In a big data processing scenario, you are required to design a data ingestion pipeline that can handle high-velocity data streams. What considerations should you take into account to ensure the pipeline's efficiency and scalability?