• I've rolled out a refactor of our wind tunnel metrics database system, which was partitioned tables with row compression and is now partitioned views across separate databases with page compression. This makes it considerably easier to move data from live to archive (was scary sliding partitions!) and reduces the backup considerably (as the data is essentially a heap, no need to daily backup the old databases every day). The data is created at a remote site on the WAN and is replicated back to HQ live in seconds, billions of rows of data in a year. Also got a performance boost as part of the deal - which was an unexpected bonus. The system is used to design the car for an F1 team.