Hello everyone,

I am developing a project in Brazil wich uses distributed computing as well and as I am in a very embrionary stage I have some doubts about the technology itself.
In my project I aim to do the reverse path that Boinc does wich is: provide processing power to end users from a server through the internet.
The question that I have is: what's the real advantage to use distributed computing?
Is it a real enhance to the processor (local cpu + remote processing) or you will just be replacing local processing for remote processing?
To a end user using 100% processing power from the server, would his local cpu be idle?
I don't know if at this point you guys understand but this project is for people with low incomes have access to computers, wich are expensive here in Brazil.

---

The BOINC model sends two items from the server to the remote client - the data to be processed and the application with which to do the processing. Upon finishing the processing the client sends the results back to the sever. The idea being that a single server can send data sets out to a very large number of clients who will run independently of each other. In your model only a few clients could be supported, and they would be very reliant on having a "good" internet connection - which is not always possible.

What you are describing is more like the cloud computing - a bit like Google & MS's offerings.

Hum... I see and well, this is frustrating.
I met this other project on https://github.com/miha-stopar/xCloud wich has a similar purpose to mine and I tought it would be possible/viable.

BOINC is intended to send out work to the millions of computers out there to crunch and send back the results. The ideology of BOINC is to let these computers out in the public serve in place of dedicated clusters of workstations and servers which are costly. However, not all work is suited for this kind of technology, particularly very time-sensitive work that needs results returned ASAP as further work is dependent on previous results.

What you are looking at is akin to thin clients attaching to the server (mainframes in olden days) which does the heavy lifting and very little processing is done on the user side.

You probably want to look into cloud and fog computing. If you're looking to experiment with such technologies with minimal investment (there's always an investment whether it's time or money), I suggest looking into Linux which can offer you a variety of server oriented distros. Check out Distrowatch: http://distrowatch.org for info and reviews.