英语论文网

The Piazza Peer Data Management ProjectIgor Tatarinov1, Zachary Ives2, Jayant Madhavan1,Alon Halevy1, Dan Suciu1, Nilesh Dalvi1, Xin (Luna) Dong1,
Yana Kadiyska1,留学生论文网 Gerome Miklau1, Peter Mork1
1Department of Computer Science and Engineering University of Washington, Seattle, WA 98195figor,jayant,alon,suciu,nilesh,lunadong,ykadiysk,gerome,pmorkg@cs.washington.e du2Department of Computer and Information Science
University of Pennsylvania, Philadelphia, PA 19103
zives@cis.upenn.edu
ABSTRACT
A major problem in today’s information-driven world is thatsharing heterogeneous, semantically rich data is incrediblydifficult. Piazza is a peer data management system that enablessharing heterogeneous data in a distributed and scalableway. Piazza assumes the participants to be interested insharing data, and willing to define pairwise mappings betweentheir schemas. Then, users formulate queries overtheir preferred schema, and a query answering system expandsrecursively any mappings relevant to the query, retrievingdata from other peers. In this paper, we provide abrief overview of the Piazza project including our work on
developing mapping languages and query reformulationalgorithms,assisting the users in defining mappings, indexing,andenforcing access control over shared data.
1. INTRODUCTION
A major problem in today’s information-driven world isthat sharing heterogeneous, semantically rich data is frustratinglydifficult. Authoring and publishing Web documents isvery easy. Peer-to-peer systems make dissemination of musicfiles a simple job. But no similar technology exists forsharing information with different schemas and representations.
This is not entirely surprising, since sharing semanticallyrich data is inherently much harder: data in differentschemas must somehow be mapped or related, queries aremuch richer, and security is an important consideration.Yet, the benefits of semantic data sharing are enormous.
Consider the problem of sharing scientific data. Until recently,individual researchers collected and analyzed data inisolation, studying small-scale phenomena and keeping theirdata proprietary. But today, by integrating and aggregatingdata from multiple sources scientists can conduct experimentsthat were not possible before. Examples include theSkyQuery project in astronomy [19], the Institute for SystemsBiology [12], and the Human Brain Project. At a global
scale, the Semantic Web [5, 10] aims at allowing universalsharing of semantically rich data.The problem of sharing semantically rich data has been
addressed by the data management community from differentangles. Mariposa [22] implemented distributed data sharingover heterogeneous resources (but not schemas). In dataintegration a centralized mediated schema is created to integratemultiple data sources; a particular problem here is howto generate the mapping between two schemas [21]. Researchon the SemanticWeb [5, 6] focuses on highly expressive
knowledge representation languages, in the hope thatthis will facilitate the creation of a universal ontology. Distributedquery processing has been intensively studied, assuminga known global schema [15]. Yet none of these techniques
addresses the two key challenges in sharing rich, heterogeneous
data: that the number of distinct schemas is very
large, and that a centralized mediated schema is not possible.
In Piazza, we address the problem of sharing semanticallyheterogeneous data in a distribut本论文由英语论文网提供整理，提供论文代写，英语论文代写，代写论文，代写英语论文，代写留学生论文，代写英文论文，留学生论文代写相关核心关键词搜索。