The following message is a courtesy copy of an article that has been posted to bit.listserv.ibm-main,alt.folklore.computers as well.
eric-ibmm...@wi.rr.com (Eric Bielefeld) writes: > This is kind of off the topic, but related. Wasn't there a discussion > on IBM-Main a couple years ago about the air traffic control system > being run on old IBM 3081s? If I remember right, and my memory isn't > as good as it used to be, it was just a few years ago that these old > machines used to control some of the air traffic control. Have these > been retired yet? re: http://www.garlic.com/~lynn/2009q.html#26 Check out Computer glitch to cause flight delays across U.S. - MarketWatch http://www.garlic.com/~lynn/2009q.html#28 Check out Computer glitch to cause flight delays across U.S. - MarketWatch we had project a decade ago with some of the people that had done the 1960s implementation ... and later left to form their own company to do various things. these were modified 360/50s that ran in triple configuration. this comes up in the virtual machine folklore. the cambridge science center was trying to get a 360/50 to build their hardware modifications for supporting virtual memory ... but because all the available 360/50s were going to air traffic control ... the science center had to settle for 360/40. the first virtual machine system then was "cp/40" (instead of "cp/50") ... morphing into cp67 (when they got a standard 360/67 that came with hardware virtual memory support) ... which later morphed into vm370. misc. past posts mentioning science center http://www.garlic.com/~lynn/subtopic.html#545tech some of that early 360 FAA software was moved along to various platforms ... including some of it eventually running in Flex-ES virtual machines on intel platforms (possibly even stratus intel ... as mentioned in previous post). in the late 80s, we started ha/cmp product ... misc. past posts http://www.garlic.com/~lynn/subtopic.html#hacmp and part of that effort, we did detailed availability studies of tcp/ip and various tcp/ip environments. there have also been some number of (failed) FAA "modernization" efforts. About the time we were doing ha/cmp ... there was such an effort using triple-RS6000s as basic component. Because of our work on high availability ... we were periodically asked to participate in some of the discussions/reviews. There were some number of interesting failure-modes that they overlooked ... and it was one of the "modernization" efforts that ran into difficulty. After we left in the 90s, we were invited in to consult with small client/server startup that wanted to payment transactions ... the startup had also invented this technology called "SSL" they wanted to use. Two of the people that we worked on ha/cmp for parallel Oracle ... referenced in jan92 meeting mentioned in this post http://www.garlic.com/~lynn/95.html#13 were at this startup in charge of something called the "commerce server". As part of doing what is now frequently called "electronic commerce", we deployed something called the "payment gateway" ... misc. past posts http://www.garlic.com/~lynn/subnetwork.html#gateway which acts as gateway for payment transactions between webservers on the internet and payment infrastructure. An early prototype of this had a situation where a merchant called up not being able to do transactions. Normally, payment trouble-call desk has 5mins elapsed time to do first level problem determination. In this case, three hours later, the trouble ticket was closed with no-trouble-found. In order to try and approach the non-internet availability ... we deployed payment gateway on HA/CMP configuration with multiple diverse routes ("telco provisioning") into different parts of the internet. we also had to deploy/invent some number of compensating procedures to compensate for vaguries of the internet ... as well as to compensate for large number of identified security vulnerabilities. One of the issues ... was that I had planned on also broadcast/advertising the different routes ... but in the period of the deployment; the internet backbone transitioned to hierarchical routing. As a result, the remaining alternate path mechanism had to rely on multiple-A record support in DNS (where the client is provided multiple ip-address records in response to request domain lookup). The client then cycles through each of the addresses until it finds one that makes a succesful connection. In any case, part of that deployment (for "electronic commerce") included compensating for large number of different kinds of failures that might happen anywhere in the internet infrastructure. -- 40+yrs virtualization experience (since Jan68), online at home since Mar1970 ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html