Hi there @Duhit Choudhary unfortunately you'll need to work closer with Ab Initio on this, as their documentation is only available for paying customers.
Typically Ab Initio integration is used to manipulate files on HDFS, and it can write out files into Hive tables, there are three main methods of integrating Ab Initio.
The first one is just to keep it as a completely separate cluster, and passing files to and from HDP.
The second one is to more tightly couple the Ab Initio instances by installing them on edge nodes, this means they have more direct access to HDFS and are closer to being part of the cluster, HDP client tools and libraries are installed on the edge nodes for easier direct access, however scaling this can be difficult depending on how you are deploying Ab Initio.
The third method is to run Ab Initio on the HDP cluster itself. Ab Initio does have some support for YARN integration, however as yet it is not fully YARN certified so your mileage may vary on your experiences.
My utmost recommendation is to speak to Ab Initio, as they should be able to point you to the integration documentation that is not in the public domain.
Good luck!
