Your company and one of its partners each have a Google Cloud project in separate organizations, and you need to establish internal IP communication between these projects. Your company's project (prj-a) runs in Virtual Private Cloud (VPC) vpc-a, while the partner's project (prj-b) runs in VPC vpc-b. There are two instances running in vpc-a and one instance running in vpc-b. The subnets defined in both VPCs do not overlap. Considering the requirements of minimizing latency and maximizing throughput for internal communication between instances, what should you do?