Solving Operational Business Intelligence with InfoSphere Warehouse Advanced Edition

386 Solving Operational Business Intelligence with InfoSphere Warehouse Advanced Edition

򐂰 Data governance and data security are more difficult to manage after data

has been extracted from the enterprise warehouse environment.

򐂰 Data duplication leads to increased storage requirements and costs. There

are also concerns about consistency.

10.3.2 In-database scoring with SAS and InfoSphere Warehouse

A solution to the problems listed in the preceding section when using SAS

Enterprise Modeler and IBM InfoSphere Warehouse is to use new in-database

features in DB2 from SAS.

In this scenario, the scoring models developed in SAS Enterprise Modeler can be

deployed using the new SAS Scoring Accelerator for DB2 that benefits from the

SAS embedded process in DB2 to accelerate deployment and performance. The

process is depicted in Figure 10-15.

Figure 10-15 SAS Scoring Accelerator and SAS embedded process in DB2

6

The process with the new accelerator and embedded process includes:

1. Export Phase

Export the scoring model from SAS Enterprise Modeler into a DB2 mining

metadata format.

6

Adapted from “Change the game with in-database analytics: how to run SAS scoring models inside

DB2”, Lin, Eileen; Nicola, Matthias; Chatterjee, Subho, Information on Demand conference 2011.

Export Phase Publish Phase Populate Phase

DB2

Model

stored in

user table

SAS EM

SAS

Enterprise

Miner

Export

Node

DS2

Format

Metadata

SAS

Publishing

Macro

SAS/ACCESS to

DB2

Client

Applications

Data

DDL