[VMC] HCX Bulk Migration transfer rate is slower than expected over Direct Connect (DX)

Products

• VMware HCX
• VMware Cloud on AWS

Issue/Introduction

• A Direct Connect (DX) connection has been configured in the VMC on AWS environment. 
• HCX Bulk Migrations using this connection are taking much longer than expected. 

Note: While this article is focused on VMC on AWS environments, this information can be used for other public cloud HCX deployments. 

Environment

VMware HCX
VMware Cloud on AWS

Cause

While DX connections should be quicker and more reliable compared to an Internet connection, there are a few possible causes for slower transfer rates:

• Packet fragmentation caused by an MTU mismatch
• L4-7 Firewalls doing packet inspection or analysis
• HCX best practices have not been followed

Resolution

• Download the source HCX logs and review the "app.log" file for the migration ID and the transfer speed entries for the Bulk Migration. 
• Check the Service Mesh being used for the Bulk Migration and verify communication between source and destination sites is utilizing the configured DX. 
• Very the HCX Bulk Migration best practices are being followed. 
• Ensure the MTU settings are the same in the source and destination sites. 
• Review any configured firewalls to ensure no packet inspection or analysis is occurring as this could cause slower transfer rates. 
• As a test, modify the Service Mesh to use the "External Network/Internet Network" profile and test the Bulk Migration to monitor the transfer speed. 

Additional Information

• Understanding VMware HCX Bulk Migration
• HCX Bulk Migration operations and best practices
• HCX is not migrating or passing traffic over network extensions at expected speeds