On Thu, 12 Sep 2013 09:13:39 -0500, Chip Grantham wrote:
>I've finally taken the time to try to understand the numbers behind the way
>EAVs were implemented. I found a great discussion in the redbook "z/OS v1.12
>Implimentation" SG24-7853-00 manual, chapter 20. Any time spend you happen
>to spend here is worth it. (not unlike all redbooks). Thanks to those that
>wrote it.
>
>I did happen into a segment that makes me go hmmm. 20.4.3 Multicylinder unit
>section says the 21-cylinder value for the MCU is derived from being the
>smallest unit that can map out the largest possible EAV and stay within the
>index architecture (with a block size of 8192), as follows:
>* It is also a value that divides evenly into the 1 GB storage segments of an
>IBM DS8000,
>* These 1 GB segments are the allocation unit in the IBM DS8000 and are
>equivalent to 1,113 cylinders.
>
>I'm sure the "index architecture" references the index vtoc architecture,
>which has always been a curious archeture to me. Has this design ever been
>made open? Just curious as to why it made 21 the magic number?
>
>I also ran into a math issue when I divided 21 into 1GB (or 1,073,741,824/21 =
>51,130,563.0476...). I suspect that's because the 1GB storage segment is a
>number used in the DS8000 degisn, and its really close to the 1GB value.
>Wondering if that's true or some other reason.
IIRC, when discussing disk storage, "the industry" uses the decimal meanings of
KB, MB, GB, etc. Thus, a 1GB disk allocation would be 1,000,000,000 bytes,
which divided by 21 yields 47,619,047.
-jc-
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
