The Tokyo regional analysis center is a computing center for the ATLAS experiment
at the Large Hadron Collider and one of the Worldwide LHC Computing Grid Tier 2 sites.
The batch system is managed by HTCondor.
For the future expansion of the center, R&D project using Google Cloud Platform (GCP), which is one of commercial cloud resources,
was launched.
The system is a hybrid system that worker nodes of HTCondor are deployed on Google Cloud Platform
while job manager nodes and file storages are placed on on-premises resources.
To manage worker nodes on GCP, GCP Condor Pool Manager (GCPM) has been developed.
The system with GCPM worked in the ATLAS production system.
In this paper, details of the R&D system are introduced and experiences on Google Cloud Platform are reported.