1. Email Backlog
2. Cheryl Watson’s TUNING Letter 2000, No. 2 Summary
1. Email Backlog
I have been waging a losing battle against my email backlog. If you have a quick question, I normally answer the same day. Anything else may take weeks, although I have some pending from 18 months ago. So I’ve decided to delete the old ones and start fresh. If you’ve sent me a question that I haven’t answered and you still want a reply, please re-send it (but only after April 30th). I’ll make every effort to keep up this time. This is my first vacation in a year and a half, so I want to leave without these hanging over my head. I hope you’ll understand. Tom and I leave on April 13th for our April 15th wedding, followed by our honeymoon in Taiwan for two weeks.
2. Cheryl Watson’s TUNING Letter 2000 No. 2 Summary
The 36-page 2000, No. 2 TUNING Letter issue was just completed. Electronic subscribers should receive their TUNING Letter today or tomorrow. Printed issues should be mailed before April 21st. A CPU Chart update containing the re-rated Amdahl processors will also be included.
Amdahl’s New CMOS
Perhaps the most important management issue in this newsletter is Amdahl’s announcement of the improved performance of their Millennium 2000C and 2000E CMOS servers (page 31). They also announced a 232-MIPS uni-processor with a 2508 MIPS 16-way scheduled for the fourth quarter of 2000. This is especially welcome in light of the HDS announcement last month of their temporary marketing halt on their S/390 line of processors. So, it DOES appear that IBM will continue to have competition in the mainframe market and the price per MIPS will continue to drop. (Now if only the software costs would come into line!)
Technical Issue
The majority of this issue contains only very technical items, but technically-oriented managers may appreciate the following items: the reason for MP overhead (page 34), a problem that occurs when response goals are set incorrectly in goal mode (page 35), an overview of UNIX terminology for MVSers (page 23), a visual representation of why you should avoid I/O (page 7), and some figures to support our contention that COBOL indexes are better than subscripts by 200 to 1 (page 6). This issue also confirms the warning that IBM 3350 disk drives, BTAM, and TCAM will no longer be supported (pages 7 and 13). Two Web tutorials on CICS are also listed on page 9.
Elsewhere
The majority of this issue covers APARs and recommendations presented at the last SHARE conference (page 10), a continuation of our HFS series that concentrates on avoiding I/O (page 14), and a continuation of the BPXPRMxx (OS/390 UNIX) parmlib member (page 26).
Our S/390 News starting on page 4 includes an RMF update for the partition data report, a new release of the RMF Performance Monitoring of OS/390, a new online parallel sysplex monitor from Amdahl (with a free trial), updated coupling facility SYNC time estimates, a performance tip for Easytrieve, APARs for IEHLIST and RACF, and two interesting user Web sites (one devoted to DB2 and the other to mainframes). If you’re looking for a list of colleges and universities that provide mainframe courses, see page 9.
Our 1991-1999 CD-ROM was mailed to our electronic subscribers on March 30. The new CD-ROM contains almost 800 additional pages. (Page 31)
Our Q&A on page 33 includes a warning about scheduling environments, clarification on LLA, and a description of page data set allocation.
Stay tuned!