W. Kent, "The Breakdown of the Information Model in MultiDatabase Systems," ACM SIGMOD Record Management of Data, vol. 20, no. 4, 1991.  Home/Search   Document Not in Database   Summary   Related Articles   Check

.... algorithms [18, 12] the application of methods known from the area of data mining and even machine learning [16] Other interesting results came from specific application areas, like for instance digital libraries [8, 14] An overview of problems related to entity identification is given in [15]. In [17] Lim et al. describe an equality based approach, include an overview of other approaches and list requirements for the entity identification process. Monge and Elkan describe an efficient algorithm that identifies similar tuples based on a distance measure and builds transitive clusters ....

W. Kent. The breakdown of the information model in multi-database systems. SIGMOD Record, 20(4):10--15, December 1991.

.... algorithms [18, the application of methods known from the area of data mining and even ma chine learning [16] Other interesting results came from specific application areas, like for instance digital libraries [8, 14] An overview of problems related to entity identification is given in [15]. In [17] Lim et al. describe an equality based approach, include an overview of other approaches and list requirements for the entity identification process. Monge and Elkan describe an efficient algorithm that identifies similar tuples based on a distance measure and builds transitive clusters ....

W. Kent. The breakdown of the information model in multi-database systems. SIGMOD Record, 20(4):10--15, December 1991.

.... algorithms [17, 11] the application of methods known from the area of data mining and even machine learning [15] Other interesting results came from specific application areas, like for instance digital libraries [7, 13] An overview of problems related to entity identification is given in [14]. In [16] Lim et al. describe an equality based approach, include an overview of other approaches and list requirements for the entity identification process. Monge and Elkan describe an efficient algorithm that identifies similar tuples D . B C D E B D F G A B F G B B B G G A ....

W. Kent. The breakdown of the information model in multi-database systems. SIGMOD Record, 20(4):10--15, December 1991.

....In addition, there is a term called computational proxy[7] that was introduced to model the legacy program in objectoriented databases. The computational proxy is used to represent an individual invocation of the program having complex inputs and long execution time. In the paper[11], objects in databases are regarded as proxy objects of the real world entities. If the same realworld entity is distributed in more than one database, it means that it has one or more proxy objects in the databases. Obviously, purposes of these terms are different from deputy objects. View is a ....

William Kent, The Breakdown of the Information Model in Multi-Database Systems, ACM SIGMOD Record, Vol. 20, No. 4, pp.10-15 (1991)

....data conflicts between values (conflict resolution) Especially during conflict resolution, IQ reasoning can greatly improve the result. Object Identification. Integrating data from different sources requires that different representations of identical real world entities be identified as such [Ken91]. This process is called object identification. Object identification is difficult, because the available knowledge about the objects under consideration may be incomplete, inconsistent, and sparse. A particular problem occurs if no natural IDs exist. For instance, the URL of a Web page is a ....

William Kent. The breakdown of the information model in multi-database systems. SIGMOD Record, 20(4):10-15, 1991.

....3: Object migration dimensions We will first specify these dimensions in more detail and then illustrate them by an example. 2.2.1 Object Kind As we have already mentioned in the object migration base model, all local objects of the various CDBS are mapped to (logical) global objects. Kent [10] recognized that multiple local objects (proxies) also may represent the same real world entity and thus map to a single global object. We analyzed the association of global objects to local objects more precisely and ended up with the following object kinds (see Fig. 4) 1. Globally new: global ....

KENT, W. The breakdown of the information model in multi-database systems. SIGMOD RECORD 20(4), pages 10--15, 1991.

....model for FDBS to globally control heterogeneous database systems, in this section we point out specific techniques to control redundant data. It represents the basis for our migration facilities. When database systems are coupled to an FDBS, in general they contain many redundant data. Kent [7] called it a mapping from many proxies to one real world entity which require specific object identification techniques [5, 6] For example, one of our project partners stores geometry information for hardware components redundantly in a layout system and a part library because it is needed by ....

....data in the various DBS. In order to supplementary control data redundancy that was created before federation resp. by local applications in the FDBS, we additionally offer techniques for object integration. They allow to identify existing data of different CDBS as a single entity for the FDBS [7]. For that, we developed a global operation same for the global external interface [16] It enables a user administrator to select some objects via queries and unify them to a single logical global object (Fig. 4) g1 global interface l1 l2 same CDBS2 CDBS1 Fig. 4. Identifying two local ....

KENT, W. The breakdown of the information model in multi-database systems. SIGMOD RECORD 20(4), pages 10--15, 1991.

....which use those old path expressions must be updated accordingly to enable them to be valid in the modified schema. Up to now, many researchers have studied issues related to avoiding database restructuring and reorganization due to schema modification ( Ber92] BKKK87] Zic91] Zic92] [Ken91], TS93] However, the issue of avoiding database reprogramming due to schema modification has received suprisingly little attention in the database research community. Why should reprogramming due to schema modification be avoided Reprogramming of object methods and database queries usually ....

W. Kent. The breakdown of the information model in multi-database systems. ACM SIGMOD RECORD on Management of Data, 20(4), 1991.

....of data relative to the entire data set. The results on the real world data validate our previous predictions as being quite accurate. 2 Previous Work Several lines of work in the database community have bearing on efficient solutions for the merge purge problem. The semantic integration problem [10] seeks to identify a multiplicity of database objects that represent the same or related real world entity, even though their database representations are different. This problem has been studied by researchers in the heterogenous multi database community. The solutions we present here for the ....

W. Kent. The Breakdown of the Information Model in Multi-Database Systems. SIGMOD Record, 20(4):10--15, December 1991.

.... of the data models that exist in the market are used as canonical data models in the existing systems, such as: Entity Relationship (DDTS [23, 91] Relational (MERMAID [14, 91, 104] DATAPLEX [14, 26, 104] ADDS [14, 104] Functional (MULTIBASE [14, 23, 91, 97, 104] Object Oriented (PEGASUS [1, 57], Heimbigner [14, 48] Hammer et al. 44, 45, 37, 38] ERC (Tari [103] and so on. As outlined by Saltor et al. 88] a data model is responsible for the representation ability of a database which is composed of two factors: expressiveness and semantic relativism. Expressiveness means the ....

.... in the literature as can be found in [5, 6, 20, 27, 30, 31, 39, 40, 62, 64, 77, 78, 80, 93, 101, 113] Other solutions have been suggested by commercial and prototype systems suchas:DATAPLEX [14, 26, 104] DDTS [23, 91] DHIM [25] MERMAID [14, 91, 104] MULTIBASE [14, 23, 91, 97, 104] PEGASUS [1, 57]. However, despite the methodologies, algorithms and heuristics that help the integration process, it is necessary to havehuman assistance to support the resolution of conflicts (there is a lack of automatic processes) This assistance is only reasonable when the number of components is small. ....

W. Kent. The breakdown of the information model in multi-database systems. SIGMOD RECORD, 20(4):10--15, December 1991.

....to determine if two proxy objects 4 represent the same real world object [LIM93] However in OO systems, generated Object Identifiers (OIDs) are used to uniquely identify objects. OIDs work perfectly well within the boundaries of a single database, but they are insufficient in a MDB environment [KEN91] as they have not been assigned in a global context. Integrity constraints may be used to resolve object identification 4 A proxy object is the database representation of a real world object [KEN91] 54 problems at the MDB level 5 [SCH94] For example, two proxy persons in different databases ....

....well within the boundaries of a single database, but they are insufficient in a MDB environment [KEN91] as they have not been assigned in a global context. Integrity constraints may be used to resolve object identification 4 A proxy object is the database representation of a real world object [KEN91] 54 problems at the MDB level 5 [SCH94] For example, two proxy persons in different databases represent the same real world person if they agree on name, address and date of birth. III. Schema (Meta Data) versus Instance (Data) Level Conflicts This type of conflict arises when what is ....

[Article contains additional citation context not shown here]

W. Kent. The Breakdown of the Information Model in Multidatabase Systems. SIGMOD RECORD, 20(4), pages 10--15, 1991.

.... report contradicting values for the same data item, such as different social security numbers for the same person, the FIS uses rules to remove the conflict [Ken 91b] Note that data fusion is very difficult; frequently it is impossible to identify objects or to decide which data value is correct [Ken 91a] Abstraction: Hereby, the federated data base on extracted data of the components, but further functions may be applied to lift or lower the source data to the abstraction level of the federation schema. The need for abstraction is in general caused by semantic conflicts. It encompasses ....

W. Kent, The Breakdown of the Information Model in Multi-Database Systems, SIGMOD Record 20(4): 10 - 15, 1991.

....Value Incompatibility Abstraction Level Incompatibility Schematic Discrepancy Figure 1: Structural Incompatibilities due to Heterogeneity (representational structural) differences. While there is a significant amount of literature discussing schematic differences, work on semantic issues (e.g. [Ken91]) in the database context is scarce. Classification or taxonomies of schematic differences appear in [DH84, BOT86, CRE87, KLK91, KS91] However, purely schematic considerations do not suffice to determine the similarity between objects [FKN91] SG89] In this paper we try to reconcile the two ....

W. Kent. The breakdown of the Information Model in Multidatabase Systems. SIGMOD Record, special issue on Semantic Issues in Multidatabases, A. Sheth, ed., 20(4), December 1991.

....define a data model independent taxonomy of semantic similarities. The possible values of the first component can be contexts constructed using the various operations mentioned above. While there is a significant amount of literature discussing schematic differences, work on semantic issues (e.g. [Ken91]) in the database context is scarce. Classification or taxonomies of schematic differences appear in [DH84, BOT86, CRE87, KLK91, KS91] In this paper we present what we believe is a comprehensive taxonomy of schematic conflicts which subsumes most of the taxonomies found in literature (Table 2 ....

W. Kent. The breakdown of the Information Model in Multidatabase Systems. SIGMOD Record, special issue on Semantic Issues in Multidatabases, A. Sheth, ed., 20(4), December 1991.

....one real world (entity) object may be represented by different database (proxy) objects in different component databases. The fundamental assumption of object oriented models, namely that one real world object is represented by exactly one database object, is therefore no longer true in FDBSs [9]. We therefore need an additional notion: we say that two local proxy objects are the same, iff they represent the same real world (entity) object. Same objects are identified in FDBSs by extending the composed schema using extend views to define so called same functions. Formally, we require that ....

W. Kent. The breakdown of the information model in multi-database systems. ACM SIGMOD Record, 20(4), December 1991.

....analyses with those of other users interested in the same domain and thereby discover anomalies. Further, applications may be written based on local assumptions that are not true across the federation because the semantics of the data have not been communicated outside of the local context [Kent91]. 7 Extracting semantic information from these various sources, determining what semantic information is interesting to users of the federation, and reconciling differences in semantics where they are relevant will require reading source code and documentation, interviewing developers and users, ....

Kent W. "The Breakdown of the Information Model in Multi-Database Systems", ACM SIGMOD Record, 20:4 (Dec. 1991). 20

....two classes (C 1 and C 2 ) are generalisation specialisation of each other, and no new class needs to be created. 6. Connecting variants: The two classes (C 1 and C 2 ) are variants of each other, and need to connect their extents. 3 PROPERTY INTEGRATION 3. 1 OBJECT IDENTITY INTEGRATION As in [Kent, 1991], we will distinguish between real world and proxy objects. Realworld objects are the conceptual objects that are modelled in the database by proxy objects. Two proxy objects may represent the same real world object. These objects are then said to be real world identical. If these proxy objects ....

....while we will assume that real world identities may be handled (i.e. compared) through user supplied functions. Sometimes, such a function may be inferred from other properties of the objects, like employee number or social security number, but often one has to rely on explicit enumerations [Kent, 1991]. Two classes to be integrated may or may not exist in the same extent graph prior to the integration. If the two classes are existing in distinct extent graphs, objects of these two classes cannot be proxy identical. In Section 4 we will present parts of a query algebra that has operators both ....

[Article contains additional citation context not shown here]

Kent W. (1991), The breakdown of the information model in multi-database systems.

....migration dimensions In the following, we first specify these dimensions in more detail and then illustrate them by an example. 3.1 Object Kind As we have already mentioned in the object migration base model, all local objects of the various CDBS are mapped to (logical) global objects. Kent [11] recognized that multiple local objects (proxies) also may represent the same real world entity and thus map to a single global object. We analyzed the association of global objects to local objects more precisely and ended up with the following object kinds (see Fig. 4) 1. Globally new: global ....

KENT, W. The breakdown of the information model in multi-database systems. SIGMOD RECORD 20(4), pages 10--15, 1991.

....section, we define same functions [SST94] a basic abstraction mechanism, to which object and schema level integration can be reduced. Object Identification. In an MDBS, entity objects (objects of the real world) are to be distinguished from proxy objects (their approximation in one database) Ken91] One particular entity object can be represented by multiple proxy objects in different component databases, and therefore, the fundamental assumption of object oriented systems (every real world object corresponds to exactly one database object) is no longer true. More formally: due to local ....

W. Kent. The breakdown of the information model in multi-database systems. ACM SIGMOD Record, 20(4), 1991.

....when two objects that represent the same real world entity have inconsistent information for example, one component database s data corresponds to meta data in other component databases) The problem is how to integrate data with semantic heterogeneity and semantic discrepancy in a HDBS. In [8], Kent gives a introduction to the factors that cause semantic heterogeneity and semantic discrepancy in a HDBS, and also presents some approaches to solve these problems. It is well known that these two problems exist even when all the component databases in a HDBS follow the same data model and ....

W. Kent. The breakdown of the information model in multidatabase systems. SIGMOD Record, special issue on Semantic Issues in Multidatabases, A. Sheth, ed., 20(4), December 1991.

....modified schema. The methods and queries which use those old path expressions must be updated accordingly to enable them to be valid in the modified schema. Up to now, many researchers have studied issues related to avoiding database restructuring and reorganization due to schema modification ([1, 3, 10, 21, 22, 23]) However, the issue of avoiding or minimizing database reprogramming due to schema modification has received surprisingly little attention in the database research community. Why should reprogramming due to schema modification be avoided Reprogramming of object methods and database queries ....

W. Kent. The breakdown of the information model in multi-database systems. ACM SIGMOD RECORD on Management of Data, 20(4), 1991.

No context found.

William Kent, "The Breakdown of the Information Model in Multi-Database Systems", SIGMOD Record 20(4) Dec. 1991.

....This simply doesn t work. Under the substitutability principle [Section 4.3.4] it is not possible for ECourse.Units(x) to be different from ECourse.Units(y) if x=y, since they are logically one and the same object. This problem requires the introduction of two distinct viewpoints (or spheres [9]) In an underlying sphere, x and y appear to be distinct objects, allowing the possibility of ECourse.Units(x)ECourse.Units(y) This sphere should only be visible to an administrator responsible for defining the reconciliation mechanisms. The end user should only see an integrated sphere in which ....

William Kent, "The Breakdown of the Information Model in Multi-Database Systems", SIGMOD Record 20(4) Dec 1991.

No context found.

W. Kent, "The Breakdown of the Information Model in MultiDatabase Systems," ACM SIGMOD Record Management of Data, vol. 20, no. 4, 1991.

No context found.

W. Kent. The breakdown of the information model in multi-database systems. SIGMOD Record, 20(4):10--15, December 1991.

First 50 documents

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