Are there any guidelines or rules of thumb for latency targets for VMRC connections with vCloud Director? I have checked the vCAT architecture guide the key statement I can see is the following – but what does it mean “issues with network latency”?
vCloud Director is neither designed nor supported for multi-site deployments in the current version of the
product due to potential issues with network latency and reliability. In an environment with multiple sites,
design each site as separate vCloud instances with considerations for future interconnectivity instead of
having a single vCloud that spans the sites.
From the available documentation the only minimum specifications we can find are for the vcloud director to have 2 vNIC on a management cluster of ESXi hosts and in turn the ESXi hosts must have a minimum of 1Gb pNICs and latency of less than 20ms for the cloud infrastructure (ESXi, vCenter, vShield, vCD)
http://www.vmware.com/files/pdf/vcat/Architecting-VMware-vCloud.pdf
http://www.vmware.com/files/pdf/VMware-vCloud-Private-Enterprise-Cloud.pdf
We believed up to 200ms was an acceptable level from live testing. We have users with 300ms/400ms connections which are deemed to be unusable for our current solution. Of course this is very subjective as user perception is the measurement and obviously if there are internal software issues or resource constraints in the cloud org the system can also be unresponsive - but if we assume there are no other performance issues i.e. a local user has no problem using the same vApp
It would be good to have a target table as we may need to deploy another vCloud environment in a location with lower latency and we would like to issue our users guidance for remote connections from home or customer sites similarly if there is a bandwidth element i.e. minimum recommended connection speed 1Mbs : For example:
Latency RTT | VMRC Response |
---|---|
100ms < | Excellent User Experience - Console Very Responsive |
200ms < | Good User Experience - Console Responsive |
250ms < | Average User Experience - Console Response delay is perceptible |
300ms < | Poor User Experience - Console Response delay is obvious |
350ms + | No User Experience - Console is unusable |
The only end user restriction that seems to be specified is the browser used to connect - http://pubs.vmware.com/vsphere-55/index.jsp#com.vmware.vmrc.pg.doc/vmrc_pg_overview.3.3.html
There is a very detailed article on troubleshooting VMRC connections but this does not mention latency or bandwidth restrictions which would be very useful.
Chris Colotti has this post that is very good but again it seems to focus on resolving no connectivity not poor connectivity!
http://www.chriscolotti.us/vmware/how-to-troubleshoot-vcloud-vmrc-connections/
There is a statement about PCoIP network requirements that indicates latency up to 100ms – see below. A similar statement for VMRC would be ideal!
PCoIP can compensate for an increase in latency or a reduction in bandwidth, to ensure that end users can remain productive regardless of network conditions. PCoIP is optimized for delivery of images, audio, and video content for a wide range of users on the LAN or across the WAN. PCoIP provides the following features:
You can use 3D applications such as Windows Aero themes or Google Earth, with a screen resolution of up to 1920 x 1200. With this non-hardware-accelerated graphics feature, you can run DirectX 9 and OpenGL 2.1 applications without a physical graphics processing unit (GPU). This feature provides acceptable performance on the WAN, up to 100ms.
So far from testing the table starts to look like:
Latency RTT | VMRC Response |
---|---|
100ms < | Excellent User Experience - Console Very Responsive |
200ms < | Good User Experience - Console Responsive |
300ms < | Average User Experience - Console Response delay is perceptible |
400ms < | Poor User Experience - Console Response delay is obvious |
400ms+ | No User Experience - Console is unusable |