6.5 vCenter Performance

[PhillipUdel]

I have a question on the maximums of 5.5 and 6.5

“I Have a friend” that has a  vCenter that has some fairly slow performance.

But it did make me give some thought to the published maximums
and they are effected by host latency.

So the vCenter Server Maximums states that you can have 2000
Hosts per vCenter Server.

The Question is simple, does host latency change that number,
and by how much?

It’s a rhetorical question, but I wanted to get a feel for
how much Latency would affect total hosts.

I plan to use this information to fight for more (PSC) as close
to the hosts as possible.

If I assume that I have 2000 hosts at 1 ms per.  What would be the count for hosts with 500ms?

I have used this rule of thumb, that for every 100 ms of latency
of a host that should count as a host, so a host with 500ms to the vcenter that’s
5 host.  Am I way off, to the right or to
the left?

And ya this “Friend” has lots of hosts on one vcenter..  “Not My Design”

[daphnissov · Accepted solution]

Host latencies do not affect the configuration maximums number. That said, the user experience and even some operations will definitely be affected for hosts that are over high-latency links which can manifest as significant delays, time-outs, and operation failures. The general object with ESXi hosts managed by vCenter is that they should be in the same or close geographic proximity to their vCenter. So, for example, if you have locations in New York and Paris, you don't want to manage hosts in Paris by a vCenter in New York and vice versa. In that design, you should be using two vCenters in each location and, optionally, joined in enhanced linked mode to manage them as one.