• pipelining mode using Spark data processing
• pipelining mode using the Hadoop
• plugin manager
• Better interface with databases or system