On Mon, 16 Feb 2004, Orlando Andico wrote: > On Mon, 16 Feb 2004, ian sison (mailing list) wrote: > .. > > per controller. The object of the exercise is to make sure that the bus > > doesn't become a bottleneck assuming a drive can churn out 33-40MBps > > from platter to the underlying drive electronics. > > but when doing random read/write, the drive's performance does not > approach the burst speed, but the I/O's per second. A drive doing 100 > I/O's per second is a fast drive (10ms seek time). > > but since UNIX generally uses 4K file blocks, and file I/O is most often > just one block at a time (due to mmap) then your actual drive bandwidth > for a very busy drive is more like 4K x 100 I/O/s per second = > 400K/second. Which is nowhere near 40Mbps. > > translation.. with SCSI you can put at least 10 drives on a single > controller, assuming the application has lots of small files and lots of > random seek (true of Squid and email servers).
True, but that discounts optimization done on the OS level as well the firmware level which can induce more data to be sent to the bus than the over all average 400K/second you mentioned. Examples of these are optimizations resulting from the Elevator scheduler algorithm used which could result in multiple read/seek/reads being optimized out to one large read to satisfy certain disk requests. Another of course would be hard disk on board cache hits, which when enabled could quickly fill up the bus with data. -- Philippine Linux Users' Group (PLUG) Mailing List [EMAIL PROTECTED] (#PLUG @ irc.free.net.ph) Official Website: http://plug.linux.org.ph Searchable Archives: http://marc.free.net.ph . To leave, go to http://lists.q-linux.com/mailman/listinfo/plug . Are you a Linux newbie? To join the newbie list, go to http://lists.q-linux.com/mailman/listinfo/ph-linux-newbie
