v2 buckets will still be the default timeseries bucket, and buckets will only be promoted to v3 if an out of order measurement is inserted. We can take advantage of this sortedness guarantee in the bucket unpacker class. Currently, v2 and v3 buckets use the same BucketUnpacker version. We should first create a base BucketUnpackerCompressed class that takes a bool isSorted as a template parameter and instantiate the two concrete version (v2, v3) unpackers from it. If needed, this part can be a different ticket but we should decide how we are going to take advantage of the sortedness in the unpacker.