Global Data Space Research Articles

Global Virtual Data Space for Unified Data Access Across Supercomputing Centers

In the wide-area high-performance computing environment, heterogeneous storage resources are geographically distributed in different supercomputing centers, which leads to the barriers between applications and data. This paper proposes a global virtual data space, named GVDS, to meet the needs of unified data access across supercomputing centers. GVDS integrates the parallel/distributed file systems of supercomputing centers to present a virtual space with tremendous storage capability for users. GVDS organizes users into groups for easy management, which allows users to share, collaborate, and perform computations on the stored data. For failure tolerance, global metadata is replicated and distributed on multiple supercomputing centers, redundant I/O service components are deployed in each supercomputing center. GVDS uses adaptive prefetching, caching, and request merging to improve access performance. Experimental results running on real-world supercomputing centers show that, GVDS can deliver excellent I/O performance running micro-benchmark, real-world traces and applications.

The Vision of the FAIR Digital Object Machine and Ubiquitous FDO Services

The Vision of the FAIR Digital Object Machine and Ubiquitous FDO Services

Editor’s Note

Editor’s Note

Unsupervised Domain Adaptation Using Robust Class-Wise Matching

Unsupervised domain adaptation (DA) enables a classifier trained on data from one domain to be applied to data from another without labels. Given that the key to transferring a classifier across domains is to mitigate the data distribution mismatch for each class, most previous works completely or partially focus on global distribution matching across domains. The global data space, however, can be complicated, which makes modeling the global distribution difficult. To mitigate this problem, we present a novel unsupervised DA framework where the DA problem is addressed by proposing a robust class-wise matching strategy. Specifically, through minimizing a maximum mean discrepancy-based class-wise fisher discriminant across domains, this framework jointly optimizes two modules: a transferable feature learning module that reduces the distribution discrepancy between the same classes as well as increasing the distribution discrepancy between different classes across domains by a linear projection, and a robust classifier that exploits both the supervised information in source domain and the unsupervised low-rank property of target domain. In experiments on three DA benchmark data sets, the proposed framework shows the state-of-the-art performance.

Student Inquiry, Networks of Knowledge and Linked Data

This paper explores the potential for the development of new learning opportunities in higher education, through students being conceptualised not as consumers, recipients or commodities, but rather as co-researchers and co-producers of knowledge. Specifically, it discusses the implications of new forms of networked knowledge enabled by the emergence of semantic web and linked data technologies and the reconceptualising of the Internet as a ‘global data space’. We draw on our experience of initiating and supporting a range of projects in UK higher education in the course of an extended programme of research and development. Some of these involved the design and development of new technology platforms, while others were focussed on the redevelopment of taught courses, assignments and assessed activities. What these projects had in common is that they all took place in the context of complex learning settings in which some variety of case based learning is used. They involved students drawn from different disciplines in higher education in ‘research-based learning’ about curriculum contexts, and about pedagogical aspects of these contexts. New digital tools were developed in the form of rich web applications that allowed learner interaction with content, in many cases underpinned by data from multiple sources and in diverse formats. In the development of these online technologies, students located, analysed, synthesised and, in some cases, generated new data, and, perhaps more significantly, participated in local or global knowledge networks. What we will argue is that these types of projects involve not only the development of specific techno-literacies, but also that they form the basis of broader ‘critical digital literacies’. These in turn equip students to enter workplaces better positioned to inquire into the particularities of the educational settings in which they work and the practices in which they are engaged. They can thus undertake ‘counter-research’ in which dominant rhetorics are challenged, and evidence bases for policy and practice are subjected to scrutiny, critique and reinterpretation. This presents the potential for students to undertake critical and politicised inquiry as part of a broader reframing of the purposes of higher education.

A Global Data Ecosystem for Agriculture and Food

Agriculture would benefit hugely from a common data ecosystem. Produced and used by diverse stakeholders, from smallholders to multinational conglomerates, a shared global data space would help build the infrastructures that will propel the industry forward. In light of growing concern that there was no single entity that could make the industry-wide change needed to acquire and manage the necessary data, this paper was commissioned by Syngenta with GODAN’s assistance to catalyse consensus around what form a global data ecosystem might take, how it could bring value to key players, what cultural changes might be needed to make it a reality and finally what technology might be needed to support it. This paper looks at the challenges and principles that must be addressed in in building a global data ecosystem for agriculture. These begin with building incentives and trust – amongst both data providers and consumers – in sharing, opening and using data. Key to achieving this will be developing a broad awareness of, and making efforts to improve, data quality, provenance, timeliness and accessibility. We set out the key global standards and data publishing principles that can be followed in supporting this, including the ‘Five stars of open data’ and the ‘FAIR principles’ and offer several recommendations for stakeholders in the industry to follow. • Finding business models that provide incentives for various entities to collect and share data. If these models provide business value directly to the data providers, the quality of the collected data will be higher. • Leading by example by providing open data sources. Syngenta has already done this by publishing data about the results of its Good Growth Plan. • Encouraging data standards that make it easier to produce and share data. In doing so, stakeholders will need to have reasonable expectations of how these standards will be used. • Automating data collection. Automatically collected data is more likely to be accurate and precise than data collected by hand. • Annotating datasets. Even automatically collected data cannot be used if it is not described in a consistent and understandable way. • Following data sharing principles. The five-star maturity model and the FAIR principles provide guidelines for creating and sharing data. • Using the data. All of the best data sharing efforts have little impact if the data is not used in a productive way. Stakeholders must encourage a cottage industry of data-backed apps that get the most value from datasets. In response to the paper, four key organisations engaged in developing data infrastructures (CGIAR, Agrimetrics, ODI and AgroKnow) state what a common data ecosystem would mean to them. Original DOI: 10.1079/CABICOMM-79-12

The Opportunity of Linked Data for the European Higher Education Area

The european higher education area (EHEA) aims to create a more comparable, compatible and coherent space of Higher Education in Europe. After a decade of reforms, most of the required instruments are already in place, so the next years are key to complete the implementation and consolidation of the EHEA. However, data challenges pose a threat to pivotal EHEA processes such as credit recognition that require institutions to cooperate and integrate their datasets scattered across diverse repositories in multiple data formats. Linked Data principles and practices have the potential to address such difficulties, as demonstrated with the emergence of a global data space –the Web of Data. This paper first discusses the benefits of Linked Data for the emergent EHEA, devising a new set of Linked Data-driven scenarios that leverage current EHEA processes. Further, a set of preliminary Linked Data deployments is examined to assess the advantages and identify the challenges ahead in the EHEA landscape.

Federated query processing on linked data: a qualitative survey and open challenges

AbstractA large number of data providers publish and connect their structured data on the Web as linked data. Thus, the Web of data becomes a global data space. In this paper, we initially give an overview of query processing approaches used in this interlinked and distributed environment, and then focus on federated query processing on linked data. We provide a detailed and clear insight on data source selection, join methods and query optimization methods of existing query federation engines. Furthermore, we present a qualitative comparison of these engines and give a complementary comparison of the measured metrics of each engine with the idea of pointing out the major strengths of each one. Finally, we discuss the major challenges of federated query processing on linked data.

Using a data-centric event-driven architecture approach in the integration of real-time systems at DTP2

Integration of heterogeneous and distributed systems is a challenging task, because they might be running on different platforms and written with different implementation languages by multiple organizations. Data-centricity and event-driven architecture (EDA) are concepts that help to implement versatile and well-scaling distributed systems.This paper focuses on the implementation of inter-subsystem communication in a prototype distributed remote handling control system developed at Divertor Test Platform 2 (DTP2). The control system consists of a variety of heterogeneous subsystems, including a client–server web application and hard real-time controllers. A standardized middleware solution (Data Distribution Services (DDS)) that supports a data-centric EDA approach is used to integrate the system.One of the greatest challenges in integrating a system with a data-centric EDA approach is in defining the global data space model. The selected middleware is currently only used for non-deterministic communication. For future application, we evaluated the performance of point-to-point communication with and without the presence of additional network load to ensure applicability to real-time systems. We found that, under certain limitations, the middleware can be used for soft real-time communication. Hard real-time use will require more validation with a more suitable environment.

Measuring the overhead of Intel C++ Concurrent Collections over Threading Building Blocks for Gauss–Jordan elimination

SUMMARYThe most efficient way to parallelize computation is to build and evaluate the task graph constrained only by the data dependencies between the tasks. Both Intel's C++ Concurrent Collections (CnC) and Threading Building Blocks (TBB) libraries allow such task‐based parallel programming. CnC also adapts the macro data flow model by providing only single‐assignment data objects in its global data space. Although CnC makes parallel programming easier, by specifying data flow dependencies only through single‐assignment data objects, its macro data flow model incurs overhead. Intel's C++ CnC library is implemented on top of its C++ TBB library. We can measure the overhead of CnC by comparing its performance with that of TBB. In this paper, we analyze all three types of data dependencies in the tiled in‐place Gauss–Jordan elimination algorithm for the first time. We implement the task‐based parallel tiled Gauss–Jordan algorithm in TBB using the data dependencies analyzed and compare its performance with that of the CnC implementation. We find that the overhead of CnC over TBB is only 12%– 15% of the TBB time, and CnC can deliver as much as 87%– 89% of the TBB performance for Gauss–Jordan elimination, using the optimal tile size. Copyright © 2012 John Wiley & Sons, Ltd.

Система интеграции данных на основе наборов RDF-связей пространства Linked Open Data

This paper describes research-in-progress work on data integration system in Linked Open Data space. Proposed system uses concept of RDF identity links to interlink heterogeneous local data sources and integrate them into global data space.

Linked Data: Evolving the Web into a Global Data Space

The World Wide Web has enabled the creation of a global information space comprising linked documents. As the Web becomes ever more enmeshed with our daily lives, there is a growing desire for direct

Linked Data - The Story So Far

The term “Linked Data” refers to a set of best practices for publishing and connecting structured data on the Web. These best practices have been adopted by an increasing number of data providers over the last three years, leading to the creation of a global data space containing billions of assertions— the Web of Data. In this article, the authors present the concept and technical principles of Linked Data, and situate these within the broader context of related technological developments. They describe progress to date in publishing Linked Data on the Web, review applications that have been developed to exploit the Web of Data, and map out a research agenda for the Linked Data community as it moves forward.