<<< HUMANE::DISK$SCSI:[NOTES$LIBRARY]DIGITAL.NOTE;1 >>> -< The Digital way of working >- ================================================================================ Note 2220.113 Poor Alpha Marketing 113 of 114 OTOU01::MAIN "Systems Integration-Canada,621-5078" 81 lines 27-APR-1997 23:49 -< It's all smoke and mirrors !! >- -------------------------------------------------------------------------------- Boy, it is a real treat to see real marketing kicking in and going on all cylinders.. previous article assumes the fight is over, and it is only just begun ! At a time when the Intel architecture is peaking out and Customers are faced with massive server recentralization, budget and headcount cutbacks, thin client with FAT servers looming, Intel still has the clout to convince the trade rag's that deploying more Intel 32 bit based servers are the way of the future. Now thats marketing ... Customer IS dept's are being told that whatever they deploy better work for 18-24 months. Because of this and the small price difference, these Cust's are deploying P166-PPRO 200 desktops with 24-32MB of memory. Am I crazy, or does deploying servers that are roughly the same speed as the current clients being deployed sound a little flawed ? Lets look underneath the fluff .. - Merced (P7) has been delayed until 1999 at the earliest. Intel is only committing to release it before the year 2000 (press release). This does not include the time to get NT ported to 64bit, or any app's such as SQL Server to be ported from Alpha. - Merced will run existing app's in "compatibility mode", but they quote performance numbers based on IA-64 based app's. Ok, so even with HW and brilliant porting folks, we are looking at the year 2001+ for native IA-64 app's (assuming no slips). Competition will be 900-1Ghz Alpha's based on EV6-7 architectures, so I would say the P7 folks have their work cut out for them .. - Klamath (Pentium II) was delayed due to performance being worse (in certain area's) than existing PPro's ..now not expected to ship in qty until fall '97. - rather than announce that the follow on to Klamath (Deschutes) has been delayed a second time (no ship until Sept 98), they announce "new" Pentium Pro's with extra cache and bus technologies ie. tweaks of existing technologies. Quote from http://www.infoworld.com/cgi-bin/displayStory.pl?970426.epenpro.htm "The processor is intended to tide the market over while Intel prepares its Deschutes chip. This processor, which has a four-way capability, has been delayed a second time and is now expected to ship in volume in September 1998, sources said. " Sooo... to quickly(?) summarize, expect more rapid announcements of "new" Intel cpu's and technologies.. all of which are designed to mask the fact that each new one is only delivering 10-15% improvements over existing technologies... Bottom line, we need to start marketing Alpha strengths (emphasize NT5 beta with encryption and VLM issues) and quit getting depressed on all of these Intel announcements.. Btw, in case there are those that think this is just a ranting Alpha nut, check out : http://www.pcweek.com/opinion/0414/14chip.html (PC Week editorial) http://www.heise.de/ct/english/9705154/ (European analysis) After looking at these 2 articles, re-read the recent Intel announcement at : http://www.infoworld.com/cgi-bin/displayStory.pl?970426.epenpro.htm Anyway, enough of my rambling - DIGITAL does have an excellent Intel story as well for those Customers who feel safe with Intel. However, IMHO, safe for a Customer is looking at where the technology requirements will be a year from now - not at what they are now. Especially if your HW decisions need to carry you a year beyond next year. Now, if we could only convince the market of this. Regards, / Kerry