|
To remain competitive in today's marketplace the value of business data cannot be over-emphasized. There are compelling needs to understand business drivers in many areas: customer knowledge, supplier and internal process management, mandated compliance analysis and reporting, to name a few. The need for gathering an ever increasing amount of data and providing for storage, reporting and analysis of this data has placed enormous pressures on both IT staff and the computer infrastructure to deliver the right information, at the right time and for the right price. Microsoft SQL Server 2005, based on 64-bit technology, has been developed to meet the needs of world-class Business Intelligence (BI) environments. Microsoft's vision for SQL Server 2005 as an open database platform allows a wide range of deployment options for customers.
To harness the power now available with SQL Server 2005 and maximize the customer's return on investment, HP supports SQL Server 2005 with two basic BI infrastructure strategies. The first strategy is with the consolidation server where all the functionality of the BI ecosystem is located on one HP Integrity enterprise-computing platform. The other is a hybrid server strategy where the high system loading applications, e.g. data integration platforms, the data warehouse, large scale OLAP, etc., are on the HP Integrity enterprise server and low demand BI tools are deployed on smaller 32- or 64-bit servers. This paper presents a discussion of these options and presents some guiding principles for determining which strategy may be best for a particular customer situation.
There are two basic approaches to BI ecosystems. The BI component approach uses the core SQL Server 2005 database but you choose other BI tool and solution offerings to build out the BI ecosystems as you see fit. The integrated BI approach uses built-in functionality available with the SQL Server 2005 offering. The functionality needed by many customers can be met with the built-in BI capability of Analysis Services (SSAS), Reporting Services (SSRS) and Integration Services (SSIS). The choice between these approaches is discussed from a Pareto Principle point of view and we highlight some of the merits of each approach.
SQL Server 2005 is taking its place in enterprise-level BI implementations and offers customers a rich choice of options and approaches. HP's Integrity computer platforms offer the most powerful and flexible option for any scope and size of Windows SQL Server 2005 BI environment. In addition to the adaptability of the Integrity servers to support the broad array of BI deployment options for SQL Server 2005, the Integrity servers support the newest database availability technologies for SQL Server 2005: database replication enhancements, on-line reindexing, partial restores, and database mirroring. With the HP Integrity servers and HP enterprise storage in combination with SQL Server 2005 it now possible to build the business and mission critical BI IT environments our enterprise Windows customers have been asking for.
Most enterprise class servers are used in one of two types of business applications. The first application area is broadly called operational and uses the database in an On-line Transactional Processing (OLTP) configuration. These systems are about running the business on a day-to-day basis and represent a significant investment in capital and IT resources. The scope of this paper is to discuss the other application area, which we broadly call Business Intelligence (BI). The BI applications are about using the data from the business, significantly from the OLTP systems, and provide reporting and analysis to improve decision making and help understand and optimize the business. Thus, OLTP and BI applications are simply opposite sides of the same business coin.
A key reason why BI and OLTP systems infrastructure are treated differently is the nature of the workloads. A typical operational system support hundreds or even thousands of users, or other instances of workload, in call centers, reservation centers, banking transactions, manufacturing processes, web transactions, etc. These types of workloads create many small transactions with the database and the database is tuned and optimized for thousands of these small concurrent transactions. The OLTP database is often used in a highly normalized form with multiple tables that are easily and quickly updated. It is possible for the systems and database administrators to tune these systems to a high degree and forecast/plan capacity.
|
