We are deploying Pega 8.x platform on AWS private cloud behind load balancers.  With Kafka streaming data, we would like to utilize data persistence using available AWS storage technologies like EFS (works with across subnets and regions with backup mechanism) to manage EC2 instances.

I was successfully implemented EFS mount on EC2 instances are the Kafka data is being written to the EFS structure.  But how ever the Streaming documentation Ref: https://community.pega.com/knowledgebase/articles/decision-management/84/configuring-stream-service, the document does mention that do not use "Important: Do not use network attached storage or shared folders to store your stream data."

EBS is not a good solution where the storage doesn't go across subnets and regions.

Is EFS correct solution and what is the downside of using EFS?  if it is not a right solution, can someone share their experience? 

Thank you