The following message is a courtesy copy of an article that has been posted to bit.listserv.ibm-main,alt.folklore.computers as well.
[EMAIL PROTECTED] (Timothy Sipples) writes: > Nobody said Parallel Sysplex and GDPS are the only high availability > clustered solutions in the market. But this whole thread got started > because of a complaint about *planned* outages. One must not be sloppy > here: "five nines" should have a business definition, and that definition > does not typically distinguish between planned and unplanned outages. (Or > at least people should say something like "five nines, excluding planned > outages of up to [X] duration [Y] times per year.") If you're down, you're > down. re: http://www.garlic.com/~lynn/2008i.html#97 We're losing the battle http://www.garlic.com/~lynn/2008i.html#99 We're losing the battle http://www.garlic.com/~lynn/2008i.html#101 We're losing the battle http://www.garlic.com/~lynn/2008j.html#7 We're losing the battle http://www.garlic.com/~lynn/2008j.html#10 We're losing the battle when we were out marketing ha/cmp http://www.garlic.com/~lynn/subtopic.html#hacmp ... against tandem (as well as s/88 aka stratus) ... there was a customer with five-nines application availability requirement (five minutes outage/annum). the non-clustered fault-tolerant solutions had software maintenance (scheduled) outages that far exceeded 5min/annum. we had also coined the term disaster survivability and geographic survivability ... i.e. clustering at a distance ... as hardware and other components become more & more reliable ... localized disturbances were becoming a larger percentage of unplanned outages. http://www.garlic.com/~lynn/subtopic.html#available as mentioned earlier in the thread ... long ago and far away, my wife had been con'ed into going to POK to be in charge of loosely-coupled architecture ... where she created peer-coupled shared data. http://www.garlic.com/~lynn/subtopic.html#shareddata Lack of uptake (at the time) resulted in her not staying long in the position. Except for ims hot-standby ... it wasn't until sysplex that you started seeing her architecture being supported. the long mainframe lead time ... was at least partial motivation for ha/cmp product (based on power platform rather than mainframe platform). it was also behind POK & Rochester objecting to ha/cmp contributions to the corporate continuous availability strategy document ... claiming that it would be years before they could have such support. some folklore x-over ... Bruce's talk last month at Jim's tribute pointed out that his formulization of transaction semantics was the real significant enabler opening up online transactions (sufficient trust in computer operations vis-a-vis manual/paper operation). This was during the days of the original relational/sql implementation project at san jose research on vm/cms platform http://www.garlic.com/~lynn/subtopic.html#systemr Following Bruce's talk was some people from tandem (corresponding to Jim having left research for tandem). Two things mentioned in that time-frame was Jim defining the TPF thruput (ACP having been renamed TPF as more non-airlines started using it for transactions) as a transaction objective for Tandem systems. The other was the study showing that hardware was becoming significantly more reliable and other factors were increasingly becoming source of outages (planned, human mistakes, disturbances in localized geographical area). Jim later left Tandem for DEC database group in San Francisco. It was in this time-frame that I had something of an argument with him at '91 Asilomar SIGOPS ... where I was claiming I could do high availability on (clustered) commodity hardware (using ha/cmp methodology as example) and he claiming that it still required proprietary hardware (somewhat reflecting the Tandem and DEC vax/cluster affiliations). I've since noted that not too long later, he then had to be up on the stage for the announcement of the m'soft availability clustering ... recent reference: http://www.garlic.com/~lynn/2008i.html#50 Microsoft versus Digital Equipment Corporation ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
