Once, the workload of data warehouse systems was relatively homogeneous with the same types of queries being executed repeatedly. However, due to the ever-expanding reporting and analytical needs of today’s business users, the characteristics of analytical workloads have transformed significantly. Users are now asking for more complex reports, new forms of analytics, more queries, and so on. In addition, the introduction of operational BI, where users require near real-time access to operational data, has resulted in systems in the need for data to be refreshed more frequently.
To support this new workload demand, data warehouse systems were originally designed around a large number of databases, with massive amounts of stored data being duplicated across classic SQL databases, multi-dimensional databases, and data warehouse appliances. In most situations, data warehouse architects opted for this type of architecture in order to handle the fluctuating workloads that are typical of today’s business demand. Unfortunately, this type of traditional data warehousing architecture can be costly, complex and difficult to sustain.
To help simplify today’s data warehousing architectures, companies should ensure their database servers meet the following three essential requirements:
- Process mixed workloads in an efficient and stable way (addressing a wide range of query types).
- Handle shifting workloads (planned/unplanned increases in query workloads).
- Support high-concurrency workloads (processing many queries and other database operations simultaneously).
Access this presentation transcript to learn IT expert Rick F. van der Lans's insight on the importance of deploying a data warehousing architecture that can withstand the many complexities of today’s dynamic business environment.