Tez No	İndirme	Tez Künye	Durumu
38439	Bu tezin, veri tabanı üzerinden yayınlanma izni bulunmamaktadır. Yayınlanma izni olmayan tezlerin basılı kopyalarına Üniversite kütüphaneniz aracılığıyla (TÜBESS üzerinden) erişebilirsiniz.	Paralel veritabanı sistemlerinde veri parçalanması ve parçalamaya bağlı olarak işlemlerin paralelleştirilmesi üzerine bir çalışma / Yazar:OĞUZ DİKENELLİ Danışman: PROF.DR. TURHAN TUNALI Yer Bilgisi: Ege Üniversitesi / Fen Bilimleri Enstitüsü Konu:Bilgisayar Mühendisliği Bilimleri-Bilgisayar ve Kontrol = Computer Engineering and Computer Science and Control Dizin:Paralel sistemler = Parallel systems ; Veri tabanı sistemleri = Database systems	Onaylandı Doktora Türkçe 1995 175 s.
ABSTRACT Paralel database systems have become a major tool for high performance information processing. These systems require efficient declustcring approaches to partition each relation and to allocate them to the parallel architecture. If the database is not partitioned in a balanced way, the execution of database operations might waste resources, reducing the performance of parallel system. In this thesis, a new declustcring approach is introduced which uses a mulliatlribulc file structure as the basis for database partitioning. In addition, unlike the previous multiattribute declustcring strategics developed approach supports variable declustering where each relation is allocated to the required number of nodes. An analytical workload model is derived to find the resource requirements of each relation. Then, balanced blocks are generated from the multiattribute file making the number of blocks equal to the number of processors specified by the analytical model. Also a parallel file structure is developed for supporting decluslering and parallel execution of queries. The overall system, therefore, combines all advantages of variable and multiattribute declustcring together with efficient processor mapping and parallel file support. In addition a join methodology, which uses the introduced declustering approach, is developed The join is efficient since the allocation strategy always achieves compatible grid partitions among the joining relations with differing grid ranges as well as different join attribute value distributions. in