You're familiar with the rationale behind business information (BI) warehousing, and there's no question that it will greatly enhance your enterprise business performance. But if your ERP world is SAP R/3, a BI warehouse may be closer than you think.
Build on the land you own
The first major truth of data warehouses is that they are most easily built on the land you already own. Part of the point is to construct a new data access mechanism on top of the environment you've already built. If you're running SAP R/3, you're simply adding on to the house. SAP Business Warehouse (BW) inherits the multitiered architecture of SAP Web Application Server and provides its own development environment. It's not actually part of SAP R/3, but it's pretty close.
Understand that you don't have to make SAP BW your warehousing software if you're running SAP R/3. Nor do you have to be running SAP R/3 to run SAP BW; it's standalone, and you can run it on top of anything, including Oracle.
Your task list
In setting up a business information warehouse, you have a number of tasks that must be accomplished before and during the implementation. These include:
- Supplement your physical facilities with additional storage.
- Identify all internal and external sources for data that will be loaded into the warehouse.
- Establish the necessary transforms and integrations that must be performed on similar data derived from differing sources.
- Create metadata to describe this data (SAP BW will give you two classes of metadata: business content, or pre-defined metadata, and client-defined metadata).
- Establish, in cooperation with those users in your company who will be developing analytics, optimum levels of granularity for the data you'll be warehousing.
Right away you're ahead of the game with SAP BW. The tasks above are facilitated by SAP BW's integrated architecture, which provides utilities for metadata services and Extract-Transform-Load (ETL). You also have an Administrator Workbench, which facilitates data modeling, testing, reporting, and monitoring of your warehouse.
Within ETL services, SAP BW provides you with a staging engine for managing the transport of data into the warehouse from diverse sources. This engine will do all the required massaging of inbound data. It also gives you a DataSource Manager for managing the different interfaces required for importing data from both SAP and non-SAP sources (these can include XML and DB Connect interfaces, along with file interfaces and SAP-based APIs).
In short, SAP BW is designed to give you one-stop shopping. These implementation/development tools are complemented with the SAP suite of process development features, described below.
Understanding the highly refined and integrated nature of data culled from data warehouses for business information analytics, SAP BW creates a layered architecture that applies physical, logical, and business-level services to stored data in a manner that is at once efficient and flexible. As you might guess, it's extremely hard to achieve this when building from scratch. The layers include, from start to finish:
- ETL Services
- Storage Services
- Analysis and Access Services
- Presentation Services
Necessarily integrated into each of these layers are the Administrative Services and Metadata Services.
This style of architecture is fairly typical of data warehouses in general. It's particularly well implemented for SAP BW because of the Web Application Services fringe benefits, which include interfaces for various operating systems and databases, integration with third-party portal infrastructures, heavy Java support, and full support for open protocols (HTTP, HTML, XML, SOAP and others)—all of which accrue to SAP BW. Consider the advantages in making your warehouse available via the Internet.
Turning SAP BW data into Business Information
SAP BW gives you every advantage in getting data into your warehouse. As intimidating as this part of the job is, it's actually the easy part. The challenge in implementing a data warehouse, SAP or otherwise, is that you don't really know ahead of time what the end result will be. Business intelligence—or analytical knowledge upon which decisions can be based and upon which your company's performance can be measured both internally and in the marketplace—is different for every company and it is probably new to you, if you've never implemented a data warehouse before.
You'll need to rely on users in various departments to develop the analytics that will deliver reliable metrics for performance at their level (and you'll ultimately base companywide performance monitoring on these lower-level metrics). The problem is they'll define it as they go—it won't happen overnight. On the other hand, their analyses will grow better and better over time.
In such an iterative process, your best friend is a set of analytical and presentation tools that accepts a wide range of data structures and offers a wide range of query, processing, and report interface options.
What SAP BW offers
What makes SAP BW a superior choice? Remember that SAP R/3 or any other operational transaction processing environment will permit almost any data warehouse to be built alongside it. But if you choose SAP BW, you have several tactical advantages.
Online analytical processing (OLAP), the core of any data warehouse analysis capability, is complemented in SAP BW by a high-efficiency query repository that defines and stores all queries, regardless of presentation source, for ready access by the OLAP engine. This metadata repository speeds data reads and collation and helps you optimize warehouse performance, particularly in a high-volume user environment.
Third-party data mining support
Data mining is particularly well adapted to data warehouses, though it can also be applied to OLTP history databases. If you have an investment in third-party data mining software, SAP BW may allow you to stay with what you have because it supports a number of third-party data mining engines. In addition, SAP BW supports several data mining interface standards, including Predictive Markup Language and OLE-DB.
This is the true gem in SAP BW and may be a decision driver for you. The Report-to-Report Interface (RRI) enables you to jump from query to query. This means you can link information together by context. This is a tremendous time-saver and organizational tool for end users.
How does it work? Analysis paths can be linked to receiver objects or other queries. But a receiver object can also be a SAP transaction, a Web application, even a remote Web link. You can use this mechanism to do Web navigation, to dump information into a report you're compiling, to build case studies—there is almost no limit to the flexibility of this feature. And the power it gives your data warehouse in servicing users and customers via the Internet is obvious.
You'll have to hit the books to fully strategize how to make SAP BW work best in your house, but as an enterprise-driven business intelligence framework, it is worthy of candidacy in your list of data warehouse possibilities.
Scott Robinson is a 20-year IT veteran with extensive experience in business intelligence and systems integration. An enterprise architect with a background in social psychology, he frequently consults and lectures on analytics, business intelligence and social informatics, primarily in the health care and HR industries.