The Arizona State University Dan Ware-house was created to give faculty and staffaccess to data previously trapped in infor-mation systems designed for on-line pro-cessing. Our data was extremely secure andvery accurate but not accessible.Development of the Warehouse startedin the summer of 1992 championed by theDepartment of DataAdministration.Twelve individuals from Data Administration and Information Technology formed ateam to "prove" the warehouse conceptin i client set ver environment. A representative group of ASU staff was selectedto serve as pilot users to test the datawarehouse and access software. The teamdesigned,a student warehouse model basedon over 200 questions considered difficultto answer by pilot users using current in-formation resources. It was now possibleto extract data from the operational IDESdatabase and load a Sybise/SQLServer(tm)database. We had the first client serversapplication in place at ASU. The Data WarehouseToday ASU's data warehouse resides in a clientserver environment where processing oc-curs on both the client (front end) andserver (back end). As seen below, we ex-tract data from the mainframe and load itonto a UNIX server running an RDBMS.ASU's warehouse server is a Sun/Sp1rc630(tm) with 512 megabytes of memoryand two processors, running the Sun Sohris2.3 operating system.The RDBMS is SybaseSQL Server release 10.x. Currently11gigabytes of data reside on the server. Us-ers connect through Ethernet to the ware-house over ASU's network backbone viaTransmission Control Protocol/InternetProtocol (TCP/IP).The front end consists of a GUI dataaccess tool running identically in both theMacintosh and Windows environment. Theminimum hardware and software recom-mended for the PC client is a 486/66 MHzprocessot, 8MB RAM for Windows 3.1(16MB for Windows 95),6MB of space onthe hard drive for software,and sufficientspace for the data, an ethernet card andTCP/IP software -- WINSOCK. The mini-mum hardware and software recom-mended for the Macintosh client is a 68030processor running System 7.1, 8MB RAM,1OMB of space on the hard drive for soft-ware and sufficient space for the data, anethernet connection,and TCP/IP software-- TCP/IP control panel.GUI tools on the front end build struc-tured query language (SQL) requests andbring the results back to the client machine.SQL is the dominant query language foraccessing relational databases ind is ac-cepted by most non-relational databasesin a standardized form. The retrieved dataresides on the client machine, empower-ing and liberating the user. They can utilizetheir favorite spreadsheet, word processor,graphic tools, etc., to format data at will.The STUDENT, FINANCIAL, HUMANRESOURCES, and COURSE databasescomprise the foundation of the warehouse.These databases contain a large volume ofdetailed data. Updates occur or, a weekly,bi-weekly, or monthly basis, depending onthe database. The OFFICIAL database con-tains census values for frequently used datathat are the basis for published reports. TheOFFICIAL database contains a dozen tablesinstead of the 100 tables of the STUDENTdatabase. The small size makes it a betterstarting point for new warehouse users.The data here are frozen at the same pointin time each semester, making it the idealdatabase for trend analysis over time. Pro-viding "official numbers" in the data ware-house greatly improves warehouse cred-ibility.The four-hour introductory training ses-sion at ASU focuses on the query tool, thelogic, the data, and its appropriate use.After 1n extensive review of data accesstools, Brio Technologies' Brio Query is ouraccess tool of choice. We found it to bcthe best for training since Macintosh andPC versions are identical and it is simpleto use. Logic training on join methods and SQL operators improve the use of users learn how to identify graduate students,and degree/major codes, forinstannce. Classes emphasize the appropri-ate use of the data. The Buckley Amend-ment and Family Educational Rights andPrivacy Act (FERPA) which govern releascof student and employee data are discussedwith examples.While training reduces the number ofdata warehouse questions, a support in-frastructure handles other support needs.There is an E-mail address (ware-q@asu.edu)for user questions or problems.Experts on warehouse data, networking,ind d1t1 access tools respond within 24haul s. Users can also chill a central helpline. A file transfer protocol (FTP) site isavailable to download postscript copies ofdocuments associated with the data ware-house, copies of the data models, and common queries built by users or the ware-house team. The Warehouse Users s Group(WUG) meets monthly to share queries,educate members,and provide feedback tothe warehouse team. There are over 250WUG members and nearly 500 Warehouseusers in total. [Next Page]