Hi,
Queries on auto summary data on tier-1 are much faster compared to fetching data from "normal" history blocks. This is especially noticable when fetching large time periods with high value count.
When data is replicated to tier-2, it seems like tier-2 creates auto summary data as well. This is confirmed when we inspect the auto summary history blocks. But the queries against tables like "AnalogSummaryHistory" do not fetch data from autosummarydata, they use the normal history blocks instead. So in certain cases we are can query the tier-1 and get results in less than 1 second, while the same query on tier-2 takes 20 seconds.
Shouldn't tier-2 use the autosummarydata blocks as to increase query performance? Are we asking the wrong table or doing something wrong? We have tested this with two customer setups, and it yields the same kind of results.