upon which to process the query, and provide the result to the user. Our results show that the number of required messages can be decreased up to 49% and processing latency by 42% in comparison with state-of-the-art approaches, including Innet.
Yu W, McCann JA, 2015, Gauging Correct Relative Rankings For Similarity Search, 24th ACM International on Conference on Information and Knowledge Management (CIKM '15), Publisher: Association for Computing Machinery, Pages: 1791-1794
One of the important tasks in link analysis is to quantify the similarity between two objects based on hyperlink structure. SimRank is an attractive similarity measure of this type. Existing work mainly focuses on absolute SimRank scores, and often harnesses an iterative paradigm to compute them. While these iterative scores converge to exact ones with the increasing number of iterations, it is still notoriously difficult to determine how well the relative orders of these iterative scores can be preserved for a given iteration. In this paper, we propose efficient ranking criteria that can secure correct relative orders of node-pairs with respect to SimRank scores when they are computed in an iterative fashion. Moreover, we show the superiority of our criteria in harvesting top-K SimRank scores and bucket orders from a full ranking list. Finally, viable empirical studies verify the usefulness of our techniques for SimRank top-K ranking and bucket ordering.
Zhao C, Shi F, Huang R, et al., 2015, Trustworthy device pairing for opportunistic device-to-device communications in mobile crowdsourcing systems, S3 Wireless of the Students, by the Students, for the Students, Publisher: ACM, Pages: 4-6
Mobile Crowdsourcing System is an emerging service paradigm base on numerous personal smart devices, where the Deviceto- Device communication among opportunistically encountered participating devices is an indispensable part of task allocation, file transmission and data collaboration. Considering that participating devices are privately held and opportunistically encountered, we design the Trustworthy Device Pairing (TDP) scheme that realizes user-transparent sharing secret key negotiation and reliable peer device determination for trustworthy spontaneous D2D transactions. TDP is demonstrated to be effective based on our proof-of-concept implementation, and a further evaluation on efficiency will be conducted.
Yang X, Zhao C, Yang S, et al., 2015, A Systematic Key Management Mechanism for Practical Body Sensor Networks, IEEE International Conference on Communications (ICC), Publisher: IEEE, Pages: 7310-7315, ISSN: 1550-3607
Security plays a vital role in promoting the practicality of Wireless Body Sensor Networks (BSNs), which provides a promising solution to precise human physiological status monitoring. A fundamental security issue in BSN is key management, including establishment and maintenance of the key system. However, current BSN key management solutions are either designed for specific phases of a BSN's life-time or restricted to strong assumptions such as homogeneous BSN composition, pre-deployed key materials, and existing secure path, which limits their applications in real-world BSNs. In this paper, we develop the Systematic Key Management (SKM) for practical BSNs, where basic human interactions are conducted for non-predeployed secure BSN initialization, and authenticated key agreement is achieved using lightweight non-pairing certificateless public key cryptography. We construct a BSN prototype consisting of self-designed motes and Android phones to evaluate the real-world performance of SKM. Through extensive simulations and test-bed experiments, we demonstrate that our lightweight SKM scheme manages to provide high security guarantee while outperforming state-of-the-art approaches in terms of both computation and storage efficiency.
This data is extracted from the Web of Science and reproduced under a licence from Thomson Reuters. You may not copy or re-distribute this data in whole or in part without the written consent of the Science business of Thomson Reuters.