D. Bell and J. Grimson. Distributed Database Systems. Addison-Wesley, Wokingham, 1992.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....concurrency control, etc. and recovery issues, these are investiagted only with respect to how they impact checkpointing. When examining recovery in a distributed DBMS we must be able to recover from transaction failures, site (system) failures, media failures, and communication failures [1, 4, 36]. As is standard in discussing distributed databases, we divide the logical functions of processing into Transaction Manager (TM) and Data Manager (DM) functions. Sites in the distributed network may function as either or both. This research was partially supportedby the National Science ....

....that writes information to stable storage during normal operation in order to reduce the amount of work at restart [2, p. 323] Checkpointing involves writing to stable storage (usually disk) status information about update operations as well as flushing memory data to the stable database [1, 2]. The first part of this information is usually recorded on the log and is used at recovery to determine which operations need to be undone or redone. This information is often kept not at the operation level but at the transaction level. The second type of checkpointing activity, flushing ....

[Article contains additional citation context not shown here]

David Bell and Jane Grimson. Distributed Database Systems. Addison-Wesley publishers, 1992.

....in the data models (e.g. relational, network, ER, object oriented etc. in the query and data manipulation languages, in the data types (e.g. text, graphics, multimedia, hypermedia, etc. in the format (e.g. structured, semistructured) and in the semantics. Database interoperability [15, 18, 3, 17] is the ability of distributed, heterogeneous databases, which are independently created and administrated and have different semantics and schemas to cooperate and interoperate in a transparent way to the user while maintaining their autonomy and objectives. The requirements and objectives for ....

Bell, D., and Grimson, J., 1994. Distributed Database Systems. International Computer Science Series, Addison Wesley, Wokingham, England.

....of the actual configuration of the environment. In the WibQuS environment, technical integration is based on the industry standard SQL. Actually, a commercial product for distributed SQL access is incorporated. Technically, the architecture conforms to a loosely coupled federated database system [1]. SQL statements for the data flow (table declarations, queries, views, triggers) are generated from abstract specifications in the conceptual design. SQL SQL Met a KB subsyst em QFD KB QFD KB FTA KB XSPC FTA SQL subsyst em XSPC TRADER NETWORK Fig. 5.1: Trader architecture for technical ....

D. Bell, J. Grimson (1992) . Distributed Database Systems. Addison-Wesley.

....by the distributed schema catalog Developing highly available information systems, we propose a distributed schema catalog in conjunction with autonomy design phases : Distributed schema catalog. In general, distributed database systems have to deal with a lot of tradeoffs [ Rahm, 1994; Bell and Grimson, 1992; Ozsu and Valduriez, 1991 ] e.g. data replication versus data transfer, reuse versus autonomy, transparency versus efficiency. Our goals in gaining distribution for an object oriented data model are: ffl Distribution transparency : For scalability support transparency of location and migration ....

....Distributed administration is performed by the events of the meta schema. 4 Autonomy Design Nowadays information systems are based on distributed infrastructures to manage the requirements of the users. To take advantage of a decentralizable platform several existing disadvantages [ Rahm, 1994; Bell and Grimson, 1992; Ozsu and Valduriez, 1991 ] of distributed database systems have to be regarded. Due to the possibility of node failures (e.g. power failure, hardware or software failure, installation or maintenance tasks or user control failure) autonomy considerations have to be taken into account. This ....

D. Bell and J. Grimson. Distributed Database Systems. Addison-Wesley, 1992.

....in the system. For example, an agent cannot know that an event, say e, occurred unless either he has participated in that event or knowledge about e was acquired from another agent during a meeting. Knowledge acquisition is modeled by introducing queries in the similar manner as in databases [3]. It is our intention that the framework sketched above can serve as a semantics for formal logics of distributed games and multi agent systems. In the theory of distributed systems, knowledge formulas are usually interpreted over infinite linear or branching runs of the systems [4, 9, 16, 17] ....

D. Bell, and J. Grimson. Distributed Database Systems. AddisonWesley Pub. Co., 1992.

....in the system. For example, an agent cannot know that an event, say e, occurred unless either he has participated in that event or knowledge about e was acquired from another agent during a meeting. Knowledge acquisition is modeled by introducing queries in the similar manner as in databases [3]. It is our intention that the framework sketched above can serve as a semantics for formal logics of distributed games and multi agent systems. In the theory of distributed systems, knowledge formulas are usually interpreted over infinite linear or branching runs of the systems [4, 9, 16, 17] ....

D. Bell, and J. Grimson. Distributed Database Systems. Addison-Wesley Pub. Co., 1992.

.... a la robotique et a certains syst emes de surveillance. Dans ce qui suit, nous d ecrivons bri evement quelques syst emes d iad, en l occurence HOPES, HECODES, ARCHON et SAHARA. On evoquera enfin un exemple de travaux sur les syst emes r eactifs. 7. 1 HOPES La structure hi erarchique de HOPES [7] (Hierarchically Organized Parallel Expert System) est bien adapt ee a la r esolution de probl eme par niveau. Le cheminement du flot d informations entre agents est effectu e au moyen de blackboards partag es par les agents a plusieurs niveaux. Par exemple, dans la figure 8, l agent Ax sera ....

....steria. De l Intelligence Artificielle Distribu ee aux Syst emes Multi Agents 29 Blackboard B Ax 1 Ax 2 Ax 3 Ax 4 Ax Blackboard H de plus haut niveau Autres agents Fig. 8 mod ele hi erarchique de HOPES 7. 2 HECODES HECODES (HEterogeneous COoperating Distributed Expert System) [7] est un environnement appropri e pour une coop eration horizontale, hi erarchique et r ecursive. Il est aussi bas e sur un syst eme de blackboard. Il pr esente un syst eme de controle centralis e, un syst eme de blackboard et des syst emes experts. Les agents coop erant peuvent etre ....

D. Bell and J. Grimson. Distributed Database Systems. Addison--Wesley, 1992.

....and use it to its full capability. 14 For the purpose of this project, we are primarily interested in the first two categories, even though we may touch on the others in passing. Our literature survey ( CDY95] CP84] Chu92] CMVN92] OV91] YMW 85] MIMH85] CP87] NCWD84] BG92] and [NM89] reveals the following main considerations for deciding on the appropriate data partitioning for a given federated database. The overriding objective is to minimize the inter site communication time and maximize the proportion of local data accesses. 1. Maximizing Local Processing ....

David Bell and Jane Grimson. Distributed Database Systems. Addison-Wesley Publishing Company, 1992.

....of efficient tools for manipulating and integrating large amounts of data. Frequently these data are distributed among different computing systems on various sites. Distributed Database Management Systems provide a superstructure which integrates either homogeneous or heterogeneous DBMS (Bell and Grimson, 1992). In recent years, there has been a convergence between Database Technology and Statistics. The major statistical packages now incorporate some database functionality, while there has been increasing pressure for statistical databases to be developed which include all the advantages of modern ....

Bell, D.A., and Grimson, J.B. (1992), Distributed Database Systems, Addison Wesley.

.... DATABASE VERSUS IR ARCHITECTURES There is also a large volume of work on architectures for distributed and parallel database systems including research on performance, e.g. Stonebraker et al. 1983; DeWitt et al. 1986; Mackert and Lohman, 1986; Hagmann and Ferrari, 1986; DeWitt and Gray, 1992; Bell and Grimson, 1992. Although the fields of information retrieval and databases are similar, there are several distinctions which make studying the performance of IR systems unique. A major difference between database systems and information retrieval systems is structured versus unstructured data. In a structured ....

Bell, D. and Grimson, J. (1992). Distributed Database Systems. AddisonWesley Publishers.

....A B = ICQ(A) ICQ(B) ECQ1 (ECQ Gamma ECQ2) 6) G(ECQ2 [ ICCQ(ECQ2) G(ECQ1 [ ICQ(ECQ1) 7) G(A) G(B) G(A [ B) 8) A Gamma B) B = A (9) ICQ(A [ B) ICQ(A) ICCQ(B) Notes 1. Other authors consider that only some federated databases provide an integrated view (Bell and Grimson, 1992); they call them tightly coupled federated databases. 2. Although they all use the same data model, semantic heterogeneity still can remain. 3. Also known as Terminological Logics or based on KL ONE (Borgida et al. 1989, Peltason et al. 1989) 4. The general form of a query is: self, ....

D. Bell and J. Grimson. Distributed Database Systems. Addison-Wesley, 1992.

.... [Lu et al. 1998] There is also a large volume of work on architectures for distributed and parallel database systems including research on performance (e.g. see [Stonebraker et al. 1983, DeWitt et al. 1986, Mackert and Lohman, 1986, Hagmann and Ferrari, 1986, DeWitt and Gray, 1992, Bell and Grimson, 1992] Although the fields of information retrieval and databases are similar, there are several distinctions which make studying the per6 formance of IR systems unique. A major difference between database systems and information retrieval systems is structured versus unstructured data. The ....

Bell, D. and Grimson, J. (1992). Distributed Database Systems. Addison-Wesley Publishers.

....to be strict subsets of the canonical schema. In principle, the canonical schema can be generated for each exchange although in practice we expect that most exchanges will reuse previous versions of the conversion tables. Thus Hypermedata is loosely federated in the sense of the classification of [BG92] 1.3.3 Modelling Environment We are presented with a modelling environment which is in a high state of flux. This arises in part from the dynamic state of the federation of systems composing Hypermedata. New systems with their own data models may be joining the federation in the future. The ....

D. Bell and J. Grimson. Distributed Database Systems. Addison-Wesley, 1992.

....of the data and are only propagated to the Concurrency Control 46 database at commit if no conflicts have been detected. In the event of conflict, the transaction is rolled back and restarted. Restarting a transaction means redoing the entire transaction and may cause considerable overhead [13, 77]. On the other hand, pessimistic concurrency control methods are based on the notion of guaranteeing a priori that no other transaction accesses or modifies those items being involved in the current operation. The main technique used to control concurrent execution in that way is based on the ....

....which are involved in a transaction have to be locked, so that all conflicting actions are halted until all locks are released. Further, it is not allowed that a transaction requests an additional lock if any lock has already been released [151] Optimistic concurrency control should be used when [13]: ffl conflicts are unlikely to happen ffl the degree of contention on an object is not high ffl a high degree of concurrency is necessary ffl read only actions predominate and don t conflict Pessimistic concurrency control should be used when: ffl the degree of contention on an object is ....

Bell, D., and Grimson, J. Distributed Database Systems. AddisonWesley, 1992. BIBLIOGRAPHY 120

....shown. Via the interface layer, one and the same parameter P is made available to the components once as float and once as int. model is enhanced by a inter process and inter host communication level (eg DCOM [4] or CORBA [11] and the status database is implemented as a distributed database [2]. In the case of a distributed database, the data are distributed to several computers. Sub)queries are processed locally (buffered get) data consistency (after place) being ensured by automatic alignment of the various subdatabases. A global transaction concept is provided, too. 2.5. ....

D. Bell and J. Grimson. Distributed Database Systems. Addison-Wesley Publ. Company, 1992.

....is employed which intercepts all update requests, from legacy or target applications, and processes them to identify whether they refer to data replicated in both databases. If they do, the update is propagated to both databases using a two phase commit protocol as for distributed database systems [6]. Analysing non decomposable legacy components can be very difficult. In the worst case the component must be treated as a black box. The best that can be achieved is to discover its functionality and try to elicit as much legacy data as possible. Sometimes using existing legacy applications, ....

D. Bell and J. Grimson, "Distributed Database Systems", Addisson-Wesley, 1992.

....store scheduled flights is that of S FRA. One possible way to integrate both schemas is to map the schema S BOS to that of S FRA. However, the next step in modeling our system requires not an algorithm for the integration procedure. 3 For more detail on federated database systems please confer [17, 11, 4]. 4 For an introduction on function modeling please confer [2] F1.1 C3: Schema integration completed C2 C1 F1.2 F1 Figure 3.4: Refined transition F1 with respect to sub functions of F1. Figure 3.4 shows the balanced refinement of F1 with respect to the sub functions F1.1 and F1.2. This ....

Bell, D., and Grimson, J. Distributed Database Systems. Addison-Wesley, 1992.

....y[i] z[iz[i] do i=1,1000 SEND z[ RECV z[ SEND x[ RECV x[ SEND y[ RECV y[ RECV x[ SEND x[ Figure 5.7 Communication placement after taking resources into account. 103 is very similar to the ones encountered in scheduling and database management (file allocation) problems [32, 10] and can be similarly shown to be NP complete: the knapsack problem [32] can be trivially reduced to the problem of optimally selecting the messages that should use the available buffer at any given node. The knapsack problem is a classical problem in operations research where a scarce resource ....

D. Bell and J. Grimson. Distributed Database Systems. Addison-Wesley Publishing Company, Great Britain, 1992.

....in our FDBS, namely: Translator, Integrator and Query Processor. In figure 1 the architecture of the FDBS is shown. The work of this author was supported by a grant of the Government of the Basque Country 1 Other authors consider that only some federated databases provide an integrated view [BG92] they call them tightly coupled federated databases. 2 Although they all use the same data model, semantic heterogeneity still can remain. 3 Also known as Terminological Logics or based on KL ONE [BBMR89, PSKQ89] Cache memory INTEGRATOR TRANSLATOR Global Query Processor Processor DBMS 2 ....

D. Bell and J. Grimson. Distributed Database Systems. Addison-Wesley, 1992.

....by the distributed schema catalog Developing highly available information systems, we propose a distributed schema catalog in conjunction with autonomy design phases : Distributed schema catalog. In general, distributed database systems have to deal with a lot of tradeoffs [ Rahm, 1994; Bell and Grimson, 1992; Ozsu and Valduriez, 1991 ] e.g. data replication versus data transfer, reuse versus autonomy, transparency versus efficiency. Our goals in gaining distribution for an object oriented data model are: ffl Distribution transparency : For scalability support transparency of location and migration ....

....Distributed administration is performed by the events of the meta schema. 4 Autonomy Design Nowadays information systems are based on distributed infrastructures to manage the requirements of the users. To take advantage of a decentralizable platform several existing disadvantages [ Rahm, 1994; Bell and Grimson, 1992; Ozsu and Valduriez, 1991 ] of distributed database systems have to be regarded. Due to the possibility of node failures (e.g. power failure, hardware or software failure, installation or maintenance tasks or user control failure) autonomy considerations have to be taken into account. This ....

D. Bell and J. Grimson. Distributed Database Systems. Addison-Wesley, 1992.

....control, and transactional behaviour (Gray Reuter 1993) is required, which should also take care of the privacy of the users being modeled. Most of this functionality is already established in other areas (e.g. distributed operating systems (Goscinski 1991) or (distributed) database systems (Bell Grimson 1992; Weikum 1988) In the next section we first describe different architectures for the adaptive application and the user modeling shell system and point out their main limitations. Finally we briefly describe a new architecture for BGP MS that avoids nearly all disadvantages, and at the same time ....

Bell, D., and Grimson, J. 1992. Distributed Database Systems. Reading, MA: Addison-Wesley.

.... authentication, access control, integrity control, and transactional behaviour (Gray Reuter 1993) is required, which should also take care of the privacy of the users being modeled. Most of this functionality is already established in other areas (e.g. distributed operating systems (Bell Grimson 1992), Weikum 1988) Architectures of today s user modeling shell systems Three different run time architectures for the adaptive application and the today s user modeling shell systems can be identified: ffl The local and monolithic architecture: The user modeling shell systems GUMS (Finin 1989) ....

Bell, D., and Grimson, J. 1992. Distributed Database Systems. Reading, MA: Addison-Wesley.

....there are existing databases in the world which users wish to access, then they must either learn the interfaces of each independent database, or use a common interface tool to access all of these heterogenous databases. We are assuming a weakly federated multibase (following the terminology of Bell Grimson, 1992) in which local operators have update rights to their own databases, and must obey local data integrity constraints; multibase data retrieval users do not have update rights, and although global integrity constraints may exist, they cannot be enforced since the autonomy of local databases cannot ....

Bell, D. & Grimson, J. (1992) Distributed Database Systems, Addison-Wesley: Wokingham, England.

....is employed which intercepts all update requests from legacy or target applications and processes them to identify whether they refer to data replicated in both databases. If they do, the update is propagated to both databases using a two phase commit protocol as in a distributed database [Bell92]. As Brodie and Stonebraker themselves point out: update consistency across heterogeneous information systems is a much more complex technical problem with no general solution yet advised, and it is still an open research challenge [Brod95] Thus, it seems that applying the Chicken Little ....

....nature of many legacy systems, transaction support will be a basic requirement for the target, enterprise wide applications. These applications seek a secure and distributed environment, 12 where many users can access simultaneously and efficiently a diversity of data sources and applications [Bell92]. Finally, giving the growing importance that the World Wide Web (WWW) has in the way companies do business, the target architecture should facilitate WWW s integration within the enterprise wide information system [Perr95] An extensive literature exists regarding possible architectures and ....

[Article contains additional citation context not shown here]

D. Bell, J. Grimson, 'Distributed Database Systems', Addisson-Wesley, 1992.

No context found.

D. Bell and J. Grimson, Distributed Database Systems, Addison Wesley Longman, Reading, Mass., 1992.

....problem at hand, that is, accessing different data sources to provide data to the Synapses Server. The Generic Adapter does not pretend to be a solution to the problems presented by more sophisticated approaches of dealing with distributed information, like distributed and heterogeneous databases [Bell92][March90] 6. ....

D. Bell, and J. Grimson. `Distributed Database Systems', Addison-Wesley Longman, Reading, Mass., 1992.

No context found.

D. Bell and J. Grimson. Distributed Database Systems. Addison-Wesley, Wokingham, 1992.

No context found.

Bell, D.: Distributed database systems. Addison Wesley (1993)

No context found.

BELL,D.,AND GRIMSON,J.B.Distributed Database Systems, Addison-Wesley, 1992.

No context found.

Bell D. and Grimson J., Distributed database systems, Addison Wesley, Reading Mass, 1992.

No context found.

Bell, D. and Grimson, J. (1992). Distributed Database Systems, Addison-Wesley.

No context found.

Bell, D., and J. Grimson, Distributed Database Systems, Addison-Wesley, 1992.

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