mp...@suse.com (Mark Post) writes: > http://en.wikipedia.org/wiki/IBM_Sequoia
long ago precursor ... working both with LLNL and other national labs as well a commercial RDBMS regarding cluster scaleup ... old email http://www.garlic.com/~lynn/lhwemail.html#medusa and old reference to commercial RDBMS cluster scaleup meeting the first part of jan92 http://www.garlic.com/~lynn/95.html#13 as part of our HA/CMP product (some past posts) http://www.garlic.com/~lynn/subtopic.html#hacmp senior vp retires oct91, followed by audit of several products he sponsored including supercomputer effort, part of the result is scoring the company for technology to use, including corporate technology conference middle of jan92. by the end of jan92 (possibly within hrs of the last referenced email in above), cluster scaleup is transferred and we are told we couldn't work on anything with more than four processors (significantly contributes to motivation deciding to leave). a couple weeks later there is supercomputer announcement (17feb1992 press reference ... but no commercial, numeric intensive & scientific *ONLY*) http://www.garlic.com/~lynn/2001n.html#6000clusters1 later that spring another press item that clusters had taken the company completely by surprise (11May1992 press reference): http://www.garlic.com/~lynn/2001n.html#6000clusters2 i've mentioned before having worked with LLNL off&on dating back to benchmarks when they were looking at large numbers of 4341s in computer farm in the late 70s ... some old 4341 email http://www.garlic.com/~lynn/lhwemail.html#4341 also later in 1988, I'm asked to help LLNL standardize some serial stuff ... that eventually morphs into fibre channel standard (FCS). Some pok channel engineers then become involved and layer protocol on top that severely restricts native/underlying FCS thruput ... that eventually morphs into FICON. some recent ficon posts: http://www.garlic.com/~lynn/2013.html#10 From build to buy: American Airlines changes modernization course midflight http://www.garlic.com/~lynn/2013.html#40 Searching for storage (DASD) alternatives for other topic drift ... past posts about original RDBMS/SQL implementation http://www.garlic.com/~lynn/submain.html#systemr it and derivatives were mainframe only. In the time of ha/cmp, some of the other rdbms vendors had "portable" implementations for unix, vax/cluster and other platforms. IBM is just starting on its implementation for OS/2 (but isn't available for other platforms or with cluster support until much later). To aid in easing deploying vax/cluster on unix platforms, I work on global lock manager that emulates the vax/cluster's API. However, the various vendors have list of dozen or things done wrong in vax/cluster ... which are easy to correct since I get to start from scratch. The mainframe DB2 folks complain that if I'm allowed to continue, I'll be at least five yrs ahead of them (which presumably contributes to cluster scaleup being transferred and being told couldn't work on anything with more than four processors). -- virtualization experience starting Jan1968, online at home since Mar1970 ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN