The following message is a courtesy copy of an article that has been posted to bit.listserv.ibm-main as well.
[EMAIL PROTECTED] (Thomas Kern) writes: > I thought the early prototype of this was the Single-System-Image code > written at University of Waterloo back in the early 1980's. I tried to > convince management that it would be cheaper to use it to glue > together a slew of surplus 4341s than some of the other alternatives. re: http://www.garlic.com/~lynn/2008o.html#55 Virtual http://www.garlic.com/~lynn/2008o.html#56 Virtual Separate from the virtual machine based commercial timesharing service bureaus http://www.garlic.com/~lynn/submain.html#timeshare As part of consolidating the several US HONE datacenters in a single location (northern cal) in the mid-70s ... there was work on supporting single-system image. http://www.garlic.com/~lynn/subtopic.html#hone By 78/79 there was front-end load balancing and other single-system-image support ... across multiple multiprocessor machines in large loosely-couple environment (at the time, possibly the largest single-system-image operation anywhere). Then because of natural disaster considerations ... the load-balancing was extended to a replicated 2nd HONE datacenter in Dallas and then a replicated 3rd HONE datacenter in Boulder (there were approaching 40k defined userid on the US HONE system complex ... and mainframe orders couldn't even be submitted w/o first having been processed by HONE). Note that while the HONE support provided load balancing across the complex and various other single-system-image transparency ... it didn't support process (virtual machine) migration between different machines in loosely-coupled complex. In the very early 80s, SJR started a 4341 vm-based cluster project using 3088/trotter (this was before moving up the hill to almaden). One of the big problems before being released as a product, they had to migrate the implementation to standard SNA protocol. This had disastrous effects on the cluster operation efficiency. For instance, the original cluster syncronization process that took very small subsecond elapsed time, increased two orders of magnitude when migrated to standard SNA protocol (over 30 seconds elapsed time). -- 40+yrs virtualization experience (since Jan68), online at home since Mar70 ---------------------------------------------------------------------- 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
