In-memory business intelligence: a Wits context

Abstract

The organisational demand for real-time, flexible and cheaper approaches to Business Intelligence is impacting the Business Intelligence ecosystem. In-memory databases, in-memory analytics, the availability of 64 bit computing power, as well as the reduced costs of memory, are enabling technologies to meet this demand. This research report examines whether these technologies will have an evolutionary or a revolutionary impact on traditional Business Intelligence implementations. An in-memory analytic solution was developed for University of the Witwatersrand Procurement Office, to evaluate the benefits claimed for the in-memory approach for Business intelligence, in the development, reporting and analysis processes. A survey was used to collect data on the users' experience when using an in-memory solution. The results indicate that the in-memory solution offers a fast, flexible and visually rich user experience. However, there are certain key steps of the traditional BI approach that cannot be omitted. The conclusion reached is that the in-memory approach to Business Intelligence can co-exist with the traditional Business Intelligence approach, so that the merits of both approaches can be leveraged to enhance value for an organisation.