Morning folks!
Written an XTDB technical deep-dive for you - “Taking the ² out of O(n²)”
It covers what “bitemporal resolution” is, how it works in XT2, and how we identified and fixed an error causing us to be an order of magnitude out on performance when working with time-series data.
If you’ve got any thoughts on this blog, feel free to discuss them below 
James
Does 1.x have similar issue?
Hey @blshao84 
It doesn’t, but only because it only supports point-in-time queries - this is an issue affecting general bitemporal queries.
James
To add a little more colour, 1.x’s entity-history API (without corrections) is outwardly similar to what’s discussed here, but in 1.x most of the temporal resolution work happens at write time whereas this algorithm for 2.x necessarily has to do more work at query time, because unlike in 1.x the index/data isn’t fully sorted.