arXiv subjects

Patrick Valduriez

Publications and source records attributed to Patrick Valduriez.

21 records · Page 2Linked to original sources

arXiv2013

On Demand Memory Specialization for Distributed Graph Databases

In this paper, we propose the DN-tree that is a data structure to build lossy summaries of the frequent data access patterns of the queries in a distributed graph data management system. These compact representations allow us an efficient communication of the data structure in distributed systems. We exploit this data structure with a new \textit{Dynamic Data Partitioning} strategy (DYDAP) that assigns the portions of the graph according to historical data access patterns, and guarantees a small network communication and a computational load balance in distributed graph queries. This method is able to adapt dynamically to new workloads and evolve when the query distribution changes. Our experiments show that DYDAP yields a throughput up to an order of magnitude higher than previous methods based on cache specialization, in a variety of scenarios, and the average response time of the system is divided by two.

cs.DB↗

arXiv2011

Principles of Distributed Data Management in 2020?

With the advents of high-speed networks, fast commodity hardware, and the web, distributed data sources have become ubiquitous. The third edition of the Özsu-Valduriez textbook Principles of Distributed Database Systems [10] reflects the evolution of distributed data management and distributed database systems. In this new edition, the fundamental principles of distributed data management could be still presented based on the three dimensions of earlier editions: distribution, heterogeneity and autonomy of the data sources. In retrospect, the focus on fundamental principles and generic techniques has been useful not only to understand and teach the material, but also to enable an infinite number of variations. The primary application of these generic techniques has been obviously for distributed and parallel DBMS versions. Today, to support the requirements of important data-intensive applications (e.g. social networks, web data analytics, scientific applications, etc.), new distributed data management techniques and systems (e.g. MapReduce, Hadoop, SciDB, Peanut, Pig latin, etc.) are emerging and receiving much attention from the research community. Although they do well in terms of consistency/flexibility/performance trade-offs for specific applications, they seem to be ad-hoc and might hurt data interoperability. The key questions I discuss are: What are the fundamental principles behind the emerging solutions? Is there any generic architectural model, to explain those principles? Do we need new foundations to look at data distribution?

cs.DB↗

arXiv2009

Reducing Network Traffic in Unstructured P2P Systems Using Top-k Queries

A major problem of unstructured P2P systems is their heavy network traffic. This is caused mainly by high numbers of query answers, many of which are irrelevant for users. One solution to this problem is to use Top-k queries whereby the user can specify a limited number (k) of the most relevant answers. In this paper, we present FD, a (Fully Distributed) framework for executing Top-k queries in unstructured P2P systems, with the objective of reducing network traffic. FD consists of a family of algorithms that are simple but effec-tive. FD is completely distributed, does not depend on the existence of certain peers, and addresses the volatility of peers during query execution. We vali-dated FD through implementation over a 64-node cluster and simulation using the BRITE topology generator and SimJava. Our performance evaluation shows that FD can achieve major performance gains in terms of communication and response time.

cs.DB↗