Business intelligence, like all long-standing technologies, has been through many cycles of hype over the years. Ten years ago, the companies that focused on ROLAP-based BI ballyhooed about how relational technology was superior to cube technology and that cubes would eventually fade away. Then five years ago, operational BI vendors began assailing the value of traditional BI, claiming the newer technology could look at data as it moved in real time throughout an organization to achieve such things as business activity monitoring. Traditional BI was doomed, they said.

Clearly, the news of the demise of these technologies was, as they say, greatly exaggerated. In fact, it can be argued that they are stronger due to their ongoing evolution.

Today, the marketing momentum is toward "in-memory" business intelligence. Like every hype cycle, there is value in the new technology, and there are also some drawbacks. This article will outline how in-memory systems differ from traditional BI, explain its value, and also discuss what type of organizations would benefit most from in-memory BI. It will also discuss a potential outcome of the push based on historical precedence.

Defining In-Memory Business Intelligence

In-memory BI is a new form of business intelligence where the BI software loads the data in memory before any query or analysis is performed. This approach gives faster query performances and shorter implementation times, as opposed to traditional disk-based BI solutions.

The in-memory systems tend to be less expensive, largely because they have less complex analytical engines. They can analyze data, but at a simpler level, making it trickier to delve more deeply into the underlying data that resulted in the query answer. They can also hit their scalability ceiling quickly when large data sources are queried for moderately complex questions.

For many companies with straightforward business needs (like tracking sales, demand and marketing, for example), in-memory systems can be a very good fit -- though it might also make sense to evaluate many of the traditional reporting tools that have begun to move upstream by improving their analytics. If, however, a mid-sized business plans to expand, or is an information-rich organization, there are still many reasons why a traditional BI vendor would be a better solution to pursue.