Times series data typically has a characteristic that only the recent data is accessed most frequently. The older the data it is, the less frequently it is needed. You can optimize this for cost saving by having a new table for each period. You set higher read/write capacity for tables with more recent data and decrease it when it gets older.