Query processing remains one of the important challenges of Object-Oriented Database Management Systems. Cost based query optimization involves creating alternative executing plans for a given query and executing the least costly one within a cost model framework. In Object-Oriented Database Management Systems (OODBMSs) objects may store references to other objects (precomputed joins), and path expressions are used in query languages. Although the cost fomulas for explicit joins and the selectivities of attributes and joins are well-known in the relational model, there is no similar work involving path expressions for OODBMSs. However in order to optimize object-oriented queries involving path expressions, a cost model is essential. This information is necessary for deciding whether to use pointer chasing or to convert the path expressions into explicit joins and also for deciding the execution order of path expressions. In this paper, we provide a cost model that includes the formulas...

In recent years, as more applications start using massive databases as their main source of information, more emphasis is placed on the performance of the database system. These require not only that the database system have good performance, but also that it be continually available. The research in this thesis makes strides in meeting these requirements: dynamically clustering data improves the database performance, and efficient on-line reorganization methods enable the database systems to be continually available. An new algorithm, Enc, for dynamically clustering hierarchical data is presented in this thesis. It uses a primary B + -tree as the main storage structure, all relations in the hierarchy are stored in the B + -tree. The hierarchical relationship is encoded into the keys of the B + -tree. The Enc algorithm maintains good clustering in the presence of insertions and deletions. Experimental results show that using the Enc algorithm, hierarchical queries can be process...

DESIGN AND IMPLEMENTATION OF A DYNAMIC FUNCTION LINKER AND AN ALGEBRA FOR AN OBJECT-ORIENTED DATABASE MANAGEMENT SYSTEM ALTINEL, Mehmet M.S. in Computer Engineering Supervisor : Prof. Dr. Asuman DOG A September, 1994, 96 pages In recent years it has become apparent that commercial database products based on relational data model are not sufficient for complex data intensive applications. These advanced, design-support applications caused database community to focus on solving problems of handling complex information and of managing system evolution. Therefore object-oriented databases appeared as a response to these needs. MOOD (METU Object-Oriented Database Management System) is one of these new generation databases with its own original features. It is developed at iii the Software Research and Development Center of Scientific and Technical Research Council of Trkiye, METU. The dynamic function linker unit and object-oriented algebra of the MOOD are designed and implemented...

In this paper we present an indexing structure for ranges. The main idea is to map a bounded range to a point in one-dimensional space, using a standard B + -tree to index such a point. The indexing structure is storage-wise efficient, requiring O(N ), where N is the number of indexed ranges. Its maintainance is also efficient, requiring O(log N ) accesses per update. We also show that it is able to efficaciously answer a number of different queries, requiring, for instance, O(logN + L) I/Os to answer an intersection type of query, where L is the length of the range given in the query. An analytical performance study is also presented. Finally, we point out two immediate applications of this indexing approach, which are (1) answering stabbing queries and (2) the indexing of a valid-time (historical) database. Contents 1 Introduction 4 2 The MAP21 approach 4 3 Answering Range-based Queries 6 3.1 Location Query : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : :...

this paper we provide two example cases for this purpose and study the performance of IFD in the conventional sequential signature file (SSF) method and a new vertical partitioning environment, the multiframe signature file (MFSF) method, that we introduced in our recent study [10, 11, 18]. For this purpose we developed a test environment and implemented the SSF and MFSF methods. We extended these methods to use IFD and tested their performance with real data. The experiments show that IFD improves the performance of the inspected methods by reducing the observed FD and the (query) response time. (Further experiments with similar results involving a generalized frame sliced signature file approach are reported in Kocberber [18].) The organization of the paper is as follows. In Section 2, the conventional FD estimation method and the proposed FD estimation method, IFD, are explained. Section 3 explains the test environment used in the experiments. In Sections 4 and 5, we apply IFD to the SSF and MFSF methods, respectively, and measure the performance improvements obtained by IFD experimentally with real data. Section 6 provides the conclusion. In the Appendix we provide a formal proof which shows that under certain conditions the number of false drop records (FD) estimated by considering the average number of terms in the records is less than or equal to the FD estimated by considering individual D values of the records

by TOLGA BOZKAYA This thesis proposes index structures for efficient evaluation of temporal queries in temporal databases and similarity based queries in multimedia databases. To support temporal operators and to increase the efficiency of temporal queries, indexing based on temporal attributes is required. A temporal database can support two notions of time. Valid time is the time when a data entity is valid in reality, and the transaction time is the time when a data entity is recorded in the database. In this thesis, methods for indexing time intervals in transaction time and valid time databases are proposed. Transaction time databases are append only databases. Data is never deleted from the database, and data versions that were deleted or modified are stored as historical versions. This thesis proposes indexing current and historical versions of temporal entities separately to exploit the behavior of transaction time data. Two structures, namely IB tree and AD*-tree, are proposed...

Query optimisers are critical to the efficiency of modern relational database systems. If a query optimiser chooses a poor query execution plan, the performance of the database system in answering the query can be very poor. In fact, the differences in cost between the least and most expensive query execution plans can be several orders of magnitude. On the other hand, it can be prohibitively expensive for the query optimiser to search exhaustively for the least-cost (strictly optimal) query execution plan. Most query optimisers, therefore, compromise by using a reasonably cheap search to obtain a reasonably cheap query execution plan. Accurate, but inexpensive, query size estimation is fundamental to the success of real query optimisers. A number of studies [Christodoulakis 1984; Ioannidis and Christodoulakis 1991, 1993] have demonstrated that optimisers can select very expensive query execution plans if they are forced to rely on poor or inaccurate query size estimates. This thesis will address the problem of how to obtain reliable and accurate query size estimation for the cost calculation of query execution plans.

With the increase in mediated applications, there are now mediated application servers that support concurrent client accesses. Multiple query optimization (MQO) is essential for efficient processing of concurrent read only queries. The MQO problem in mediated environments differs from that of ordinary relational MQO systems in two ways: first, in mediated environments, the MQO system must be capable of interacting with the "local" query optimizers of the various sources being accessed. Second, in mediated environments, the notionof what constitutes a "common" subexpression varies dramatically from one source to another, and the MQO system must be capable of working with these diverse notions of "common" subexpression. In this paper, we extend classical relational DBMS MQO techniques to the case of mediated applications. Our architecture has a compile-time and a runtime component. In the compile-time component, the developer describes common subexpressions via a set of structures calle...

In this paper we complete the analysis done by Ramakrishna and Mukhopadhyay for a data node in the Bounded Disorder (BD) file organization of Litwin and Lomet, by introducing the B-tree index into the model. Also, we extend the analysis to the case of BD files with two partial expansions as proposed by Lomet. Our main contribution is a detailed analysis of search and insertion costs, and its comparison with B + -trees. 1 Introduction Nowadays there are two main file organizations: hashing and tree indexing. New hashing techniques achieve single access retrieval, but are very inefficient for range search or key sequential access. On the other hand tree indices preserve the key order with a higher search cost. Litwin and Lomet [9] proposed the Bounded Disorder (BD) file organization to combine the advantages of both methods. This paper complements the analysis presented by Ramakrishna and Mukhopadhyay [14] concerning the performance of BD files, by including the index in their model. ...

The availability of large main memories and the new technologies for disk drives have modified the models for external sorting and have renewed interest in their study. Little is known about the performance of traditional and more recent sorting methods on nearly sorted files although such files are common in practice. ffl We confirm mathematically that the lengths of the runs created by replacement selection during the first phase of external sorting increases as the order in the input file increases. Previous work has concentrated on the expected length of initial runs when all input files are equally likely to occur. It has long been accepted that when an input file has little disorder, the lengths of the generated runs will be long. We establish such results for two measures of disorder, namely, the number of ascending runs and the maximal distance between inversions. ffl We demonstrate that, during the merging phase, the floating-buffers technique not only reduces the sorting ti...