A Storage Server for the Efficient Support of Complex Objects

Complex objects are a major advance achieved in modern database systems (DBS) to provide powerful data modelling capabilities. A complex object is highly structured, containing many dependent or independent subobjects. Most of today's DBS are based on the traditional two-level DBS architecture consisting of a storage and a logical data model level. In these systems, the structural aspect of a complex object is managed on the data model level, but lacks speci c support from the storage level which could enhance overall DBS performance. In this paper, we present the concepts and implementation of the KIOSK storage object server designed to e ciently load complex objects from secondary storage. KIOSK introduces the notion of complex storage objects which facilitates modelling the structural aspect of complex objects at the storage level. KIOSK provides operators to load a complex storage object in its entirety from disk in one access and additionally allows clients to retrieve dynamically-de ned portions of a complex storage object. Benchmark results presented suggest that using complex storage objects accelerates the processing of complex objects in a DBMS by a reasonable factor due to minimized load and store time.

[1]  J. Eliot B. Moss,et al.  Working with Persistent Objects: To Swizzle or Not to Swizzle , 1992, IEEE Trans. Software Eng..

[2]  Silvia Nittel.  Einsatz und Nutzen von objektorientierten Betriebssystemen für die Implementierung von Datenbanksystemen , 1993, Datenbank Rundbr..

[3]  Klaus R. Dittrich,et al.  DAMOKLES - A Database System for Software ENgineering Environments , 1986, Advanced Programming Environments.

[4]  Jack A. Orenstein,et al.  The ObjectStore database system , 1991, CACM.

[5]  Guy Bernard,et al.  The O2 Object Manager: An Overview , 1989, Building an Object-Oriented Database System, The Story of O2.

[6]  Alfons Kemper,et al.  Adaptable pointer swizzling strategies in object bases , 1993, Proceedings of IEEE 9th International Conference on Data Engineering.

[7]  David J. DeWitt,et al.  The oo7 Benchmark , 1993, SIGMOD Conference.

[8]  Vivek Singhal,et al.  Texas: An Efficient, Portable Persistent Store , 1992, POS.

[9]  Hans-Jörg Schek,et al.  Architecture and implementation of the Darmstadt database kernel system , 1987, SIGMOD '87.

[10]  Jay Banerjee,et al.  Operations and implementation of complex objects , 1987, 1987 IEEE Third International Conference on Data Engineering.

[11]  Elisa Bertino,et al.  Composite objects revisited , 1989, SIGMOD '89.

[12]  David J. DeWitt,et al.  Object and File Management in the EXODUS Extensible Database System , 1986, VLDB.

[13]  Alexandros Biliris.  The performance of three database storage structures for managing large objects , 1992, SIGMOD '92.

[14]  Steven Sinofsky,et al.  Managing Persistent Data with Mneme: Designing a Reliable, Shared Object Interface , 1988, OODBS.

[15]  Klaus Meyer-Wegener,et al.  PRIMA - a DBMS Prototype Supporting Engineering Applications , 1987, VLDB.

[16]  Alexandros Biliris.  An efficient database storage structure for large dynamic objects , 1992, [1992] Eighth International Conference on Data Engineering.

[17]  Eugene J. Shekita,et al.  Cricket: A Mapped, Persistent Object Store , 1990, POS.