Folding Forum

I have contacted the IT department at our school district, and they are interested in exploring the possibility of running Folding@home. Their primary concern before running a trial program at a small number of computers is security with data transfers. I am not very knowledgeable on this subject. The director of IT infrastructure here requested information "to characterize the interaction between the protein folding program and the host computer as well as what is included in the outbound data stream from the host computer (data content, data transfer file sizes, etc)".

Ideas on how to find out this information? I assume that there are certain ports that may be blocked at school that Folding would need for data transfer. I also cannot seem to find contact information on the Folding website (like an email).

Welcome to the F@H Forum TemurAmir,

This is a good overview that you can present to your director of IT Infrastructure (http://www.stanford.edu/group/pandegrou ... ingFAQ.pdf).

Most support for Folding@home is here. The researchers are busy doing their work so support is best taken care of on this forum.

In Folding@home, the client software first downloads a workunit (a packet of work), processes it to completion, uploads it, and then repeats. Workunits can vary in size, but I would expect each to be one or two megabytes. The upload/download occurs rarely because most of the time is spent doing the actual protein folding calculation. I haven't measured it myself, but I would probably put the transfer sizes at perhaps five or 10 megabytes per day per computer, which is not very much. The Stanford servers then organize the returned results and prepare new ones for other people. Communication only occurs over ports 80 and 8080, which should be allowed by your firewall by default since those are standard HTTP ports for web traffic. The client software only communicates with F@h servers, which are in one of several academic locations, including at Stanford. The software also compresses the data before sending it, minimizing network requirements. Digital signatures are used to verify the authenticity of workunits downloaded from Stanford.

Thus, in a nutshell, the clients will download workunits and uploaded completed workunits back. The software will also download the latest calculation cores, which do the actual calculations on the workunits. Cores are downloaded quite rarely, so usually that only happens at startup.

One thing to keep in mind is what resources F@h will use and how and when it will use them. You may want to take a look at http://folding.stanford.edu/home/faq/fa ... ion/#ntoc3

Please let us know if you have any more questions. School computers should be very helpful to F@h. Thanks for pursuing that!

In a school setting, I would choose to fold on idle, that should do at least 16 hours a weekday, perhaps 23.

Thanks for the suggestions everyone. I will link them to the article panther provided as well as the information in this thread.
Our school computers are on 24/7 so it seemed logical to have them do work during idle time, which is the setting it would run on. The next hurdle to overcome after security is the electrical cost, however.

Electrical costs are a potential issue that must be carefully considered. For FAH to actually perform scientific calculations, we recommend that the power-savings features in each computer be disabled -- the monitor should sleep but the processor should continue to perform calculations. In fact, the setting to only do work during idle time may have the unexpected effect of preventing calculations if the idle state also puts the processor to sleep.

If you are in a part of the world where classroom heating is more important than air conditioning, giving computers something to do during off-hours may be an auxiliary benefit of producing heat.