Re: OpenVMS Seminar in Toronto (2005-02-24) a few points
From: Robert Deininger (rdeininger_at_mindspringdot.com)
Date: 02/28/05
- Next message: Nigel Barker: "Re: OpenVMS Seminar in Toronto (2005-02-24) a few points"
- Previous message: Robert Deininger: "Re: OpenVMS Seminar in Toronto (2005-02-24) a few points"
- In reply to: JF Mezei: "Re: OpenVMS Seminar in Toronto (2005-02-24) a few points"
- Next in thread: JF Mezei: "Re: OpenVMS Seminar in Toronto (2005-02-24) a few points"
- Reply: JF Mezei: "Re: OpenVMS Seminar in Toronto (2005-02-24) a few points"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Date: Mon, 28 Feb 2005 06:44:28 GMT
In article <1109564155.365a38142e9cca1b572b7c75bacb949f@teranews>, JF
Mezei <jfmezei.spamnot@teksavvy.com> wrote:
>One more thing about the environmental issue:
>
>EV7 came out last year, the Galaxy-class machines to support EV7 came
>out last year. Not a decade ago, not 5 years ago. Last year.
But the design work was started much earlier. There were prototype
systems in 2001. Once you start building protos, wholesale churning of
the BOM would be very bad. It would be like starting over.
>These machines were conceived at a time where those environmental
>guidelines were known.
Conceived? I don't think so. The development cycle is fairly long. I
don't know when the regulations were changing, but the Marvel designers
did not seem to be planning for this. It was not a consideration on the
table.
>If Superdomes are able to support those
>standards, then Alphas should too.
There is a great deal of turmoil around the HP-classic products to deal
with the changing regulations, now. Some stuff will EOL sooner than
originally planned.
>When they cam out with the DS15, they could have ensured that any order
>for chips was made for components that didn't have the nasty stuff that
>was to be banned shortly (whatever the "ban" really means).
It's not just the chips. It's virtually everything in the system: ***
metal, plastics, knobs, fasteners, wire. Checking an existing BOM for
compliance is a big job. Many of the suppliers don't know if their stuff
complies. They have to go investigate. Replacing non-compliant
components requires an engineering team with roughly the same skills as
the original design team. The reality is that the Alpha design teams have
been mostly dissipated. They are doing other work, either inside HP or
elsewhere.
This problem was not on the radar when DS15 was designed. Maybe it should
have been.
>It isn't as if those standard just popped out without warning. If HP had
>time to prepare Pa-Risc, 8086 and IA64 machines for thsoe standards, it
>had time to do so with Alpha which also got new models recently.
Work is ongoing to bring the Integrity servers into compliance. (They
aren't in total compliance today.) The good news is, there are teams in
place that can do the work. The bad news is, they have to postpone
interesting work on future products to deal with this.
>Please stop finding excuses NOT to do things with Alpha/VMS and instead
>find ways to make it possible. That is what succesful companies do.
This being the real world of 2005, the profitable options for AlphaServers
are rather limited. If you think that is an excuse, well, excuse me.
>If even the VMS folks are resigned to believe the EXCUSES to stop Alpha,
>EXCUSES not to market VMS, EXCUSES about Alphas being more expensive,
>then will any of these supposed problems ever get fixed ?
The solution to too-expensive AlphaServers running VMS is less-expensive
Itanium servers running VMS. I don't know anyone who has any realistic
ideas to reduce the cost of AlphaServers today. It's too late.
>You make it look like it is physically impossible to build a low cost
>alpha machine and this pointless to even try.
No, see my other posts. I don't think you could build the existing
designs much cheaper than the current costs. Could you redesign for lower
cost? Yes, probably. Would you recover your design investment at this
late date? That seems doubtful to me.
Designing much lower cost AlphaServers would have been a wonderful idea 3,
5, or 10 years ago. Alas, they didn't seem to be able to accomplish it,
and they did try. I don't think there's much point today. Now, the
effort would be better directed at improving the Itanium server line.
>For a DS10, you don't
>need 2" steel plating and bulltproofness that yoU,d expect from a galaxy
>class machine.
That is true, and I think Digital struggled with this concept for many years.
It is also true that for a DS10, many VMS folks don't need a DS10. An
rx1600 will do just fine. That reality seems to infuriate some folks.
- Next message: Nigel Barker: "Re: OpenVMS Seminar in Toronto (2005-02-24) a few points"
- Previous message: Robert Deininger: "Re: OpenVMS Seminar in Toronto (2005-02-24) a few points"
- In reply to: JF Mezei: "Re: OpenVMS Seminar in Toronto (2005-02-24) a few points"
- Next in thread: JF Mezei: "Re: OpenVMS Seminar in Toronto (2005-02-24) a few points"
- Reply: JF Mezei: "Re: OpenVMS Seminar in Toronto (2005-02-24) a few points"
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]