Application factors

Two factors affect hardware requirements. First is the user load distribution. At most customer sites, the load distribution varies according to the nature of the application, but all applications have peaks and troughs.

Peak periods may have a daily or weekly rhythm for analytic or reporting applications. Budgeting applications experience high peaks during the budgeting process, with the heaviest loads occurring just before deadlines. Management reporting applications experience their highest demands during the period after new results are available. For correct sizing, these peak periods must be anticipated, and equipment sized accordingly.

The second factor is the complexity of the application. There are several factors that can affect the suitability of these recommendations, including, but not limited to, amount or level of data, number of dimensions, number of elements in dimensions, formulas, number of databases and applications, complexity of report definitions and number of users.

The hardware sizing recommendations were derived by using a series of customer reports of average complexity to estimate peak user load. In the current architecture, this memory consumption was recorded:


Component	Scaling server memory requirement per user
Web-based main access to dashboards and d/EPM Administration using OLAP.	200 - 2840 MB A simple report opened in Application Studio Service and dashboards takes 20 - 40 MB. A large report with a lot of data can be much higher, depending on the size. For example, each user may open multiple reports inside a dashboard.