Database
Design Issues in Peer-to-Peer Data Warehousing
Abstract:
Data warehousing is an active research area due to the
popularity and widespread use of data warehouse and OLAP technology in the industry
today. There are many interesting research problems attractive to the database
community. With business environment and requirements changing, there is a
potential trend of moving centralized data warehouse to distributed data
warehouse to improve the decision making efficiency and gain computing resource
cost-effectively.
Peer-to-Peer (P2P) data warehouse is a distributed data warehousing
architecture, which allows cooperative data interchange and dynamic matching
functionalities between peers without a global schema. In a P2P data
warehousing architecture, a collection of local data warehouses is distributed
in a P2P network. Each data warehouse holds only the dimensions of
interest at that node and the fact table is constructed from the local
dimensions. Each fact table is populated locally and queries need to be
evaluated globally.
This presentation first will illustrate the concept and architecture of P2P
data warehousing. Then we will emphasize on the data model selection for P2P
data warehouses by analyzing different data model techniques for data warehouse
design as well as the characteristics of P2P data warehouse environments.
Finally, we will highlight the potential research issues on P2P data
warehousing.