The paper presents BOINC, a software similar to SETI@HOME for using home users computational resources.
A common problem with this approach are the incentives for home users, which the current paper addresses through the use of "credits" for computational power (which may be rewarded) and by allowing users to select computational topics they feel stronger about.
An interesting discussion point is the use of total energy of this approach compared to a dedicated cluster (a single organizational control). For example, BOINC launches multiple redundant tasks to avoid malicious results, which is clearly a waste of energy. Also, these tasks can wake up processors in low states of sleep (a global optimization could optimize this energy use) and transfer data on long distances.
Subscribe to:
Post Comments (Atom)
I'm not sure when that paper was published, but some of the points you have stated are beginning to be addressed.
ReplyDeleteI believe the current BOINC behavior has changed from the multiple redundant tasks all the time to the use of redundant tasks until the trust a computer's resources are established (not sending back corrupted results, modified results, errors, etc). A project then is able to send one task for a work unit and periodically the computer is verified to maintain its trust level. Helping to better utilize the current resources, reducing inefficiency, and maintaining trusted results.
GridRepublic a nonprofit working in collaboration with BOINC is working to create ease of use of volunteer computing with BOINC through a simple to use website for joining, discovery, and management of projects. It also makes it easier to manage multiple computers from a single login. They are working to spread the word, creating an interest in volunteer computing, and most importantly trying to keep people involved. Check them out at: http://www.gridrepublic.org
I and probably many others would like to find out the energy efficiency between normal computer use and clusters.