A new VMware ESX Server upgrade at the horizon?

I’ll really appreciate it if you guys answer the foll. questions for me. If
you don’t want to broadcast the answers, you can send them to svaghani at
vmware dotcom. Also, you can choose to answer a subset of the questions.
Excuse the technicalities, but they matter to me 🙂

– Do you use VMFS-1 volumes on your ESX2.x.x production servers? If yes, why?

– If you have > 1 ESX server:
* can the VMs on one ESX server ping VMs on the rest of the ESX servers?
Please note: this is different from asking whether the Service Console on
one of your ESX servers can ping the Service Console IPs on other ESX servers.
* Are all your VMs on the same subnet or inside the same AS? If yes, what
kind of network do you use (10G/1G/100M)? Also, if inside an AS, do you
cross a router when sending packets between such remote VMs?
* Do you use a non-conventional MTU? Either way, whats your MTU size in the
datacenter?
* What is the worst case ping time from VM1 on ESX1 to VM2 on ESX2?
* Do you have any empirical evidence on what is the median packet size on
your network? You might need to look at switch management utilities, etc.
* Do you use NIC teaming on ESX?
* How do you cope against network failures? (You can just give me a buzz
word, I’ll understand)
* Do you bandlimit traffic? Do you have level 4 bandlimiters? level 2, 3?
If you do, what kind of policy do you use?

– For people who use SANs with ESX: (Please provide impulsive answers –
don’t think for more than a second)
* In your opinion, what has failed more often: your SAN or your network?
What component, in particular?
* If and when your clients complain about performance problems using VMs or
physical machines, which one is the more frequent culprit: your SAN or your
network? What component, in particular?