Putting data to work

In what ways do the latest generation of enterprise data warehouses, data marts and analytics applications better respond to the business intelligence needs of user?

Michel Kisfaludi. Most data warehouses and data marts are better today only because computer hardware is bigger, faster and cheaper. The potential benefits to the business user are slight in the sense that more data can be used to develop business information, and more people can query and analyse the data faster. However, in order to keep up with and adapt to ever changing business requirements, data warehouses and data marts are now being built using technology which not only stores all of company's data on less disk space, but allows faster data access, better data flexibility, and user driven ad hoc analysis.

How important is it that data warehouse appliances allow for complete ad hoc analysis and data analytics flexibility?

MK. This is of paramount importance. Complete ad hoc analysis is the ultimate frontier from a technological perspective but we should remind ourselves that it is a very basic expectation of the business user. Wouldn't it be strange if a car buyer was told he can only drive his car on certain streets and roads at certain times of the day? Business success is measured by a company's ability to outsmart and better execute than its peers. Only with complete ad hoc analysis and flexible data analytics can a company challenge its own assumptions and move beyond the obvious to find new value for its business.

What are the main advantages of value-based storage as opposed to record or column-oriented data storage structures?

MK. The number one advantage to value-based storage (VBS) is the 100 percent indexing of incoming data values allowing each value to be stored just once and the correlation of every unique value. It is on this cornerstone that the main advantages over record or column-oriented data storage structures are built upon: the ability to load all of a company's data assets into a data warehouse and being able to ask any and all business questions, and better leveraging those answers into information and business insights; elimination of planning and structure design;  data warehouse schema is data-generated during the loading process which drastically shortens an otherwise very long and costly development process: In fact, data can be loaded and ready for ad hoc querying and analysis in days or weeks rather than many months;  Every query, regardless of the combination of fields used, is processed as quickly as an optimised index can locate the information, usually in less than one second. In addition to simple yet powerful ad hoc querying, advanced analytics, such as Google-type data searches and dynamic segmentation querying, are put at the user's fingertips.

How does agile data warehousing allow businesses to make better decisions faster, keep abreast of the latest trends, anticipate customer needs and better manage operations? 

MK. The purpose of agile data warehousing is to provide unfettered access to company data. To accomplish this, a system must be provided that supports all company data assets, is able to incorporate new data sources as they come on-line, and adapts to changing business requirements and trends. Once a user can gain access to this singular data source free of limiting data architectures, true agility will be achieved. Simply put, companies should first solve the data container problem with appropriate technologies then support it with simple, flexible yet powerful set of interfaces.  

How do you see data warehousing platforms developing in the future? Are there any game changing developments or technologies on the horizon that you are particularly excited about? 

MK. One of the most exciting developing technologies that will influence and enhance data warehousing is multi-core chips. Once on the market, they will put massive computing power into a single machine which will be a performance boon for the data warehouses built on technologies which can leverage that additional power without suffering from the restrictions of a single machine's I/O capacity. In the near future, data warehouses will be expected to contain all of a company's data rather than data subsets that can only answer specific questions. The increased data accessibility and usability will become a commodity rather than a luxury. The emphasis will then be on improved flexibility, agility, and greater ad hoc analysis and data analytics. As these developments take root, metadata will become the key component of the data warehouse and drive developments in technologies to provide metadata that is more scalable, better adaptable to change, and is easy and quick to develop or better still, generated.

Biography

Michel Kisfaludi, 44, is a former international telecom and banking executive. Over the last 10 years, he has run the US$500 million retail banking division of Deutsche Bank in Spain after managing its marketing department. Previously, he ran business development projects for Dassault and France Telecom in China and Mexico.