In spite of the fact that we are now living in the information age, one of the most interesting paradoxes has been the failure to turn mountains of data into validated decisions useful as cornerstones of business enterprise. Without digressing into a history lesson of file storage methods, suffice it to say that corporations have largely fallen short of delivering on the revered "IT promise" to put useful information in the hands of the business users and executives for making sound and reliable decisions. Today, companies are taking the next step toward making good on the promise the enterprise data warehouse. What business user would not want access to information beyond the limited view of departmental data to which they have been confined? What executive would debate the value of having a well-defined and consolidated repository of corporate data for the business planning process? The answer is hopefully as obvious as the need.
Sadly, a common outcome from early warehouse implementations is one of disappointment. But before we relegate the data warehouse to the realms of the Edsel, we must consider the warehouse for what it is and is not. Costly? Yes. Time-consuming? Yes. Painful? Usually. Necessary? Definitely. The last step? An emphatic NO.
Without debate, the outcomes of data warehousing initiatives come at a price. However, once completed properly, the unified data dictionaries, the cleansed data, the integration of disparate data sources into a common view and the sheer convenience of having a consolidated repository from both functional and operational views all add to the list of benefits that substantiate the definitive need. But it is not the last step.
The resultant gigabytes or terabytes of information are still not, and cannot be, maintained in a manner conducive to decision support activities. At last though, a handful of products are now available that make effective decision support just one step away.
There is growing momentum in the marketplace for what is being labeled the "exploration warehouse" or "exploration mart." The key theme of data exploration is, "Ask any question of any data at any time efficiently."
Since these products are indeed quite new, a quick overview of key features is needed.
- Complete Indexing: One of the main issues relating to traditional databases in the world of decision support systems (DSSs) has been that the data is not indexed in a manner that allows efficient processing of truly ad hoc queries. Since ad hoc queries promote open thinking and open thinking is an iterative process, questions asked in exploration (the new DSS) systems are not constrained by what the database is tuned to handle. Thus, an "index everything" approach is required to effectively and efficiently support open thinking.
- Ease of Administration: Owing to the complexity of data storage implementations and the inherent need for constant monitoring and tuning, traditional database administration is done by a very specialized, highly skilled group of persons known as database administrators (DBAs). The operational data stores (ODSs), on-line transaction processing (OLTP) systems and the newly created data warehouses (DWs) have all added to the workload of the DBAs, thus enlarging the backlog of requests for information. Due to this, the exploration DSS systems must provide a facility by which the data to be explored can be administered with little or no DBA requirement.
- Scalable: The tiered computing architecture that is so pervasive requires a level of scalability that is not typically found even in systems that tout it. Scalability of storage, memory and CPU are all critical to the point of being base requirements. A few products in the market today are reaching the point where "in-memory" databases of respectable size are attainable. By accomplishing this, linear and at times super-linear scalability can be achieved.
- Open: Tools in the marketplace that support a variety of market niches and specialized applications are many. Most of these tools support open access via Microsoft's de facto standard 16- or 32-bit ODBC interface. While many databases tout a list of proprietary front- end data access tools or tools that write to a proprietary API, the ODBC interface (especially in 32-bit) provides enough throughput to support DSS activity.
- Fast: This is perhaps the single-most important attribute of exploration systems. In today's competitive world, the questions that have the greatest impact usually have a short window of opportunity. Due to this, each tick of the clock from the time a question is conceived until the answer is received has a cost. Therefore, speed of implementation, formulation and execution must all be considered. In many cases, customers currently engaging in exploration activities carry this speed to the point of one-off "exploration marts create it, use it, obtain the value, dispose of it then do it again for the next question.
The environment of exploration promotes a level of decision support that finally delivers on "the promise." The delivery of these tools combined with the higher degree of computer- knowledgeable users creates an opportunity for those "power users" to explore data in ways previously unimaginable. Decisions can be made more accurately by looking at the data in ways never before possible and, thanks to the data warehouse and the exploration warehouse, that is cleaner and more well- defined.
Register or login for access to this item and much more
All Information Management content is archived after seven days.
Community members receive:
- All recent and archived articles
- Conference offers and updates
- A full menu of enewsletter options
- Web seminars, white papers, ebooks
Already have an account? Log In
Don't have an account? Register for Free Unlimited Access