Originally posted by: Ronin
Ease of setup and maintenance would also be helpful, as deploying a new DC project to 70 or so systems manually can be time consuming, and I'd rather minimize my headache.
You can
customize the installation of BOINC, and use active directory to deploy to how many computers you wants.
If having installed the BOINC-client as a service and you later wants to add another project, just "attach" on one computer, copy the resulting account_url.to.new.projects.boinc.page.xml, example account_setiathome.berkeley.edu.xml, to all the other computers and remotely re-start the BOINC-service. If you wants to remove a project, just delete the account_*-file and re-start.
Preferences and so on are web-based, so it's enough to change one place. This also makes it very easy to change shares between projects if multi-project.
The computer-overview on web-pages is sorted after last time a computer contacted, so is very easy to spot any computers having problems. If you looks closer into results will also see if anyone have problem with validation and so on.
To also check what individual computers are currently doing, you can use 3rd-party tools like BoincView to monitor/control all computers in a lan at once.
BOINC have built-in caching, multi-cpu-support, service-install, multi-project-support and cross-project-stats, with currently 5 released projects, CPDN, Einstein@home, LHC@home, Predictor@home and SETI@Home. Many other projects is currently in
alpha/beta-testing and even some non-listed, so there's upcoming projects from various parts like "space-science" with example Orbit@home and PlanetQuest, life-science with Folding@home, Rosetta@home and The Lattice Project, and even non-science with BURP.
The current team-rank is 14 in Einstein@home, 19 in CPDN and SETI@Home, 20 in Predictor@home and 44 in LHC@home, so all of the released projects needs a boost. With all the upcoming BOINC-projects, there'll most likely be ample opportunity to help-out.
It's recommended to have atleast one "backup-project", this is especially true for LHC@Home since work comes in batches and between batches it can be days even weeks with no work, but also for other projects it's recommended so any long outage will not lead to idle cpu. Setting a project as "backup", just set resource-share for this project to 1, and other project(s) to example 1000.