Harmful, pervasive SQL Server performance myths I start a new series around disproving prevalent myths regarding SQL Server performance.