Arbor Software is putting the finishing touches to Arbor Integration Server, the company`s response to the data mart-data warehouse conundrum. At the heart of this problem is the need for rapid deployment of multiple, discrete data marts to address urgent business requirements, set against the need for centralised control of the data warehouse.
Arbor Integration Server, due for commercial release shortly, employs the hub-and-spoke, or multi-tier data warehouse architecture, where application-specific OLAP data marts run against an application-neutral data warehouse.
"Arbor Integration Server will help clear up confusion over the respective roles of the relational database and the OLAP server in the data warehouse market," says Dave Fawcett, director of the IBD Group, local distributor for Arbor Software. "Used as complementary rather than competitive products they can now offer data warehouse users a complete solution."
Industry and analyst response has been overwhelming.
"We have always believed it is vital to deliver initial successes quickly, through data marts," says Fawcett. "The marts can also help define the requirements for the data warehouse, which is architected in parallel. With Arbor Integration Server we can recycle the mart feeds into the data warehouse, and feed the data marts from the data warehouse - it becomes a two-way process."
Arbor Integration Server is a suite of graphical tools and scalable data integration services. Together they help manage OLAP metadata, provide reusable OLAP design, and minimise the time and expense needed to create, deploy and manage analytic applications, says Fawcett.
"Using Arbor Integration Server, customers will be able to create centralised, reusable enterprise OLAP metadata; create on-demand analytic applications; and leverage data warehouses and packaged applications."
Arbor Integration Server does this through the rapid assembly of reusable dimensions, hierarchies and calculation logic centrally maintained in a shared, enterprise OLAP metadata catalogue. This metadata catalogue ensures synchronisation between the data marts and the data warehouse, and ensures all business users have access to one version of the truth, no matter how many data marts they have deployed.
Arbor Integration Server consists of two distinct components:
- OLAP Architect transforms relational data structures into multidimensional models using industry standard techniques. Data challenges such as datatype conversion, relational recursion and denormalisation are invisible to the end-user. To provide access to metadata across the enterprise, all OLAP models are stored in a central repository.
- OLAP Builder, which turns OLAP models into working Essbase databases.
Industry reaction has been enthusiastic, with more than 20 leading data warehousing and applications partners endorsing the new offering while it is still in beta. These include Red Brick, SAP, PeopleSoft, Oracle Applications, Sagent, IBM, Walker Financials, Hyperion, Lawson and Arthur Andersen.
Praise from analysts
Analysts have overwhelmingly praised Arbor Integration Server.
"Arbor Integration Server is a welcome solution to the dichotomy between centrally administered data warehouses and functional OLAP data marts," says Nigel Pendse, lead author of The OLAP Report.
"This new direction ensures that Arbor`s customers will not fall into the potholes of data mart chaos," observes Claudia Imhoff of Intelligent Solutions.
"Arbor Integration Server makes enormous sense. It really solidifies Arbor`s position as an enterprise player," says Bob Craig of Hurwitz Group.
Share
Editorial contacts