Thanks for the response. I'm still a little confused by the 21 thing. By my calculations 21*47,619,047 yields 999,999,987.
The 8192 block size looks like a nice bit map for 65536; 65536/8 is 8192, so the first area fits nice into a bitmap of 8K. Also when the theoretical limit of 255tb (268,434,453 cyl) and divid it by 8192, the result is 7FFF + a little, or a nice half word. The current limit of 262,668 cyl on a bit map bases would fit nicely into 3+, 8192 blocks plus the original bit map for the first 65520 tracks or four total; (262,668-65520)/(8*8192) = 3.008+. All of this with the assumption that allocation is in 21 bit chunks. OK, my brain hurts. The CA sizes makes a lot of sense to me. 315 (21*15) does yield some nice number for CAs; (1,3,5,7,9,15). IMHO, a CA larger than 15 tracks doesn't make a lot of sense and that really opens up the use of the 000 of the CCW (CCCC000H) for EVA cylinder values. Hmmm. In a world of powers of two, this is a interesting venture. Chip Grantham Sr. Software Engineer Syncsort Incorporated P: 201-882-8337 | C: TBD | F: 201-573-5176 E: [email protected] www.syncsort.com INTEGRATING BIG DATA… SMARTER -----Original Message----- From: IBM Mainframe Discussion List [mailto:[email protected]] On Behalf Of John Chase Sent: Monday, September 16, 2013 7:58 AM To: [email protected] Subject: Re: Makes you go hmmmm, EVA MSU of 21 Cyls 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 ATTENTION: ----- The information contained in this message (including any files transmitted with this message) may contain proprietary, trade secret or other confidential and/or legally privileged information. Any pricing information contained in this message or in any files transmitted with this message is always confidential and cannot be shared with any third parties without prior written approval from Syncsort. This message is intended to be read only by the individual or entity to whom it is addressed or by their designee. If the reader of this message is not the intended recipient, you are on notice that any use, disclosure, copying or distribution of this message, in any form, is strictly prohibited. If you have received this message in error, please immediately notify the sender and/or Syncsort and destroy all copies of this message in your possession, custody or control. ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
