- The Spark integration involves two key components: the read path and the write path. - The priority is correct functionality, not peak performance. - Support will first be introduced in Spark 3.5.