FHIN: an efficient storage structure and access method for object-orienteddatabases

While relational database technology has dominated the database field for m ore than a decade, object-oriented database (OODB) technology has recently gained a lot of attention in the database community. Many researchers are c oncerned about the performance of OODBs. This paper proposes an OODB design methodology called fragmented hash-indexed (FHIN) that is aimed at improvi ng the operating performance of OODBs. The FHIN model's storage structure c ontains an Instances-Classes Table (ICT) with a two-segment data design. Qu ery processing is done by accessing data segments through ICT with an algor ithm introduced here. The FHIN model uses three access methods: hashing, in dexing, or hash-indexing. The database performance of FHIN is compared to t wo previous access methods using 1050 simulation runs. Results indicate tha t the FHIN model is 43% better than either of the other models in smaller d atabases, 65% better in larger databases, 50% better under conditions of hi gh updating, and 72% better under conditions of low updating. These results suggest that FHIN methodology has promise and is worthy of exploration and OODB software development. (C) 1999 Elsevier Science B.V. All rights reser ved.