Just how long does it take me to provision a VM?  I wonder how many hours I've had my hosts in maintenance mode?  How long does it take to failover with HA?  If you have ever had any of these questions then you might want to head over to VMware Labs and download their latest fling, vBenchmark.

vBenchmark is a great tool that can be used to quantify and really sell the benefits of virtualization to both yourself and to your managers.  Broken down into four categories (Infrastructure Configuration, Efficiency, Agility, and QOS) vBenchmark parses through your vCenter tasks and event logs, rolls up all those numbers and metrics into very high-level and useful dashboards. Most of the coolness of this application however doesn't come from the fancy dashboards, it comes from the ability to share the data and see just how you rank up against similar type deployments in your geographical region, industry, and business size.  I'd highly recommend checking that out as you may be surprised.  

Installation is probably the easiest thing I have ever done.  Go to VMware Labs and download the ovf and proceed to import it as you would any other ova/ovf (File->Deploy OVF Template).  From here simply follow the deploy wizard answering the questions.  Honestly I let DHCP handle the networking for my install and it took me about 5 minutes from when I imported it to when I was firing up a browser to do the initial configuration.

Once the appliance is installed and powered on simply open up a browser and point to http://IP_OF_APPLIANCE:8080/. First off you need to specify the vCenter Server (or servers) that you want to connect to, provide some login credentials for it and click 'Add'.  Then, select the number of months of data you would like to retrieve (Anywhere between 1 and 6 months).  Once ready, click 'Initiate Query and Proceed to Dashboard'.  The appliance should now go out to your specified vCenter and start grabbing all the task and event information that it has.  Obviously, depending on how far back you requested to collect the data, the process could take a bit longer.

The next screen that you see allows you exclude certain hosts or clusters from the collection.  It recommends on the screen to definitely exclude and hosts or clusters that are running virtual desktops.  Basically, you are now setup!  That's really all there is to it!  Once the system is done collecting data you will be presented with the big selling point of this application; The dashboards.  These dashboards are broken down into the following four categories, each displaying a bit of information on the dashboard tab, with some having a little bit more detailed breakdown in their own tabs.

Infrastructure Configuration

First we see Infrastructure configuration.  This gives you a some nice calculations around the average number of physical CPUs and average amount of memory per host as well as taking a look on the virtual side of things and giving you the average number of vCPUs and average configured memory and storage per VM.  As with every single item on this page you can click to share your results and in return you will be able to see how you stack up in all of the categories to others in the same industry, location, etc…


Next up is Efficiency.  Efficiency is further broke down into obviously VM density and Memory efficiency as well as a little bit of operational efficiency.  You can see that this page shows you things like your VM Density or your average number of VMs per physical CPU (NOT CORE) as well as your ratio of total vRAM configured to the total available physical RAM in the hosts.  The admin productivity section allows you to break down operational efficiency by calculating the total number of VMs per VI-Administrator.


Agility takes yet another different look at your environment and determines how long it takes to make changes or spin up new VMs into your environment.  First off you can see the average in minutes that it takes to provision a VM.  From what I can find this metric includes the average duration in time that it takes to configure a host, add it to a cluster or vCenter, clone or deploy the VM and power it on.  The other metric displayed here Reconfiguration is basically the average duration of tasks revolving around reconfiguring a VM, Host, Network, or Storage.

Quality Of Service

Last, but certainly not least is QOS.  Here we can see some time savings of different metrics revolving around maintenance mode, vMotion, Fault Tolerance, etc.  Firstly, the Hours of Host Server Maintenance Downtime is exactly what it states, an average of the time that a host was in maintenance mode.  Somewhat related to that is the Hours of Application Downtime Avoided which is basically the average monthly hours of downtime that are avoided by utilizing vMotion during host maintenance events.  The unplanned downtime reduction (you can see I have no data here) will essentially display the average duration of failover events resulting by HA or FT.  And finally, the RTO section is the average duration of DR failovers or test failovers.

So there you have it, kind of a small intro to the latest fling.  I can certainly see value in this application, especially for those high level people who would like to see some hard numbers around the benefits of the platforms they are paying for.  Also, for that VI Admin, it's a great tool to see how you stack up against your peers, so I would encourage everyone to share their data.  You can actually see what data your will be transmitting before doing so using preview functionality and honestly, the only identifier being sent back to VMware is a license key, and really, I can't see VMware using this for any auditing purposes anyways…:)