SAS, a leader in business intelligence, announced plans to deliver SAS Intelligence Storage with SAS 9.1. SAS Intelligence Storage is an open, flexible platform that provides a variety of storage options to ensure optimal performance of business intelligence and analytic intelligence applications while keeping the cost of ownership and maintenance under control, and being open and easy to administer. Early adopters of SAS Intelligence Storage have experienced performance gains of as much as 30 percent over earlier releases.

Traditionally IT managers have focused their storage strategy on the timely and reliable collection of data. This focus resulted in the adoption of relational databases, which were designed and tuned to enable the efficient and reliable collection of data and processing large numbers of small transactions, such as banking ATM transactions and individual customer sales. Relational architecture is optimized for entering one transaction at a time, with more and more emphasis placed on the reliability of the data collection process at the price of performance. SAS Intelligence Storage is designed and optimized specifically for business intelligence and analytic applications.

Business growth and the emergence of new business units have led to IT data collection storage capabilities being put in place in isolation throughout the enterprise, creating what might be referred to as data islands to support the operational needs of individual business units. To further complicate matters mergers and acquisitions have become the norm leaving IT to manage heterogeneous operational systems, possibly on multiple system platforms such as UNIX or mainframe.

As organizations have decentralized the decision making process IT has been presented with the challenge of enabling individual business units, and their many users, to have access to the same unified view of all information held throughout the whole organization, through business intelligence and analytic intelligence applications, to make reliable and effective decisions. IT managers need to provide information dissemination capability that does not interfere with the operational systems currently in place capable of providing a unified view of multiple operational systems.

New usage patterns caused by the emergence of analytic and business intelligence applications are placing unprecedented strain on relational systems that were not designed to access large tables of information in search of particular user requested data and certainly were not designed to produce the summarized reports demanded by today’s modern business intelligence applications or intensive simulation and modeling requirements of modern analytic intelligence applications.

IT managers need to manage both a data collection and a data dissemination storage strategy. Relational systems have evolved to solve the data collection problem but now IT needs to find a way to address the dissemination issues critical to the future of most companies:

  • Lack of flexibility in enterprise storage options,
  • Lack of storage performance and scalability as enterprise data grows exponentially,
  • High cost of ownership and maintenance of traditional enterprise storage systems Increased Storage flexibility.

SAS Intelligence Storage was designed from its outset to provide relational, OLAP and parallel storage options to suit all types of application profiles from simple reporting and OLAP dissemination through to advanced forecasting and data mining. All these options are provided transparently to applications within an integrated meta data framework with meta data available to all applications and exposed to third-party environments via meta data interoperability.
Meta data interoperability is achieved through the use of the Common Warehouse Metamodel (CWM) standard. Hybrid OLAP makes it easy to configure and maintain multicube architectures with dynamic drill through to atomic detail and SAS’ powerful calculation. Statistical procedures always allow optimum choice of dynamically calculated or pre-aggregated measures.

The internal architecture of SAS Intelligence Storage fundamentally lowers the performance-scalability cost curve. Whereas the internal architecture of a general purpose RDBMS is optimized for the secure handling of concurrent updates, SAS Intelligence Storage is designed to minimize overheads and maximize I/O performance. This is achieved by using fixed record lengths, advanced indexing, large block sizes and by eliminating ‘two-phase-commit’ processing. Together with multi-threading and physical partitioning, these features typically produce order-of-magnitude improvements in loading and query performance compared to orthodox RDBMS configurations.

SAS Intelligence Storage is integrated with SAS’ powerful ETLQ capabilities in the SAS Intelligence Value chain that can access and read all relevant source data formats and can perform the necessary pre-processing needed, so that data is efficiently collected, transformed, cleansed and organized in preparation for storage. This guarantees end-to-end integrity of metadata so set-up and configuration of data extraction and transformation is much faster and ongoing enhancements are much easier. SAS Intelligence Storage’s reduced- overhead design not only enhances performance; it also simplifies installation and ongoing maintenance by the database administrator. SAS Intelligence Storage reduces the cost of information dissemination to the user by supporting MDX access so customers can re-use their existing query/reporting tools. These interoperability features ensure the availability of lowest cost infrastructure options. SAS Intelligence Storage is comprised of a number of interoperable components: SAS OLAP Server, SAS Scalable Performance Data Server, Base SAS Data Storage Engine, SAS Scalable Performance Data Engine.

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

Don't have an account? Register for Free Unlimited Access