Hierarchical power management in disruption tolerant networks using traffic-aware optimization

Recent efforts in Disruption Tolerant Networks (DTNs) have shown that mobility can be a powerful means for delivering messages in highly-challenged environments. DTNs are wireless mobile networks that are particularly useful in sparse environments where the density of nodes is insufficient to support direct end-to-end communication. Unfortunately, many mobility scenarios depend on untethered devices with limited energy supplies. Without careful management, depleted energy supplies will degrade network connectivity and counteract the robustness gained by mobility. A primary concern is the energy consumed by wireless communication, and in particular the energy consumed in searching for other nodes to communicate with. In this paper we examine a hierarchical radio architecture in which nodes are equipped with two complementary radios: a long-range, high-power radio and a short-range, low-power radio. In this architecture energy can be conserved by using the low-power radio to discover communication opportunities with other nodes and waking the high-power radio to undertake data transmission. We develop a generalized power management framework for controlling the wake-up intervals of the two radios. In addition, we show how to incorporate knowledge of the traffic load, and we devise approximation algorithms to control the sleep/wake-up cycling to provide maximum energy conservation while discovering enough communication opportunities to handle that load. We evaluate our schemes through simulation under various mobility scenarios. Our results show that our generalized power management scheme can tune wake-up intervals of the two radios to balance energy efficiency and delivery performance. Also, when traffic load can be predicted, our approximation algorithms reduce energy consumption from 60% to 99% compared to no power management.     

Dynamic slack allocation algorithms for energy minimization on parallel machines

We explore novel algorithms for DVS (Dynamic Voltage Scaling) based energy minimization of DAG (Directed Acyclic Graph) based applications on parallel and distributed machines in dynamic environments. Static DVS algorithms for DAG execution use the estimated execution time. The estimated time in practice is overestimated or underestimated. Therefore, many tasks may be completed earlier or later than expected during the actual execution. For overestimation, the extra available slack can be added to future tasks so that energy requirements can be reduced. For underestimation, the increased time may cause the application to miss the deadline. Slack can be reduced for future tasks to reduce the possibility of not missing the deadline. In this paper, we present novel dynamic scheduling algorithms for reallocating the slack for future tasks to reduce energy and/or satisfy deadline constraints. Experimental results show that our algorithms are comparable to static algorithms applied at runtime in terms of energy minimization and deadline satisfaction, but require considerably smaller computational overhead.     

Iterative MILP methods for vehicle-control problems

Mixed-integer linear programming (MILP) is a powerful tool for planning and control problems because of its modeling capability and the availability of good solvers. However, for large models, MILP methods suffer computationally. In this paper, we present iterative MILP algorithms that address this issue. We consider trajectory-generation problems with obstacle-avoidance requirements and minimum-time trajectory-generation problems. These problems involve vehicles that are described by mixed logical dynamical equations, a form of hybrid system. The algorithms use fewer binary variables than standard MILP methods, and require less computational effort.     

Radio resource allocation problems for OFDMA cellular systems

Orthogonal frequency division multiple-access (OFDMA) manages to efficiently exploit the inherent multi-user diversity of a cellular system by performing dynamic resource allocation. Radio resource allocation is the technique that assigns to each user in the system a subset of the available radio resources (mainly power and bandwidth) according to a certain optimality criterion on the basis of the experienced link quality. In this paper we address the problem of resource allocation in the downlink of a multi-cellular OFDMA system. The allocation problem is formulated with the goal of minimizing the transmitted power subject to individual rate constraint for each user. Exact and heuristic algorithms are proposed for the both the single-cell and the multi-cell scenario. In particular, we show that in the single-cell scenario the allocation problem can be efficiently solved following a network flow approach. In the multi-cell scenario we assume that all cells use the same frequencies and therefore the allocation problem is complicated by the presence of strong multiple access interference. We prove that the problem is strongly NP-hard, and we present an exact approach based on an MILP formulation. We also propose two heuristic algorithms designed to be simple and fast. All algorithms are tested and evaluated through an experimental campaign on simulated instances. Experimental results show that, although suboptimal, a Lagrangian-based heuristic consisting in solving a series of minimum network cost flow problems is attractive for practical implementation, both for the quality of the solutions and for the small computational times.     

Adaptive Localized QoS-Constrained Data Aggregation and Processing in Distributed Sensor Networks

In this paper, an efficient Quality of Service (QoS)-constrained data aggregation and processing approach for distributed wireless sensor networks is investigated and analyzed. One of the key features of the proposed approach is that the task QoS requirements are taken into account to determine when and where to perform the aggregation in a distributed fashion, based on the availability of local only information. Data aggregation is performed on the fly at intermediate sensor nodes, while at the same time the end-to-end latency constraints are satisfied. Furthermore, a localized adaptive data collection algorithm performed at the source nodes is developed that balances the design tradeoffs of delay, measurement accuracy, and buffer overflow, for given QoS requirements. The performance of the proposed approach is analyzed and evaluated, through modeling and simulation, under different data aggregation scenarios and traffic loads. The impact of several design parameters and tradeoffs on various critical network and application related performance metrics, such as energy efficiency, network lifetime, end-to-end latency, and data loss are also evaluated and discussed.     

Media Flow Rate Allocation in Multipath Networks

We address the problem of joint path selection and source rate allocation in order to optimize the media specific quality of service in streaming of stored video sequences on multipath networks. An optimization problem is proposed in order to minimize the end-to-end distortion, which depends on video sequence dependent parameters, and network properties. An in-depth analysis of the media distortion characteristics allows us to define a low complexity algorithm for an optimal flow rate allocation in multipath network scenarios. In particular, we show that a greedy allocation of rate along paths with increasing error probability leads to an optimal solution. We argue that a network path shall not be chosen for transmission, unless all other available paths with lower error probability have been chosen. Moreover, the chosen paths should be used at their maximum available end-to-end bandwidth. Simulation results show that the optimal flow rate allocation carefully adapts the total streaming rate and the number of chosen paths, to the end-to-end transmission error probability. In many scenarios, the optimal rate allocation provides more than 20% improvement in received video quality, compared to heuristic-based algorithms. This motivates its use in multipath networks, where it optimizes media specific quality of service, and simultaneously saves network resources at the price of a very low computational complexity.     

Slack allocation algorithm for parallel machines

Energy consumption is a critical issue in parallel and distributed embedded systems. We present a novel algorithm for energy efficient scheduling of Directed Acyclic Graph (DAG) based applications on Dynamic Voltage Scaling (DVS) enabled systems. Experimental results show that our algorithm provides near optimal solutions for energy minimization with considerably smaller computational time and memory requirements compared to an existing algorithm that provides near optimal solutions.     

A case for centrally controlled wireless sensor networks

In this article we present the Intelligent, Manageable, Power-Efficient and Reliable Internetworking Architecture (IMPERIA), a centrally managed architecture for large-scale wireless sensor networks (WSNs). We discuss the advantages of a centralized management over distributed approaches and derive our design by rigorously minimizing the amount of state information on individual sensor nodes and all sources of message collision during network operations. The result is a clustered multi-hop TDMA protocol that globally synchronizes the network and collects data at ultra-low power consumption.We present the end-to-end architecture and detail the algorithms we developed for (a) efficient network topology discovery and link quality estimation, (b) combined routing and clustering for pre-defined basestations, and (c) the scheduling of the medium access for multi-cluster and multi-channel data collection.IMPERIA has been implemented on TinyOS and IBM’s Mote Runner and successfully deployed in applications for vibration sensing as well as datacenter energy management. This article summarizes the performance results from simulations, laboratory experiments, and deployment measurements that support our design decisions.     

Online Feature Selection of Class Imbalance via PA Algorithm

Imbalance classification techniques have been frequently applied in many machine learning application domains where the number of the majority (or positive) class of a dataset is much larger than that of the minority (or negative) class. Meanwhile, feature selection (FS) is one of the key techniques for the high-dimensional classification task in a manner which greatly improves the classification performance and the computational efficiency. However, most studies of feature selection and imbalance classification are restricted to off-line batch learning, which is not well adapted to some practical scenarios. In this paper, we aim to solve high-dimensional imbalanced classification problem accurately and efficiently with only a small number of active features in an online fashion, and we propose two novel online learning algorithms for this purpose. In our approach, a classifier which involves only a small and fixed number of features is constructed to classify a sequence of imbalanced data received in an online manner. We formulate the construction of such online learner into an optimization problem and use an iterative approach to solve the problem based on the passive-aggressive (PA) algorithm as well as a truncated gradient (TG) method. We evaluate the performance of the proposed algorithms based on several real-world datasets, and our experimental results have demonstrated the effectiveness of the proposed algorithms in comparison with the baselines.     

A generalized model for investigating scheduling schemes in computational clusters

In this paper, we present a generalized model for the performance evaluation of scheduling compute-intensive jobs with unknown service times in computational clusters. We propose the application of parameters defined in the SPECpower_ssj2008 benchmark of the Standard Performance Evaluation Corporation to construct a performance evaluation model. In addition, we also define a method to rank physical servers based on either the high performance priority or the energy efficiency priority, and measures to characterize the performance of computational clusters.We investigate three schemes (separate queue, class queue and common queue) for buffering jobs in a computational cluster that is built from Commercial Off-The-Shelf (COTS) servers. Numerical results show that the buffering schemes do not have impact on performance measures related to the energy consumption of the investigated cluster. However, the buffering schemes play an important role in ensuring the quality of service parameters such as the waiting time and the response time experienced by arriving jobs. Furthermore, Dynamic Voltage and Frequency Scaling should be carefully applied to reduce the energy consumption of computational clusters.     

Opportunity-Based Topology Control in Wireless Sensor Networks

Topology control is an effective method to improve the energy efficiency of wireless sensor networks (WSNs). Traditional approaches are based on the assumption that a pair of nodes is either "connected” or "disconnected.” These approaches are called connectivity-based topology control. In real environments, however, there are many intermittently connected wireless links called lossy links. Taking a succeeded lossy link as an advantage, we are able to construct more energy-efficient topologies. Toward this end, we propose a novel opportunity-based topology control. We show that opportunity-based topology control is a problem of NP-hard. To address this problem in a practical way, we design a fully distributed algorithm called CONREAP based on reliability theory. We prove that CONREAP has a guaranteed performance. The worst running time is O(vert Evert ), where E is the link set of the original topology, and the space requirement for individual nodes is O(d), where d is the node degree. To evaluate the performance of CONREAP, we design and implement a prototype system consisting of 50 Berkeley Mica2 motes. We also conducted comprehensive simulations. Experimental results show that compared with the connectivity-based topology control algorithms, CONREAP can improve the energy efficiency of a network up to six times.     

Data aggregation framework for energy-efficient WirelessHART networks

WirelessHART has become an industrial standard for robust and real-time wireless monitoring and control. While energy-efficiency is one of the key design considerations for networks with battery-operated devices, data aggregation has been widely studied in the wireless sensor network (WSN) environments to reduce the traffic and prolong the lifetime of the network. However, existing data aggregation techniques cannot be applied directly to WirelessHART networks due to the multi-channel Time Synchronized Mesh Protocol (TSMP) and the superframe-based communication slot scheduling in WirelessHART. In this work, we propose a data aggregation framework for energy-efficient and real-time WirelessHART communication. In particular, we consider aggregation as a factor during the link selection procedure of the graph routing to improve the chance of data fusion at intermediate routing nodes and reduce the total number of message transmissions. Furthermore, a greedy-based heuristic is applied during the superframe construction phase to allocate package transmissions whose data can be aggregated at intermediate routing nodes into nearby time slots. During the superframe re-scheduling, we make sure that the predefined end-to-end deadline for each package is satisfied as long as the entire network is schedulable without data aggregation. Experimental results show that compared with existing WirelessHART routing algorithms, our proposed framework has significantly improvement on the energy saving and prolongs the overall lifetime of the network.     

WSNs clustering based on semantic neighborhood relationships

We propose a semantic clustering model based on a fuzzy inference system to find out the semantic neighborhood relationships in wireless sensor networks in order to both reduce energy consumption and improve the data accuracy. As a case study we describe a structural health monitoring application which was used to illustrate and assess the proposed model. We conduct experiments in order to evaluate the proposal in two different scenarios of damage with different data aggregation methods. We also compared our proposal, using the same data set, with a deterministic clustering method and with the LEACH algorithm. The results indicate that our approach is an energy-efficient clustering method for WSNs, outperforming both the deterministic clustering and LEACH algorithms in about 70% and 47% of energy savings respectively. The energy saving comes from the fact that we have a more efficient in-network data aggregation process since by exploiting the semantic relation between sensor nodes we can potentially aggregate more similar data and consequently, decrease the data redundancy (thus minimizing transmissions). Nodes that are semantically unrelated can operate in low-duty cycle, further reducing the energy consumption. Moreover, our proposal has the potential to improve the data accuracy provided for the application where accuracy is a QoS requirement in typical WSN applications.     

A survey of key management schemes in wireless sensor networks

Wireless sensor networks have many applications, vary in size, and are deployed in a wide variety of areas. They are often deployed in potentially adverse or even hostile environment so that there are concerns on security issues in these networks. Sensor nodes used to form these networks are resource-constrained, which make security applications a challenging problem. Efficient key distribution and management mechanisms are needed besides lightweight ciphers. Many key establishment techniques have been designed to address the tradeoff between limited memory and security, but which scheme is the most effective is still debatable. In this paper, we provide a survey of key management schemes in wireless sensor networks. We notice that no key distribution technique is ideal to all the scenarios where sensor networks are used; therefore the techniques employed must depend upon the requirements of target applications and resources of each individual sensor network.     

EClass: An execution classification approach to improving the energy-efficiency of software via machine learning

Energy efficiency at the software level has gained much attention in the past decade. This paper presents a performance-aware frequency assignment algorithm for reducing processor energy consumption using Dynamic Voltage and Frequency Scaling (DVFS). Existing energy-saving techniques often rely on simplified predictions or domain knowledge to extract energy savings for specialized software (such as multimedia or mobile applications) or hardware (such as NPU or sensor nodes). We present an innovative framework, known as EClass, for general-purpose DVFS processors by recognizing short and repetitive utilization patterns efficiently using machine learning. Our algorithm is lightweight and can save up to 52.9% of the energy consumption compared with the classical PAST algorithm. It achieves an average savings of 9.1% when compared with an existing online learning algorithm that also utilizes the statistics from the current execution only. We have simulated the algorithms on a cycle-accurate power simulator. Experimental results show that EClass can effectively save energy for real life applications that exhibit mixed CPU utilization patterns during executions. Our research challenges an assumption among previous work in the research community that a simple and efficient heuristic should be used to adjust the processor frequency online. Our empirical result shows that the use of an advanced algorithm such as machine learning can not only compensate for the energy needed to run such an algorithm, but also outperforms prior techniques based on the above assumption.     

Processing approximate aggregate queries in wireless sensor networks

In-network data aggregation has been recently proposed as an effective means to reduce the number of messages exchanged in wireless sensor networks. Nodes of the network form an aggregation tree, in which parent nodes aggregate the values received from their children and propagate the result to their own parents. However, this schema provides little flexibility for the end-user to control the operation of the nodes in a data sensitive manner. For large sensor networks with severe energy constraints, the reduction (in the number of messages exchanged) obtained through the aggregation tree might not be sufficient. In this paper, we present new algorithms for obtaining approximate aggregate statistics from large sensor networks. The user specifies the maximum error that he is willing to tolerate and, in turn, our algorithms program the nodes in a way that seeks to minimize the number of messages exchanged in the network, while always guaranteeing that the produced estimate lies within the specified error from the exact answer. A key ingredient to our framework is the notion of the residual mode of operation that is used to eliminate messages from sibling nodes when their cumulative change to the computed aggregate is small. We introduce two new algorithms, based on potential gains, which adaptively redistribute the error thresholds to those nodes that benefit the most and try to minimize the total number of transmitted messages in the network. Our techniques significantly reduce the number of messages, often by a factor of 10 for a modest 2% relative error bound, and consistently outperform previous techniques for computing approximate aggregates, which we have adapted for sensor networks.     

Artificial life techniques for load balancing in computational grids

Load balancing is a very important and complex problem in computational grids. A computational grid differs from traditional high performance computing systems in the heterogeneity of the computing nodes and communication links, as well as background workloads that may be present in the computing nodes. There is a need to develop algorithms that could capture this complexity yet can be easily implemented and used to solve a wide range of load balancing scenarios. Artificial life techniques have been used to solve a wide range of complex problems in recent times. The power of these techniques stems from their capability in searching large search spaces, which arise in many combinatorial optimization problems, very efficiently. This paper studies several well-known artificial life techniques to gauge their suitability for solving grid load balancing problems. Due to their popularity and robustness, a genetic algorithm (GA) and tabu search (TS) are used to solve the grid load balancing problem. The effectiveness of each algorithm is shown for a number of test problems, especially when prediction information is not fully accurate. Performance comparisons with Min-min, Max-min, and Sufferage are also discussed.     

A computational comparison of scaling techniques for linear optimization problems on a graphical processing unit

Preconditioning techniques are important in solving linear problems, as they improve their computational properties. Scaling is the most widely used preconditioning technique in linear optimization algorithms and is used to reduce the condition number of the constraint matrix, to improve the numerical behavior of the algorithms and to reduce the number of iterations required to solve linear problems. Graphical processing units (GPUs) have gained a lot of popularity in the recent years and have been applied for the solution of linear optimization problems. In this paper, we review and implement ten scaling techniques with a focus on the parallel implementation of them on GPUs. All these techniques have been implemented under the MATLAB and CUDA environment. Finally, a computational study on the Netlib set is presented to establish the practical value of GPU-based implementations. On average the speedup gained from the GPU implementations of all scaling methods is about 7×.     

Energy-efficient cooperative data aggregation for wireless sensor networks

Recently, cooperative communication mechanism is shown to be a promising technology to improve the transmit diversity only by a single transceiver antenna. Using this communication paradigm, multiple source nodes are able to coordinate their transmissions so as to obtain energy savings. As data aggregation is one of the most important operations in wireless sensor networks, this paper studies the energy-efficient data aggregation problem through cooperative communication. We first define the cooperative data aggregation (CDA) problem, and formally prove that this problem is NP-Hard. Due to the difficult nature of this problem, we propose a heuristic algorithm MCT for cooperative data aggregation. The theoretical analysis shows that this algorithm can reach the approximate performance ratio of 2. Moreover, the distributed implementation DMCT of the algorithm is also described. We prove that both centralized and distributed algorithms can construct the same topology for cooperative data aggregation. The experimental simulations show that the proposed algorithms will decrease the power consumption by about 12.5% and 66.3% compared with PEDAP and PEGASIS algorithms respectively.     

Systematic performance modeling and characterization of heterogeneous IP networks

Accurate measurement and modeling of IP networks is essential for network design, planning, and management. Efforts are being made to detect the state of the network from end-to-end measurements using different techniques and paradigms. In this paper we propose a novel concept to use in the modeling of real network scenarios under measurement and analysis. We called this new concept Service Condition. We explain our proposal's motivations and we use some simple examples to show how to apply the Service Condition concept to the study of real heterogeneous network scenarios. To show the real applicability of our proposal, preliminary results from a performance evaluation study over real heterogeneous networks (where the integration of LAN, WLAN, ADSL, UMTS, and GPRS is present) are given.