Abstract: OLAP model is the foundation for implementing efficient on-line analysis processing,data mining and advanced business intelligence applications.ROLAP,MOLAP and OOLAP models are well known models in the current data warehousing and business intelligence research areas.In this paper,the authors propose a novel multi-data warehouse model called Multi-OROLAP model based on multicast services,which has the advantages such as scalability,efficient computation,and distributability.The authors focus on the Multi-OROLAP model and algorithms used within this model,including the incremental refreshing algorithms for data warehouses,storage optimization algorithms,transparent exchange algorithms,and adaptive selection algorithms for main and secondary memory objects.Typical user query and analysis patterns and their performance impact on the Multi-OROLAP model and algorithms,group membership management and maintenance of multi-data warehouses are also discussed.

Key words: multi-data warehouse, multi-OROLAP model, multicast services, distributed computing, optimization algorithms