Comments on: Database machines and data warehouse appliances – the early days http://www.softwarememories.com/2008/09/15/database-machines/ History of software, by somebody who lived it Fri, 30 Jul 2010 14:36:06 -0400 http://wordpress.org/?v=2.8.4 hourly 1 By: Roger Reid http://www.softwarememories.com/2008/09/15/database-machines/#comment-43890 Roger Reid Wed, 17 Feb 2010 21:17:27 +0000 http://www.softwarememories.com/?p=34#comment-43890 Following up months later - I can check the Z80XX chip, I have a Britton Lee CPU board back in my office. (We were an early customer). I keep it to show young folks what a "patch" is, since BL fixed often fixed bugs by adding a patch wire to the board. That, and then I show them the memory expansion - a full megabyte of core in a 14" square board. Dunno why they call me gramps. Roger Reid (who still misses QUEL, but loves 64GB caches) Following up months later – I can check the Z80XX chip, I have a Britton Lee CPU board back in my office. (We were an early customer).

I keep it to show young folks what a “patch” is, since BL fixed often fixed bugs by adding a patch wire to the board. That, and then I show them the memory expansion – a full megabyte of core in a 14″ square board.

Dunno why they call me gramps.

Roger Reid (who still misses QUEL, but loves 64GB caches)

]]> By: Gartner 2009/2010 Data Warehouse Magic Quadrant comments | DBMS2 -- DataBase Management System Services http://www.softwarememories.com/2008/09/15/database-machines/#comment-43777 Gartner 2009/2010 Data Warehouse Magic Quadrant comments | DBMS2 -- DataBase Management System Services Wed, 10 Feb 2010 23:29:21 +0000 http://www.softwarememories.com/?p=34#comment-43777 [...] Gartner correctly says that Teradata has been a data warehouse appliance vendor from the getgo. [...] [...] Gartner correctly says that Teradata has been a data warehouse appliance vendor from the getgo. [...]

]]> By: Bill Robertson http://www.softwarememories.com/2008/09/15/database-machines/#comment-43380 Bill Robertson Thu, 28 Jan 2010 05:18:12 +0000 http://www.softwarememories.com/?p=34#comment-43380 "...with each parent node talking to two children, until it got down to the lowest-level nodes that actually talked to disk. (This is what was called Ynet.) The architecture meant that almost exactly half the microprocessors talked to disk..." This isn't quite right. The "tree structure" of the YNet (to the extent that description is correct - it's arguable) had absolutely nothing to do with which processors could access disks or the ratio of processor types. There were two types of processor types: AMPs (Access Module Processors) and IFPs (InterFace Processors). All AMPs had disks and all IFPs were diskless. One could have as few as one IFP with as many AMPs as your checkbook could buy - not that one would likely do that for fault tolerance reasons if nothing else. There was a restriction that the number of IFPs couldn't exceed the number of AMPs (as each IFP needed a "buddy AMP"). However, for a large real world system, anything close to a 1:1 ratio would have likely been irrational anyway. “…with each parent node talking to two children, until it got down to the lowest-level nodes that actually talked to disk. (This is what was called Ynet.) The architecture meant that almost exactly half the microprocessors talked to disk…”

This isn’t quite right. The “tree structure” of the YNet (to the extent that description is correct – it’s arguable) had absolutely nothing to do with which processors could access disks or the ratio of processor types.

There were two types of processor types: AMPs (Access Module Processors) and IFPs (InterFace Processors). All AMPs had disks and all IFPs were diskless. One could have as few as one IFP with as many AMPs as your checkbook could buy – not that one would likely do that for fault tolerance reasons if nothing else.

There was a restriction that the number of IFPs couldn’t exceed the number of AMPs (as each IFP needed a “buddy AMP”). However, for a large real world system, anything close to a 1:1 ratio would have likely been irrational anyway.

]]> By: Jon Forrest http://www.softwarememories.com/2008/09/15/database-machines/#comment-42532 Jon Forrest Sat, 12 Dec 2009 04:23:20 +0000 http://www.softwarememories.com/?p=34#comment-42532 If I remember correctly, Britton-Lee's database machine used Z8000s, not Z80s. If I remember correctly, Britton-Lee’s database machine used Z8000s, not Z80s.

]]> By: Curt Monash http://www.softwarememories.com/2008/09/15/database-machines/#comment-31416 Curt Monash Wed, 17 Sep 2008 22:41:16 +0000 http://www.softwarememories.com/?p=34#comment-31416 Andrew, Thanks for the corrections! CAM Andrew,

Thanks for the corrections!

CAM

]]> By: Andrew McLaren http://www.softwarememories.com/2008/09/15/database-machines/#comment-31384 Andrew McLaren Wed, 17 Sep 2008 05:01:04 +0000 http://www.softwarememories.com/?p=34#comment-31384 CAFS - the ICL product was called CAFS: "Content Addressable File Store". Yeah, it was one of those early ideas where the structure of the data and the structure of the hardware was closely aligned, like database machines. ICL won the Queen's Award for Technological Achievement for developing CAFS, sometime in the early 80s - and few other computer firms can claim a Queen's Award :-) IDMS was a popular DB in ICL systems but I don't think CAFS was in any way directly derivative of IDMS. CAFS – the ICL product was called CAFS: “Content Addressable File Store”. Yeah, it was one of those early ideas where the structure of the data and the structure of the hardware was closely aligned, like database machines. ICL won the Queen’s Award for Technological Achievement for developing CAFS, sometime in the early 80s – and few other computer firms can claim a Queen’s Award :-)

IDMS was a popular DB in ICL systems but I don’t think CAFS was in any way directly derivative of IDMS.

]]> By: Teradata decides to compete head-on as a data warehouse appliance vendor | DBMS2 -- DataBase Management System Services http://www.softwarememories.com/2008/09/15/database-machines/#comment-31324 Teradata decides to compete head-on as a data warehouse appliance vendor | DBMS2 -- DataBase Management System Services Mon, 15 Sep 2008 17:00:50 +0000 http://www.softwarememories.com/?p=34#comment-31324 [...] Teradata, which long avoided the “appliance” term, now says it sells both “data warehouse appliances” and “data mart appliances.” Indeed, it claims to have “invented the original appliance” — which is pretty close to being true.* [...] [...] Teradata, which long avoided the “appliance” term, now says it sells both “data warehouse appliances” and “data mart appliances.” Indeed, it claims to have “invented the original appliance” — which is pretty close to being true.* [...]

]]>