Krishnamurthy, R., W. Litwin, and W. Kent. Language Features for Interoperability of Databases with Schematic Discrepancies. Proc. ACM SIGMOD, pp. 40-49, Denver, CO, USA, 1991.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....views. Federated views are populated by a set of queries on the heterogeneous source databases. Those views can be materialized in a central data repository, or serve as a virtual interface to the user, masking the heterogeneity of the data sources. It has been shown by Krishnamurthy et al. [KLK91]) that to express federated queries over heterogeneous databases in a relational language requires special syntactic features. They take the form of variables ranging over schema elements: database, tables and attributes. Introducing these rneta variables is necessary to bridge discrepancies ....

....as T or H. Sets however will be noted using a single capital Greek letter such as Z, A, N (one exception is ( Set transformations, such as functions, and bijections will be noted with small Greek alphabet letters such as 2. 3 Example The same running example is used as Krishnamurthy et al. in [KLK91 ]. In that context there are three separate relational databases, each with a distinct schema. Despite their differences all three databases store the same information: stock market closing prices for every day. Assume for simplicity that only two companies, IBM (ibm) and Microsoft (msft) are ....

[Article contains additional citation context not shown here]

Krishnamurthy, Litwin, Kent: "Language Features for Interoperability of Databases with Schematic Discrepancies". ACM SIGMOD 1991:40-49

....evaluation shows, yield an effective execution by native query engines. Prior efforts do not provide feasible guarantees on the size of the compiled programs or require the development of new query engines encompassing higher order query operators. 1 Introduction Krishnamurthy, Litwin and Kent [KLK91], hereafter referred to as Krishnamurthy, showed that to concisely express federating views declaratively over heterogeneous databases requires special, higher order, syntactic features. The features consist of metadata variables ranging over schema elements: database, tables and columns. ....

.... Graphical View Definition Interface User Operator Feedback Declarative specifications Data Warehouse 3 Higher Order Database Languages In this paper, for accessibility, we speak to the syntax of SchemaSQL introduced by Lakshmanan in [LSS96] Like other higher order languages such as IDL ([KLK91]) SchemaLog ( LSS97] its syntax is characterized letting variables substitute in place of meta data elements: database, table and columns. Our presentation will leverage the same running example presented by Krishnamurthy. The example speaks to integrating stock market data from three actual ....

[Article contains additional citation context not shown here]

Krishnamurthy, Litwin, Kent: "Language Features for Interoperability of Databases with Schematic Discrepancies". ACM SIGMOD 1991:40-49

....revealed the difficulties of integrating data from multiple heterogeneous sources. After nearly 30 years of extensive research efforts and hundreds of commercially developed tools, this integration remains the province of engineers and ad hoc system development efforts. In an informal proof [KLK91], Krishnamurthy, Litwin and Kent demonstrated that merging data from multiple sources requires using higher order syntactic constructs to overcome schematic inconsistencies. They deduce that first order relational languages such as SQL or Datalog are not general enough to describe this class ....

....queries are pushed down to the source databases if possible and optimized. They are also often rule based approaches (i.e. the views are defined in a logic syntax rather than a SQL syntax) They are all capable of integrating sources with schematic discrepancies, as the problem was introduced in [KLK91]. Content Description Systems Another common approach is content based source descriptions. Instead of using a view definition, those systems rely on the user to describe the schema of each source, and annotate those descriptions. The purpose of those annotations is to let the system know what ....

Ravi Krishnamurthy, Witold Litwin, and Kent. Language features for interoperability of databases with schematic discrepancies. In Proceeedings of the ACM SIGMOD Conference, 1991.

....system and each presents a serious research challenge. Consider Scalability. A large scale interoperable database environment (e.g. three hundreds databases as opposed to three databases) presents challenging questions to the viability of loose coupling approach to system development (cf. [16, 17]) and the tight coupling integration framework that concentrates primarily on circumventing schematic and semantic heterogeneity (cf. 14, 29, 30, 31, 32] As the number of databases participating the interoperable database system increases, the design of an in tegrated schema involving n ....

....towards system scalability and evolution, tight coupling is generally considered a small system strategy. In contrast, loose coupling goes to another extreme by favoring 21 zero schema integration, for example, by promoting integration and interoperability via a multidatabase query language [16, 17]. It does not require the existence of an integrated schema, leaving many re sponsibilities, such as resolving semantic mismatch, dealing with multiple representations of data, and coping with dynamic system evolution, to the users. Recent research in database interoperability has started to pay ....

T. Landers, W. Litwin, and W. Kent. Language features for interoperability of databases with schematic discrepancies. In Proceedigs of ACM/'IGMOD Aual Coferece o Maagemet of Data, 1991.

....database systems immediately revealed the diculties of integrating data from multiple heterogeneous sources. After nearly 30 years of extensive research efforts and hundreds of commercially developed tools, this integration remains a complex and labor intensive tasks. In an informal proof [KLK91], Krishnamurthy, Litwin and Kent demonstrated that merging data from multiple sources requires higher order syntactic constructs to overcome schematic inconsistencies across data sources. They deduce that rst order relational languages such as SQL or Datalog can be extended with a higher order ....

....much work the system requires of the user to specify a federated database of a given complexity. Other existing languages or systems used to specify data transformations can be used as benchmarks. 9 4 Related Work Related e orts have focused on de ning semantics for speci cation languages. In [KLK91], a class of higher order query languages was proven necessary to specify federating queries in the most general case. Several second order languages were de ned in [LSS96] SchemaSQL, SchemaLog, expressive enough to precisely specify federating queries over data in semantically heterogeneous ....

Ravi Krishnamurthy, Witold Litwin, and Kent. Language features for interoperability of databases with schematic discrepancies. In Proceeedings of the ACM SIGMOD Conference, 1991.

....we wish the database system had provided. These new capabilities should be of interest to the database engine architects and implementors. There is rich heterogeneous database research literature on transformations between schematically disparate schemas and types of schematic differences. In [10], Krishnamurthy et al. eloquently elucidated how data values in one data source may be modeled as schema (attribute or relation) labels in another. Several languages have beenproposed for querying over schema labels, including [13] 17] There is also work on defining higher order views for ....

....elucidated how data values in one data source may be modeled as schema (attribute or relation) labels in another. Several languages have beenproposed for querying over schema labels, including [13] 17] There is also work on defining higher order views for integrating heterogeneous data sources [10] [12] 14] Closest to this paper is the interesting work presented in [11] on the implementation of SchemaSQL. We share with them the goal of a non intrusive implementation (i.e. without requiring changes in the database engine code) The extended algebra we use in our query transformations ....

R. Krishnamurthy, W. Litwin, and W. Kent. Language features for interoperability of databases with schematic discrepancies. In Proceedings of the

....language to be XML. 2.1.1 Issues with Data Heterogeneity The classes of conflict that spring from data heterogeneity partitions into three rough sections, all well documented in literature. Their major di#erences are in: 1. Schematic Conflict: How data are structured or logically organized [23] [24]. 1 Semantic interoperability means the meaningful exchange of information . 19 2. Semantic Conflict: How data are interpreted (i.e. what does this mean ) 31] 28] 3. Intensional Conflict: What context (sources of information available to the sender and receiver) does the data require ....

R. Krishnamurthy, W. Litwin, and W. Kent (1991). Language features for interoperability of databases with schematic discrepancies. In Proceedings of the ACM SIGMOD Conference, pages 40-49.

....common data model and the data models of the heterogeneous databases. Higher order logics. The second approach for schema integration involves defining a higher order language that can express relationships between the meta information corresponding to the schemata of the individual databases [22, 25]. Thus, a common data model is not required, but the higher order language plays the role of the data model. The major advantage of this approach is the declarativeness it derives from its logical foundation. The approaches above, however, are targeted towards the integration of heterogeneous ....

R. Krishnamurthy, W. Litwin, and W. Kent, "Language features for interoperability of databases with schematic discrepancies," Proc. ACM SIGMOD Int. Conf. on the Management of Data, 1991, pp. 40-49.

....and naming, attribute and constraint conflicts have to be detected and solved. A taxonomy of conflicts is given in [19, 10] For our discussion, we use a very simplified example focusing on extensional and intensional overlappings. We assume that other conflicts are resolved (e.g. meta conflicts [11, 5] and attribute conflicts [12] Both schemata are object oriented, and attributes with identical name have identical semantics. The example is taken from two library applications and is shown in Figure 6. Using the library example, we will now discuss the integration process informally. A more ....

R. Krishmamurthy, W. Litwin, W. Kent. Language Features for Interoperability of Databases with Schematic Discrepancies. In Y. Kambayashi, M. Rusinkiewicz, and A. Sheth, editors, Proc. of the 1st Int. Workshop on Interoperability in Multidatabase Systems (IMS'91), Kyoto, Japan, pages 144--151. IEEE Computer Society Press, April 1991.

....and eliminated 2 . 2 A taxonomy of conflicts is given in [SK93, Kim95] 3 For our discussion, we use a very simplified example focusing on the problem of solving extensional and intensional overlappings. We assume that other conflicts are resolved (e.g. naming conflicts, meta conflicts [KLK91, CL94] and attribute conflicts [LNE89] Both example schemata are object oriented, and attributes with identical name have identical semantics. The example schemata are taken from a library application and from a university administration, respectively (see Figure 1) The filled subclass symbol ....

R. Krishmamurthy, W. Litwin, and W. Kent. Language Features for Interoperability of Databases with Schematic Discrepancies. In Y. Kambayashi, M. Rusinkiewicz, and A. Sheth, editors, Proc. of the 1st Int. Workshop on Interoperability in Multidatabase Systems (IMS'91), Kyoto, Japan, pages 144--151. IEEE Computer Society Press, April 1991.

....and methods. Unlike OQL, it is possible to query data without complete knowledge of the schema. However, the complex nature of XSQL raises concerns about effective and efficient implementation, a concern not addressed in their work. Also in the relational world, several papers [KLK88,CKW89,KLK91,LSS93,KLJ95] have appeared in the literature that address the meta data dependency problem. The solutions proposed in [KLK88,CKW89,KLK91] augment the query language with mechanisms that allow it to query both metadata and ordinary data. These solutions are embedded in very powerful ....

....raises concerns about effective and efficient implementation, a concern not addressed in their work. Also in the relational world, several papers [KLK88,CKW89,KLK91,LSS93,KLJ95] have appeared in the literature that address the meta data dependency problem. The solutions proposed in [KLK88,CKW89,KLK91] augment the query language with mechanisms that allow it to query both metadata and ordinary data. These solutions are embedded in very powerful obeject oriented query languages. Following the work [LS93] schemaSQL [LSS96,LSS99] is a recently proposed extension to SQL designed for ....

R. Krishnamurthy, W. Litwin, and W. Kent. Languages features for interoperability of databases with schematic discrepancies. In Proceedings of the ACM SIGMOD International Conference on Management of Data, pages 40--49, 1991.

....approach is similar in that we produce a KIF universal structure representation for each local F logic database. Other approaches may be classified as follows: those advocating a unified global schema [1, 2, 9, 17, 18] those requiring the use of a higher order language or logic or a meta model [28, 3, 7, 15, 20, 21]; and those advocating a federated approach or a mapping approach based on a canonical representation [4, 8, 9, 19, 22, 24, 25, 26, 27] We will compare our research with these projects in section 6. Mediators resolve conflicts so that data from multiple information sources can be integrated. The ....

Krishnamurthy R., et al ., "Language Features for Interoperability of Databases with Schematic Discrepancies." ACM SIGMOD 1991.

....of measure, etc. and schema mismatch (different constructs in the data models representing the same concept) They use integrator functions to solve the mismatch. They do not address the problem of query transformation. Discrepancies that may occur within the same data model is investigated by Krishnamurthy et al. . 1991). Kim and Seo (1991) provide an exhaustive list of conflicts that may occur in a multidatabase environment, expressed in the context of the relational data model. Ahmed et al. (1993) Albert et al. (1993) describe Pegasus, a heterogeneous DBMS that provides access to native and external ....

Krishnamurthy, R., W. Litwin, and W. Kent (1991) "Language features for interoperability of databases with schematic discrepancies," Proceedings of the ACM Sigmod Conference.

....between two entities can be solved by creating a subclass relationship among the entities. Other problems that are studied are aggregation and composition conflicts. Alternately, the capability of a mediator is supported by the use of higher order query languages or meta models [22] 23] 24] [25], 26] 27] The language or model provide constructs to resolve conflicts among the sources. Mediators are also implemented through the use of mapping knowledge bases that capture the knowledge required to resolve conflicts among the local schema, and mapping or transformation algorithms that ....

....defined over the extents of the data sources. We support incorporating new data sources, with no type mismatch, or simple type mismatch, by modifying the extent of the mediator type, and by mapping types to resolve simple type conflicts. In related research [17] 3] 24] 19] 18] [25], the mismatch of the data types, formats, values, etc. with respect to data source types and mediator types was resolved by obtaining a single unified type. In Disco, our first objective is to support scale up by easing the introduction of new data sources. Each addition of a type and resolution ....

R. Krishnamurthy, W. Litwin, and W. Kent, "Language features for interoperability of databases with schematic discrepancies," in Proceedings of ACM SIGMOD International Conference on Management of Data, Denver, CO, May 1991.

....of measure, etc. and schema mismatch (different constructs in the data models representing the same concept) They use integrator functions to solve the mismatch. They do not address the problem of query transformation. Discrepancies that may occur within the same data model is investigated by [Krishnamurthy et al. , 1991]. Kim and Seo, 1991] provide an exhaustive list of conflicts that may occur in a multidatabase environment, expressed in the context of the relational data model. This is later extended to the object data model in [Kim et al. , 1993] In [Albert et al. , 1993] Pegasus, a heterogeneous DBMS that ....

Krishnamurthy, R., W. Litwin, and W. Kent (1991) "Language features for interoperability of databases with schematic discrepancies," Proceedings of the ACM Sigmod Conference.

....queries transparently, so the local user (or legacy application) does not need explicit knowledge (about the global schema or mapping knowledge among the schema) to pose a query. In companion papers (e.g. Raschid et al. (1993) we have described other research approaches (Kent (1991) Krishnamurthy et al. (1991), Ahmed et al. (1993) etc, and compared them with our research. The research that is closest to our work is described by Lefebvre et al. (1992) and Qian (1993) Lefebvre et al. (1992) consider the problem of interoperable query processing, but their approach is limited to relational schema. ....

Krishnamurthy, R., W. Litwin, and W. Kent (1991) "Language features for interoperability of databases with schematic discrepancies," Proceedings of the ACM Sigmod Conference.

....Researchers need tools that will make it possible to query and analyze data across heterogeneous databases in a user transparent way. While it is a relatively recent area of research in bioinformatics, schema integration, or interoperability, is a fairly well studied discipline in databases [18, 19]. In the context of genomic research, databases may be classi ed as loosely integrated or tightly integrated. In loose integration, databases are not obliged to participate in the federation. Other databases may refer to its contents by some sort of reference, such as hyper links. The database ....

R. Krishnamurthy, W. Litwin, and W. Kent. Language features for interoperability of databases with schematic discrepancies. In Proc. ACM SIGMOD, 1991.

No context found.

Ravi Krishnamurthy, Witold Litwin and William Kent, "Language Features for Interoperability of Databases with Schematic Discrepancies", Proc ACM SIGMOD Int'l Conf on Mgmt of Data, Denver, Colorado, May 29-31 1991.

No context found.

Krishnamurthy, R., W. Litwin, and W. Kent. Language Features for Interoperability of Databases with Schematic Discrepancies. Proc. ACM SIGMOD, pp. 40-49, Denver, CO, USA, 1991.

No context found.

R. Krishnamurthy,W.Litwin, and W. Kent. Language features for interoperability of databases with schematic discrepancies. In J. Clifford and R. King, editors, Proceedings of ACM SIGMOD Conference, pages 40--49. ACM, May1991.

No context found.

R. Krishnamurthy, W. Litwin, and W. Kent. Language features for interoperability of databases with schematic discrepancies. In Proc. Intl. Conf. on Management of Data (SIGMOD), pages 40-49. ACM Press, 1991.

No context found.

R. Krishnamurthy, W. Litwin, and W. Kent, "Language features for interoperability of databases with schematic discrepancies." in Proc. of SIGMOD, Denver, Colorado, 1991, pp. 40--49.

No context found.

Ravi Krishnamurthy, Witold Litwin, and William Kent. Language features for interoperability of databases with schematic discrepancies. In Proceedings of the SIGMOD International Conference on Management of Data, pages 40--49. ACM Press, 1991.

No context found.

R.Krishnamurthy, W.Litwin, W.Kent: `Language Features for Interoperability of Databases with Schematic Discrepancies ', Proc. ACM SIGMOD Conf., 1991.

No context found.

R. Krishnamurthy, W. Litwin and W. Kent. (1991) Language Features for Interoperability of Databases with Schematic Discrepancies, Proc. ACM SIGMOD 1991.

First 50 documents  Next 50

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC