cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
1556
Views
2
Helpful
4
Replies

Performance comparisons btw Appliance and VM Image

blandrum
Cisco Employee
Cisco Employee

I have a customer who would like real world performance comparisons (not maximum number of concurrent endpoint support) between our appliances and VM image.  My assumption to them is that the hypervisor wouldn't perform as well as a bare metal customized Linux OS on our appliances.  They're willing to throw whatever VM resources are required to match our 3595 appliances, so if we have comparisons at that tier that would be excellent.  Thank you.

1 Accepted Solution

Accepted Solutions

thomas
Cisco Employee
Cisco Employee

Brad, The ISE Installation Guide is pretty clear in the Virtual Machine Requirements section:

"To achieve performance and scalability comparable to the Cisco ISE hardware appliance, the VMware virtual machine should be allocated system resources equivalent to the Cisco SNS 3415 and 3495 appliances."

To this end, we enforce minimum hardware requirements with the respective OVA files by naming them according to our hardware appliances : ISE-2.0.0.306-virtual-SNS3415.ova vs. ISE-2.0.0.306-virtual-SNS3495.ova


From a performance perspective, you should consider ISE appliances and VMs the same.

View solution in original post

4 Replies 4

Timothy Abbott
Cisco Employee
Cisco Employee

Hi Brad,

Sorry, but I don't think I completely understand the ask here.  Are you looking for comparisons that we have done or is the customer looking to test physical vs. VM themselves?  It also sounds like the customer is looking for specs on the new 3500 series appliances.  You can find more information on them here.  If customers decide to deploy virtual machines instead of physical appliances, we recommend that the VMs be modeled after the physical appliances in terms of compute resources.  That way we know the VMs will operate according to our performance and scale metrics.

Regards,

-Tim

They're looking for our internal testing to determine if the hypervisor being in place reduces the performance of the box if it's specd identical to our appliance

Sent from my iPhone

thomas
Cisco Employee
Cisco Employee

Brad, The ISE Installation Guide is pretty clear in the Virtual Machine Requirements section:

"To achieve performance and scalability comparable to the Cisco ISE hardware appliance, the VMware virtual machine should be allocated system resources equivalent to the Cisco SNS 3415 and 3495 appliances."

To this end, we enforce minimum hardware requirements with the respective OVA files by naming them according to our hardware appliances : ISE-2.0.0.306-virtual-SNS3415.ova vs. ISE-2.0.0.306-virtual-SNS3495.ova


From a performance perspective, you should consider ISE appliances and VMs the same.

I understand that matching VM specs to appliance hardware specs should achieve the same capacity in regards to number of concurrent connections.  My question was more around measured response times at load.  Have we tuned our drivers, software buffers, etc. well enough inside of the OVA to interact as quickly / smoothly with a VMWare hypervisor as we have with our custom built ADE-OS / Linux distribution that ships on the appliance?  It sounds like based on your answer above that we have. 

Thank you for the response.

Getting Started

Find answers to your questions by entering keywords or phrases in the Search bar above. New here? Use these resources to familiarize yourself with the community: