As the world shelters at home during this pandemic, some people are called upon to use their skills to heal others, some use their knowledge and ingenuity to find solutions, and others stay home, helping to reduce the potential for spreading the virus. It might seem like some people can’t help, but that isn’t exactly true, especially for VMware partners who typically have computing resources available.
Researchers today use massive super computers to crunch numbers, simulate scenarios, and find answers to complex questions. These machines are specialized and VERY expensive. But today, there are millions upon millions of computers in our pockets, at our desks, and on our laps. A long time ago, some researchers started asking for people to lend their computers capabilities towards research. Essentially, a researcher send out a question, and hundred if not thousands of computers jump in to help solve it collectively.
There have been different initiatives throughout the years for this distributed super computer model. A common one today is called Folding@Home. This project helps by installing an application on a persons computer and utilizes their computers processor or graphics cards to run the calculations for the researcher. This distributed model makes it so much easier for researchers to get computational help without buying time on a massive super computer. It’s like letting your computer be it’s own super hero for saving the world.
Today, with the COVID-19 outbreak, the Folding@Home initiative is blowing up with hundreds of thousands of new machines helping to understand and find answers about the virus. Some very bright minds at VMware also made it easier to deploy the tool in a VMware based infrastructure, by building an easy to deploy virtual appliance! You can find out more about it on the OCTO blog here, or go straight to downloading the virtual appliance on the fling page here.
So what about partners? VMware has partners all over the world, partners running labs for their Professional Services teams, Partners running hosted services and building their own public clouds, partners that build labs for solving VMware solutions. How can these partners utilize their space server capacity for good? How can they donate their CPU and GPU resources for running more of these Folding appliances without the potential for VMware licensing costs? Simple, NFR licenses.
All VMware partners, from Cloud Providers, Solution Providers, to Professional services, have access to VMware licenses for use in a test environment. These licenses typically amount to enough servers that you can do two sites of four servers, each server with two sockets. This is through a combination of vSphere Enterprise Plus and vCloud Suite licenses (which include vSphere). There are also a couple of vCenter Server licenses as well.
Now, for Cloud Providers, I know you may not have available any additional hosts for deploying this on, but let me suggest a few things.
First, always make sure that you are dedicating physical hosts for this. Do not put the appliance on the same hosts as your production hosting environment. Usage Meter will still meter this appliance. It doesn’t use much resources in memory, but it will end up costing about 3.5 points a month per appliance at minimum (1GB RAM * 50% for no RAM reservation * 7pt bundle).
Second, if you need extra hosts, take a look at using DRS and other tech to increase your VM to Host density to a point you are comfortable with. Then you can free up some physical hardware. Remember, we are after CPU cycles, not RAM.
Third, if you are going to manage it from the same vCenter as production, use a separate cluster, and use an NFR license key. (For those using Usage Meter 3.6.1 or older) Once you have the host in another cluster with the NFR key, you can go to Usage Meter and tag the license key as DEMO. (For those using Usage Meter 4.1) deploy a separate vCenter or use standalone ESXi. UM 4.1 doesn’t have license categories yet in this rewrite of the appliance.
Make sure to grab the most recent version of the Fling, it will set specific flags on the VM that will make it stand out as a larger potential worker when you give the appliance over 16 cores.