Summary In previous work, we introduced a distributed collections library for the APGAS for Java programming model. This library makes it possible for programmers to develop complex distributed programs thanks to the many abstractions and computation patterns supported. In particular, programmers can fully and dynamically change the distribution of data entries through high‐level abstractions. However, the problem of balancing the load between processes remains, especially in cases where multiple processes may be concurrently executing on a single host, or when the performance of the hosts used differs. To address this issue and to relieve the burden of programming a load balancing strategy for a specific application, we created a dynamic load balancer integrated into our library. This load balancer operates within a specific context in a manner which does not interfere with the program legibility. Internally, we implement a scheme inspired by the lifeline‐based global load balancer scheme first introduced in 10. We evaluate the performance of our integrated load balancer on a small‐scale Beowulf cluster.

Summary Global load balancer libraries should be easy to use and allow users to easily obtain good performance for their applications on a variety of distributed systems. In this article, we introduce a new tuning mechanism to our Java implementation of the lifeline‐based global load balancer which automatically adjusts the task granularity to reach good performance based on some selected runtime metrics. We evaluate our system against four backtrack‐search problems on both a many‐core supercomputer environment and on a beowulf server, achieving ideal performance with our tuning mechanism on the supercomputer. We also identify the limits of our mechanism in handling situations with reduced imbalance.

Localization using low-power wireless devices is one of the promising methods even in outdoor environments because the Global Navigation Satellite System (GNSS) devices generally consume much power to result in frequent battery replacements. Received Signal Strength Indicator (RSSI) is often utilized in such methods due to availability in most wireless devices. Even in outdoor open space, however, various factors such as path-loss characteristics and antenna directivity possibly cause difficulty in RSSI-based localization. This paper investigates the error factors of Deep Neural Network (DNN)-based fingerprinting localization in terms of antenna directivity. Most existing studies have analyzed the error factor in localization algorithms or methods to improve accuracy. On the other hand, the error factors on the environments had not been paid attention yet at all. In this paper, we investigate distance errors of the DNN-based fingerprinting method in terms of antenna directivity as one of the environmental factors using virtual space reproduced in a computer. Through experiments where an actual Bluetooth Low Energy (BLE) tag system was emulated, we found that the tag antenna directivity could be the factor to worsen the localization accuracy while receiver antenna directivity to better.

The importance of real-time notification has been growing for social services and Intelligent Transporting System (ITS). As an advanced version of Pub/Sub systems, publish-process-subscribe systems, where published messages are spooled and processed on edge servers, have been proposed to achieve data-driven intelligent notifications. In this paper, we present a system that allows a topic to be managed on multiple edge servers so that messages are processed near publishers and subscribers, even when publishers are spread over a wide area. However, permitting a topic to use multiple edge servers can cause exhaustion of resources as the number of topics and consumed resources by each topic increase. We formulate the delay caused by the depletion of edge server resources to find the optimal allocation of topics on limited edge server resources. As the optimization problem is NP-hard, we propose heuristics leveraging the given locality and the pub/sub relationships observed between clients to use the edge server resources efficiently while delivering realtime notifications. The experimental results demonstrate that the proposed method reduces the delay to deliver notifications and the effectiveness of the strategy exploiting the relationships between clients.

In recent years, wireless mesh networks (WMNs) have become popular because of their wide coverage areas and high fault tolerance. Although unlicensed-band off-the-shelf wireless devices based on IEEE 802.11 are widely deployed, WMNs based on IEEE 802.11 suffer from degradation of throughput owing to the hidden node problem. To mitigate this problem, in this study, we developed a new routing metric called the “Asymmetric Hidden node problem Aware routing Metric (AHAM).” AHAM enables a new flow to establish a source-destination path that largely avoids not only interference with hidden nodes of existing flows but also making hidden nodes on that path interfere with existing flows. This is because the hidden node problem can asymmetrically impact throughput between flows when new flows are established. As a response, AHAM aims to estimate the influence of hidden nodes accurately and compensate for how both the new and existing flows influence each other. Simulation results showed that AHAM realizes high throughput while maintaining network fairness compared with other metrics.

The field of parallel computing has experienced an increase in the number of computing nodes and parallel computing has widened its application to include computations that have irregular features. Some parallel programming languages handle object data structures and offer marshaling/unmarshaling mechanisms to transport them. To manage data elements spread over computing nodes, some research on distributed collections has been conducted. This study proposes a distributed collection library that can handle multiple collections of object elements and change their distributions while maintaining the associativity between their elements. This library is implemented on an object-oriented parallel programming language, X10. We suppose pairs of associative collections such as vehicles and streets in a traffic simulation. When many vehicles are concentrated on streets assigned to certain computing nodes, some of those streets should be moved to other nodes. Our library supports the programmer in easily distributing the associative collections over the computing nodes and re-allocating their elements while maintaining the data sharing relationship among associative elements. The programmer can describe the associativity between objects using both declarative and procedural methods.

Recent years have witnessed the emergence of Web services that publish data that can be used as globally shared data resources, and Web data stores offer data resources that can be shared and modified by users or groups. We think associative data representation among Web data elements is important for the reuse of the data for various purposes even when treating personal/group data. This paper aims to provide a general-purpose widget for Web data browsing/editing where users can easily create and update their data elements that are associated with other data elements, and proposes a table editor for joined views of Web services. With our editor, the user can join multiple Web services, and edit the left/right elements on the joined view in a tabular grid. To support easy manipulation of associated data during the modification, our editor continuously provides the renewed view reflecting the modification. The user can also try out multiple modifications while checking their impact on the view, and decide later whether to commit the modifications to the Web data stores or to discard the modifications. This paper presents some illustrations that demonstrate the use of this joined view editor, discusses the design, and briefly describes its implementation.

This paper proposes a mashup framework for creating flexible mashup applications in which the user can selectively browse through mashup items. Our framework provides a data management engine for on-demand data generation, and GUI components called widgets that can be used to browse through mashed-up data selectively. The application developer has to only prepare a mashup relation specifying the web service combinations and widget configurations specifying how to display the mashed-up data. On the basis of these configurations, widgets monitor user interactions and requests data from the data management engine that processes the demand-driven creation of mashed-up data. To enable selective browsing, a table widget, for instance, allows selection of columns to be displayed, provides a limited view with scroll bars, and filtering facilities. Our framework also offers a mechanism for widget coordination where a widget can change the display target according to states or events of other widgets. We introduce a sample application for tour planning using five cooperative widgets, and discuss the usability and performance advantages of our framework.

This paper proposes a programming environment for Java that processes network XML data in a demand-driven manner to return quick initial responses. Our system provides a data binding tool and a tree operation package, and the programmer can easily handle network XML data as tree-based operations using these facilities. For efficiency, demand-driven data binding allows the application to start the processing of a network XML document before the arrival of the whole data, and our tree operators are also designed to start the calculation using the initially accessible part of the input data. Our system uses multithread technology for implementation with optimization techniques to reduce runtime overheads. It can return initial responses quickly, and often shortens the total execution time due to the effects of latency hiding and the reduction of memory usage. Compared with an ordinary tree-based approach, our system shows a highly improved response and a 1-28% reduction of total execution time on the benchmark programs. It only needs 1-4% runtime overheads against the event-driven programs.

On distributed-memory high-performance massively parallel computers (MPPs) where processors are interconnected by an asynchronous network, efficient garbage collection (GC) becomes difficult, due to inter-node references and references within pending, unprocessed messages. Our parallel global GC algorithm (1) takes advantage of reference locality, (2) efficiently traverses references over nodes, (3) admits a minimum pause time for the ongoing computations, and (4) has been shown to scale up to 1024-node MPPs. The algorithm employs a global weight counting scheme to substantially reduce message traffic. Two methods for confirming the arrival of pending messages are used: one counts the number of messages and the other uses network 'bulldozing'. Performance evaluations in actual implementations on a multicomputer with from 32 to 1024 nodes, the Fujitsu AP1000, reveals various favorable properties of the algorithm.>