Last week I was prototyping a sample statistics view. This view will be the highest level of a trace view and helps to understand where are the potential problems of the system.
Now I have finished a phase of that and it works fine (and it is fast enough!) with even large trace files(I have tested with 1 TB Trace).
Now I am providing a report for that and trying to measure memory usage, speed and on the other hand, the overhead costs of that.