IBM Message Board - Msg: 4346872

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Technology Stocks : IBM

IBM 312.75

+0.7%

Dec 10 3:59 PM EST

Public Reply Prvt Reply Mark as Last Read File

Previous 10 Next 10 Previous Next

To: J R KARY who wrote (3044)	5/6/1998 11:15:00 AM
From: Arrow Hd.	Read Replies (1) of 8218

Greiner, though a former IBM Mainframe Product Manager, is a bit off here. The Symphony ten engine cluster will produce about 900 MIPs with factoring in the multi-engine degradation which means when a job is run in single image over multiple engines the machine does not generate the total sum of the engine MIPs but due to the inefficiency of using multiple engines you get only some percentage of the MIPs in performance. That is why a large single engine size is preferred by some customers like the one quoted here. Symphony's engine size is more like 115 MIPs which is still almost doubling the 60 MIPs per engine of last year's Overture announcement. So a very significant improvement. Actually Hitachi's Skyline is not an advanced technology but older Bi-polar technology that due to its architecture produces via packaging and chip design a heavy duty engine which is currently at around 125 MIPs. They may have another turn still so getting to 150 MIPs may be possible but it is very costly to produce and IBM chose to switch to the less expensive CMOS implementation. Further, with multiple models you can sort chips for performance putting the higher MIP chips in the biggest "turbo" machines and utilizing what may have been a "sort failed" chip in smaller machines where the engine performance is not paramount. This avoids chip scrap. This is less an opportunity on the Amdahl and Hitachi lines. Hitachi has taken significant market share in the past couple of years from IBM (IBM 70+, Hitachi 20+, Amdahl 10-). This announcement will allow for reversing that trend and will assist in stemming price attrition due to the fact that IBM now has an equal offering so the need to give a better price to offset the fact that there was an inferior product no longer exists to the extent it did exist. This is only for single engine scenarios where the customer wants the machine with the largest engine. As Scott points out, what customers should be doing is implementing the Parallel Syplex where jobs are not only spread among engines within a machine but between machines. Still, the old paradigm in the mainframe business is engine size and this mostly closes that gap. There are many other exciting pieces to this announcement that we will see tomorrow including a feature that flags Yr. 2000 problems automatically on the fly as the machine runs the customer's application logging them in the system. Though the first customer ship isnt until the third quarter, the G3 and G4 are upgradeable to G5 via a "net priced" upgrade where the MCMs (MCM is the part that has the imbedded engines) are swapped. So IBM can sell the current product family in the second quarter with an upgrade to Symphony in the second half and the "two step" pricing can be made to be indifferent versus waiting for a full Symphony machine later in the year. This is not a new strategy but has been used since the 70s. This is one of the year's most significant announcements and any analyst worth his salt should be recalculating mainframe market-share erosion and price attrition.

Report TOU Violation

Share This Post

Public Reply Prvt Reply Mark as Last Read File

Previous 10 Next 10 Previous Next