The current implementation of the DataSource only allows for footer aggregates during the fetch of data. When we f.i. use virtual scrolling in a grid for each fetch of data from the server we also have to calculate the aggregates. In many cases that is a time consuming process. F.i. we have a project that has an activitylist containing over 100.000 rows. We only show 25 rows with skip(page) and take(25) but on each roundtrip we calculate the aggregate for all 100.000 rows.

A better solution would be to get the aggregate asynchronous from the data fetch and only fetch the first time and when a change has taken place in one or more fields that have an aggregate result.