Andrew Perkins calls himself an IT specialist on IBM's Z Warehouse SWAT team. He performs technical presales for InfoSphere Warehouse and DB2. InfoSphere is a superbrand; InfoSphere Warehouse is a product providing a core basic warehousing set of tooling for System Z: physical data modelling; a a DB2-based ETL tool; an OLAP server provides Cubing Services, a multidimensional caching system.

 

Andrew is at SHARE to present an overview of InfoSphere Warehouse on System Z. Architecturally, InfoSphere on Z is a hybrid. The runtime server runs in an IFL under Linux on Z. The data store is DB2 for Z. The development tool is an Eclipse -based IDE running on a workstation. From a programming point of view, another flavor has been added to the versatile and layered liqueur that is mainframe programming.

Andrew enjoys his job. He's happy to be back in mainframing after a decade-long stint in distributed systems. A decade ago, data warehousing and business intelligence on mainframes was expensive and the tools were lacking. In the distributed world you had all the GUIs and ETL tooling. On mainframes it was hand-coding. The approved way of doing data warehousing was 1) extract data from mainframe; 2) throw workstation-based reporting tool at data. The data itself was not considered critical; it was not live data, but duplicate data off to the side.

But over the past 3-5 years customers were still doing data warehousing on the mainframe despite industry recommendations. The data had become more valuable to the enterprise. The reporting needed to be integrated with the transaction processing. Take online shopping: You find something you want, the web server gives you the customer recommendations garnered in realtime. All that ability to guide and enhance the customer's shopping experience comes from live access to warehoused data. The upsell, the cross-sell is facilitated thereby.

Also, in terms of data security, on-host warehousing means all data is processed the same way with the same security infrastructure in place.