Currently this witsml client manages last query time and always query from last to now time. This is not flexible enough if applications like to manage the offset themselves. Plus, if there is a large gap in log data fetching, we might want to split the large queries into parallel smaller ones for performance consideration.