EXPERT RESPONSE
I'm abridging what could be a protracted answer to a complex question. But you're right; it's not all about the addition of linear processing capabilities.
Indeed, there a few different dimensions to consider when planning to scale your BI environment. For instance, there is the "mixed workload" factor -- query processing versus ETL. Adding additional processing power and memory is the most effective means of addressing incremental growth of data. It's important to realize however, that there's a point at which the physical table structures -- not to mention the physics of disk I/O -- will prevent the linearity of queries (e.g., in order to support x percent new data, add x percent new storage and CPU). Because ETL is much more I/O intensive than standard query processing of most BI environments, simply adding more storage and CPU can be impractical. Bottom line, you can't just throw more horsepower at more data.
|
