On Tue, 20 Jun 2006 10:33:33 EDT, IBM Mainframe Discussion List <[EMAIL PROTECTED]> wrote:
> > >In a message dated 6/20/2006 9:12:47 A.M. Central Daylight Time, >[EMAIL PROTECTED] writes: > >>I thought that todays disks have more capacity on the outer >>tracks than on the inner ones. Transfer rates vary with the >>track position being read or written to. > >Transfer rates from the disk into the read/write transducer do indeed vary >depending on where the track is, but then the data is slowed down or speeded >up as needed by the controller's microcircuits and by buffering so that data >transfer to or from the channel is at a consistent rate. > > >Today's disks are not different from yesterday's disks in this regard. Yes, they are. The old SLEDs were built with large diameter disks and only a narrow band contained all the tracks. Every track had the same capacity, and the transfer rate remained constant. >The >circumference of the outermost concentric circle is always larger than that >of the innermost concentric circle. In the extreme case, the circumference of >any circle is pi*diameter, and the circumference at its center is 0. Data >is recorded on tracks at X number of bytes per linear inch. No two tracks >have exactly the same number of linear inches. The engineers who design the >disks, read/write transducers, and control units take this into consideration. On a 3390, the inner tracks and outer tracks were only about two inches apart. The electronics transferred the data to all tracks at the same rate, with the result that the data on the outer tracks was recorded at a slightly lower bit density. >The control unit will transfer data to or from the channel at a rate no >higher than the channel's maximum transfer rate regardless of how many bytes are >on the track or how fast the disk is spinning. Controllers these days also Disk drives have device level buffers, and control units (not the same thing as controllers) have additional buffers. >have device-level buffers, so even if the disk is spinning too slowly the data >transfer rate from the buffer to the channel can still take place at the >channel's maximum rate. > ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
