Aiming to ease the parallel programming for heterogeneous architectures, we propose and implement a high-level OpenCL runtime that conceptually merges multiple heterogeneous hardware devices into one virtual heterogeneous compute device (VHCD). Moreover, automated workload distribution among the devices is based on offline profiling, together with new programming directives that define the device-independent data access range per work-group. Therefore, an OpenCL program originally written for a single compute device can, after inserting a small number of programming directives, run efficiently on a platform consisting of heterogeneous compute devices. Performance is ensured by introducing the technique of virtual cache management, which minimizes the amount of host-device data transfer. Our new OpenCL runtime is evaluated by a diverse set of OpenCL benchmarks, demonstrating good performance on various configurations of a heterogeneous system.

Recently, Hadoop has attracted much attention from engineers and researchers as an emerging and effective framework for Big Data. HDFS (Hadoop Distributed File System) can manage a huge amount of data with high performance and reliability using only commodity hardware. However, HDFS requires a single master node, called a NameNode, to manage the entire namespace (or all the i-nodes) of a file system. This causes the SPOF (Single Point Of Failure) problem because the file system becomes inaccessible when the NameNode fails. This also causes a bottleneck of efficiency since all the access requests to the file system have to contact the NameNode. Hadoop 2.0 resolves the SPOF problem by introducing manual failover based on two NameNodes, Active and Standby. However, it still has the efficiency bottleneck problem since all the access requests have to contact the Active in ordinary executions. It may also lose the advantage of using commodity hardware since the two NameNodes have to share a highly reliable sophisticated storage. In this paper, we propose a new HDFS architecture to resolve all the problems mentioned above.

In recent years, wireless power transfer has been attracting a great deal of attention. In order to realize efficient power transfer, it is necessary to communicate data such as a frequency, required power, or error tolerance. In the proposed system, because of the use of the same antennas for power transmission and data transmission, the frequency response of a channel for the data transmission changes owing to load fluctuation and the distance between antennas. This paper investigates data transmission performance in the wireless power transfer system with frequency response estimation at the transmitter side. The numerical results obtained through computer simulation show that the proposed scheme can estimate the frequency response of the channel and the difference between the expected bit error rate (BER) and the BER with the estimation error is less than 0.5dB at the BER of 10-3.

Most existing deterministic multithreading systems are costly on pipeline parallel programs due to load imbalance. In this letter, we propose a Load-Balanced Deterministic Runtime (LBDR) for pipeline parallelism. LBDR deterministically takes some tokens from non-synchronization-intensive threads to synchronization-intensive threads. Experimental results show that LBDR outperforms the state-of-the-art design by an average of 22.5%.

Video streaming uploading over vehicular ad hoc networks (VANETs) can support many interesting applications. Due to the high mobility and dynamic topology of VANETs, how to support video streaming using wireless communications between vehicles and road-side access points still remains an open issue. In this paper, we propose a geographical uploading scheme, called MPVUS, which uses the moving prediction to keep the stable forwarding and reduce the high link failure probability over VANETs. The scheme also decides the AP switch opportunity by traffic flow estimation, so as to adjust the forwarding direction timely to avoid the short-sighted switch decision. Simulation results demonstrate the effectiveness of our scheme, which can achieve good performance in terms of the start-up delay, playback interruption ratio and video frame distortion.

The endurance time of smartphone still suffer from the limited battery capacity, and smartphone apps will increase the burden of the battery if they download large data over slow network. So how to manage the download tasks is an important work. To this end we propose a smartphone download strategy with low energy consumption which called CLSA (Concentrated Download and Low Power and Stable Link Selection Algorithm). The CLSA is intended to reduce the overhead of large data downloads by appropriate delay for the smartphone, and it based on three major factors: the current network situation, the length of download requests' queue and the local information of smartphone. We evaluate the CLSA using a music player implementation on ZTE V880 smartphone running the Android operation system, and compare it with the other two general download strategies, Minimum Delay and WiFi Only. Experiments show that our download algorithm can achieve a better trade-off between energy and delay than the other two.

This paper investigates the performance of an overloaded multiple-input multiple-output (MIMO) orthogonal frequency division multiplexing (OFDM) system with a repetition code. It has been demonstrated that diversity with block coding prevents the performance degradation induced by signal multiplexing. However, the computational complexity of a joint decoding scheme increases exponentially with the number of multiplexed signal streams. Thus, this paper proposes the use of a repetition code in the overloaded MIMO-OFDM system. In addition, QR decomposition with M-algorithm (QRM) maximum likelihood decoding (MLD) is applied to the decoding of the repetition code. QRM-MLD significantly reduces the amount of joint decoding complexity. In addition, virtual antennas are employed in order to increase the throughput that is reduced by the repetition code. It is shown that the proposed scheme reduces the complexity by about 1/48 for 6 signal streams with QPSK modulation while the BER degradation is less than 0.1dB at the BER of 10-3.

As the quantity of mobile application traffic keeps increasing, operators are facing the scalability limits of VoIP protocols. Higher queuing delays at the Session Initiation Protocol (SIP) server create significantly more retransmissions in the network. When the message arrival rate including retransmissions exceeds the message serving capacity of a SIP server, the queue size increases and eventually the SIP server can crash. Our analysis demonstrates that server crash can be prevented if the buffer size of the SIP server is limited. However, having smaller buffer sizes yields side effects such as lower successful transaction ratio for bursty traffic. In this paper, we propose a new SIP server scheduling mechanism in which the original incoming SIP requests have strict priority over the retransmitted requests. The priority based scheduling mechanism provides network administrator with the ability to configure the buffer size of a SIP server to a moderately high value. We implement the proposed priority-based scheduling mechanism in the JAIN-SIP stack and confirm that the implementation requires minimal changes to the SIP standard. Numerical experiments show that the proposed scheduling mechanism provides significantly and consistently better scalability at high buffer sizes compared to the heavily used first-in-first-out scheduling, thus enabling us to avoid server overloads.

Internet traffic classification is an essential step for stable service provision. The payload signature classifier is considered a reliable method for Internet traffic classification but is prohibitively computationally expensive for real-time handling of large amounts of traffic on high-speed networks. In this paper, we describe several design techniques to minimize the search space of traffic classification and improve the processing speed of the payload signature classifier. Our suggestions are (1) selective matching algorithms based on signature type, (2) signature reorganization using hierarchical structure and traffic locality, and (3) early packet sampling in flow. Each can be applied individually, or in any combination in sequence. The feasibility of our selections is proved via experimental evaluation on traffic traces of our campus and a commercial ISP. We observe 2 to 5 times improvement in processing speed against the untuned classification system and Snort Engine, while maintaining the same level of accuracy.

Recently, radar resource management of multifunction radar is a challenging issue in electronically scanned array radar technology. This paper deals with radar beam scheduling, which is a core issue of radar resource management. This paper proposed stochastic scheduler algorithm using Simulated Annealing (SA) and Hybrid scheduler algorithm which automatically selects two different types of schedulers according to the radar load: Rule based scheduler using modified Butler algorithm for underload situations and SA based scheduler for overload situations. The proposed algorithms are evaluated in terms of scheduling latency, the number of scheduled tasks, and time complexity. The simulation results show that the performance of rule based scheduler is seriously degraded in overload situation. However, SA based scheduler and Hybrid scheduler have graceful performance degradation in overload situation. Compared with rule based scheduler, SA based scheduler and Hybrid scheduler can schedule many more tasks on time for the same operation duration in the overload situation. Even though their time complex is relatively high, it can be applied to real applications if the parameters are properly controlled. Especially, Hybrid scheduler has an advantage of low time complexity with good performance.

At present, vast numbers of information resources are available on the Internet. However, one emerging issue is how to search and exploit these information resources in an efficient and flexible manner with high scalability. In this study, we focused our attention on the design of a distributed hash table (DHT)-based search system that supports efficient scalable multi-attribute queries of information resources in a distributed manner. Our proposed system, named D-AVTree, is built on top of a ring-based DHT, which partitions a one-dimensional key space across nodes in the system. It utilizes a descriptive naming scheme, which names each resource using an attribute-value (AV) tree, and the resource names are mapped to d-bit keys in order to distribute the resource information to responsible nodes based on a DHT routing algorithm. Our mapping scheme maps each AV branch of a resource name to a d-bit key where AV branches that share a subsequence of AV pairs are mapped to a continuous portion of the key space. Therefore, our mapping scheme ensures that the number of resources distributed to a node is small and it facilitates efficient multi-attribute queries by querying only a small number of nodes. Further, the scheme has good compatibility with key-based load balancing algorithms for DHT-based networks. Our system can achieve both efficiency and a good degree of load balancing even when the distribution of AV pairs in the resource names is skewed. Our simulation results demonstrated the efficiency of our solution in terms of the distribution cost, query hit ratio, and the degree of load balancing compared with conventional approaches.

This letter proposes a wideband compact DC block design technique. This DC block has a wide pass-band and wide stop-band and transforms termination impedances. It comprises a pair of coupled lines on a defected ground structure (DGS) with capacitor loading. A periodic DGS pattern increases coupling, and, consequently, a wideband DC block design is allowed with a microstrip process on a high dielectric low height substrate. A DC block with equal termination impedances of 50,$Omega$ and another that transforms 50 into 30,$Omega$ are fabricated. The measured fractional bandwidths are 48% and 47%. The size of the DC block is 16.8$ imes$ 15,mm$^2(0.057lambda_0 imes 0.051lambda_0)$.

This paper describes experiments on passive Multiple-Input Multiple-Output (MIMO) transmission with load modulation. PIN diodes are used as the variable impedance element at the tag side to realize multi-level modulation. The results indicate that the transmission rate of passive MIMO is up to 2 times higher than that of Single-Input Single-Output (SISO) with the same transmission power when the distance between the reader and the tag is 0.5m. Also, when the distance is 1m, MIMO offers up to 1.7 times higher transmission rate than SISO. These results indicate that passive MIMO offers high-speed data transmission even when the distance is doubled.

We propose an architecture for offloading processes in applications to support low-performance devices. Almost all applications based on standardized web technologies are compatible with our architecture. We discuss how interfaces should be used properly to offload processes in JavaScript and argue that an interface for offloading should only be used for defining complex processes. We also propose a method for applying our architecture to web applications that use web workers. Our method automatically offloads some worker processes to the cloud. We also compare the processing times achieved with and without our method. Our architecture exhibits good efficacy with regards to the N-Queen problem, although it is influenced by network latency between a device and the cloud.

This paper presents a low complexity joint decoding scheme of block coded signals in an overloaded multiple-input multiple-output (MIMO) orthogonal frequency division multiplexing (OFDM) system. In previous literature, a joint maximum likelihood decoding scheme of block coded signals has been evaluated through theoretical analysis. The diversity gain with block coding prevents the performance degradation induced by signal multiplexing. However, the computational complexity of the joint decoding scheme increases exponentially with the number of multiplexed signal streams. Thus, this paper proposes a two step joint decoding scheme for block coded signals. The first step of the proposed scheme calculates metrics to reduce the number of the candidate codewords using decoding based on joint maximum likelihood symbol detection. The second step of the proposed scheme carries out joint decoding on the reduced candidate codewords. It is shown that the proposed scheme reduces the complexity by about 1/174 for 4 signal stream transmission.

Mobility load balancing (MLB) is a key technology for self-organization networks (SONs). In this paper, we explore the mobility load balancing problem and propose a unified cell specific offset adjusting algorithm (UCSOA) which more accurately adjusts the largely uneven load between neighboring cells and is easily implemented in practice with low computing complexity and signal overhead. Moreover, we evaluate the UCSOA algorithm in two different traffic conditions and prove that the UCSOA algorithm can get the lower call blocking rates and handover failure rates. Furthermore, the interdependency of the proposed UCSOA algorithm's performance and that of the inter-cell interference coordination (ICIC) algorithm is explored. A self-organization soft frequency reuse scheme is proposed. It demonstrates UCSOA algorithm and ICIC algorithm can obtain a positive effect for each other and improve the network performance in LTE system.

Online Social Networks (OSNs) have recently been playing an important role in communication. From the audience aspect, they enable audiences to get unlimited information via the information feeding mechanism (IFM), which is an important part of the OSNs. The audience relies on the quantity and quality of the information served by it. We found that existing IFMs can result in two problems: information overload and cultural ignorance. In this paper, we propose a new type of IFM that solves these problems. The advantage of our proposed IFM is that it can filter irrelevant information with consideration of audiences' culture by using the Naïve Bayes (NB) algorithm together with features and factors. It then dynamically serves interesting and important information based on the current situation and preference of the audience. This mechanism helps the audience to reduce the time spent in finding interesting information. It can be applied to other cultures, societies and businesses. In the near future, the audience will be provided with excellent, and less annoying, communication. Through our studies, we have found that our proposed IFM is most appropriate for Thai and some groups of Japanese audiences under the consideration of audiences' culture.

The number of networked devices of sensors and actuators continues to increase. We are developing a data-sharing mechanism called uTupleSpace as middleware for storing and retrieving ubiquitous data that are input or output by such devices. uTupleSpace enables flexible retrieval of sensor data and flexible control of actuator devices, and it simplifies the development of various applications. Though uTupleSpace requires scalability against increasing amounts of ubiquitous data, traditional load-distribution methods using a distributed hash table (DHT) are unsuitable for our case because of the ununiformity of the data. Data are nonuniformly generated at some particular times, in some particular positions, and by some particular devices, and their hash values focus on some particular values. This feature makes it difficult for the traditional methods to sufficiently distribute the load by using the hash values. Therefore, we propose a new load-distribution method using a DHT called the dynamic-help method. The proposed method enables one or more peers to handle loads related to the same hash value redundantly. This makes it possible to handle the large load related to one hash value by distributing the load among peers. Moreover, the proposed method reduces the load caused by dynamic load-redistribution. Evaluation experiments showed that the proposed method achieved sufficient load-distribution even when the load was concentrated on one hash value with low overhead. We also confirmed that the proposed method enabled uTupleSpace to accommodate the increasing load with simple operational rules stably and with economic efficiency.

This paper presents a joint decoding scheme for the overloaded multiple input multiple output (MIMO)-orthogonal frequency division multiplexing (OFDM) system. In the overloaded MIMO system, the number of receive antenna elements is less than that of transmit antenna elements. It has been shown that under the overloaded condition the performance of joint detection deteriorates while diversity reduces the amount of performance degradation caused by signal multiplexing. Thus, this paper proposes a maximum likelihood joint decoding scheme of block coded signals in the overloaded MIMO-OFDM system. The performance of joint decoding over Rayleigh fading channels is evaluated through simulation and experiments. The simulation shows that the diversity through block coding prevents any performance degradation in the joint decoding of 2 Hamming coded signal streams. However, there are differences between numerical results obtained through computer simulation and experiments owing to channel estimation errors.

In wireless sensor networks, energy depletion of bottleneck nodes which have more data packets to relay than others, dominates the network lifetime referred to as the funnel effect problem. To overcome this problem, multiple sink methods have been proposed where sensor nodes send observed data packets toward several sinks to distribute traffic load of bottleneck nodes. If both of the topology and the traffic pattern are symmetric, bottleneck nodes are located near sinks. However, in a general sensor network with an asymmetric topology and/or an asymmetric traffic pattern, bottleneck nodes may exist any place in the network. In this paper, we propose DCAM (DispersiveCast of packets to Avoid bottleneck nodes for Multiple sink sensor network), which is a load balancing method to improve lifetime of a sensor network with an asymmetric topology and an asymmetric traffic pattern. DCAM first finds bottleneck nodes, and then balances the load on the bottleneck nodes. Selected nodes send data packets to several sinks dispersively according to some criteria. The criteria classify DCAM into three variations: DCAM with probability (DCAM-P), DCAM with moving boarder (DCAM-MB), and DCAM with round-robin (DCAM-RR). This paper gives details of the DCAM methods, and thereafter evaluates them with asymmetric topologies and asymmetric traffic patterns. To deal with these dynamic asymmetry, the topology is modeled by a grid network with virtual holes that are defined as vacant places of nodes in the network. Asymmetry of traffic pattern is modeled by defining a hot area where nodes have heavier data traffic than the others. The evaluations are conducted as changing hot-area traffic patterns as well as fixing hot-area patterns. The results show that DCAM improves network lifetime up to 1.87 times longer than the conventional schemes, (i.e., nearest sink transmissions and optimal dispersive cast of packet). We also discuss DCAM on several aspects such as overhead, energy consumption, and applications.