does boinc use my computer for something besides what's advertised? are projects spread via p2p and I'm now a server for boinc? can I disable that? I'm using a cell phone for internet, it really is using 100% of my upload and download I have to close boinc for anything else to work


I'd like to avoid this and leave it on run after 1 minute of computer not in use

No, BOINC doesn't use public, external, bandwidth - except when you see transfers in process.

How are you measuring the traffic? If you're measuring it at the fone connection, then you have some other problem - I'd check for malware.

But if you're using internal monitoring on the host computer, be aware that the Client and the Manager are in constant internal communication over TCP/IP 'loopback' address 127.0.0.1, port 31416. That traffic is innocuous - it doesn't use any external (paid-for) bandwidth.

I've found some Projects I believe BOINC@TACC acts as a P2P every time I see it running I see my bandwidth is used constantly. Never shows up in BOINC Transfer.
From Task Manager.

I've found some Projects I believe BOINC@TACC acts as a P2P every time I see it running I see my bandwidth is used constantly. Never shows up in BOINC Transfer.
From Task Manager.

Does any of this usage show up in the Event log?
(cpdn sends some data as an RPC, but it's not often, and only a few hundred bytes.)

Does any of this usage show up in the Event log?

I would try enabling http debug and http transfer debug to look at this. - Disable them again afterwards as they generate a lot of data which makes it time consuming to look for other things in the event log.

If you are 100% sure about this, I would suggest posting on the project's forums.

There's only one other possibility I can think of: is either of our complainants running tasks which require a Virtual Machine?

If the VM is communicating directly with its project base (as I know LHC does), this would show up in a network traffic scan but not in BOINC.

Could either of you supply any precise detail about the projects, tasks and applications involved, and the measuring tools used to observe the bandwidth usage?

There's only one other possibility I can think of: is either of our complainants running tasks which require a Virtual Machine?

If the VM is communicating directly with its project base (as I know LHC does), this would show up in a network traffic scan but not in BOINC.

Could either of you supply any precise detail about the projects, tasks and applications involved, and the measuring tools used to observe the bandwidth usage?

I've personally seen where a data throttle is not respected by the VM for LHC.

The other issue is VM's sometimes don't stop crunch when BOINC asks them to. No checkpoint is one reason. BOINC unfortunately assumes that the VM did stop so it will start another job. A few of these and suddenly all your bandwidth is being used, never mind more cores than your chip and perhaps more RAM than your system has.

My advice on VM projects if you use the computer for anything but crunch is:
<max_concurrent>1</max_concurrent>
and/or
<project_max_concurrent>1</project_max_concurrent>

You can try higher numbers if you monitor with O/S tools to keep your system sane.