Hello, @randyramandani  This indeed can be done, as long as both switch are directly connected to each other and can learn MAC address for each other there should not be much latency. 
you can monitor latency between CVM in file. 

/home/nutanix/data/logs/sysstats/ping_hosts.INFO, as long as its below 5ms it should be good.  

Hy @Sudhir9 thanks for your best answer, but in real production environment, is there any experience with the topology as above (using 2 different switches in one cluster)

@randyramandani Yes, I have seen this being used in production environments as well. 

Please refer to https://portal.nutanix.com/page/documents/solutions/details?targetId=BP-2071-AHV-Networking:BP-2071-AHV-Networking

When multiple uplinks from the AHV host connect to multiple switches, ensure that adequate bandwidth exists between these switches to support Nutanix CVM replication traffic between nodes

Regards

Sudhir

Hi @Sudhir9 great, thanks for your answer

Hi @randyramandani In the topology what you shared in that here is no high availability, many singly points of failure, and potential oversubscription between the two switches. Below is the topology you can have a look

@Neel Kotak Hi Neel, thanks for your response

yes I understand that, but now currently my customers have limited infrastructure and will be improve in the next year