David Brown wrote: > On Sun, Mar 23, 2008 at 02:59:28PM -0700, Tracy R Reed wrote: > >> I really like this idea. Memory is plentiful and cheap. Good hardware >> is quite reliable these days and if you put a UPS behind it I could >> conceive of eventually putting a production database or some other >> seek-heavy application on such a storage system. > > I'm not seeing how this isn't just a poorer way of implementing caching. > There's nothing novel here, except that it requires full memory for the > size of the disk, instead of using it dynamically. > > There are also a lot of very important ordering issues with cachine and > disk that likely mean that actually loss with this system would be > completely catastrophic. > > Almost makes me want to re-subscribe to LKML just to see the idea thrashed. >
While there may be plenty of gotchas to worry about, and a long time between now and a reliable implementation, I think you may be a little too quick to dismiss it. In (perhaps only?) my imagination, it may in fact, be so much better at (say) caching functionality, or versatile at dealing with (or redefining) ordering, that new paradigms may pop out of it -- who can forsee? Regards, ..jim -- [email protected] http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-list
