"The Turning of the Wheel" –
IEEE Computer Magazine
Neville Holmes, University of Tasmania; July, 2005
To: Neville Holmes <firstname.lastname@example.org>
From: "David B. Tuttle" <DaveTuttle@comcast.net>
Date: July 12, 2005 10:50 PM
Subject: The Turning of the Wheel, Indeed...
Thank you for the fine article in July's "Computer"
magazine -- in spite of it's costing me a couple of evenings reading time! I find it very hard to
resist re-reading Melinda Varian's paper when I run across a topical reference to it.
On the subject of virtual machines and other such beasts, it is actually difficult to describe the technical distinctions between a "true" virtual machine, a "hybrid" virtual machine, an emulator, an interpreter, and a "pure" simulator, in terms that a user would care about. The differences are very real but often not very significant to the consumer of the service. Over the years I have used, helped create, and exploited all of the different breeds. As was true with the evolution of CP-40 -> CP-67 -> VM/370 -> VM/XA, they are sometimes varying aspects of the same underlying system. The current taxonomy is flawed, certainly, but perhaps there are few of us left that have the benefit of the history first hand.
My first encounter with a time sharing or "virtual" system was the Summer of 1966 in New York City, with a terminal connected to the M44/44X system at IBM Yorktown Research. Next was the fall of the same year, when I wheedled my way into the MIT Computation Center as a freshman and worked as a part-time user support consultant for CTSS and IBSYS. Summer 1967 found me spending 60 or so hours a week as a system programmer for the MIT S/360-65 and S/360-40 machines, bringing up ASP 1.0. Fall 1968 found me with a part-time job at the IBM Cambridge Scientific Center. And so it has gone, though IBM derailed one possible path when they moved the VM/370 crew to Poughkeepsie.
After 20 years or more working in the data communications realm, I was pleased (if not too surprised) to become a daily user of VMWare 1.0 in the Fall of 1999. The small group I joined was doing Linux-based development in a larger company that had its IT built around WindowsNT and Lotus Notes, so we ran NT in a virtual machine under Linux (RedHat 5.2, initially). Now I find myself doing exotic Linux kernel work with an eye to better exploiting multiprocessor systems and multi-core processors, for communications (do you hear the wheel turning?). The next set of chips we are planning to use are currently embodied as an SDK and companion simulator / emulator, of course. One of the bread-and-butter uses for "virtualization" has always been concurrent development of hardware and software. Eighteen months ago, though, I was part of a startup team that did it precisely backwards, using software and generic hardware to make multiple "real" machines into one or more "virtual" machines of whatever size you want.
One of the aspects of our profession that is seldom discussed is the 'journeyman' effect. The people who are most recognized are the ones that can stay with a technology thread.or a particular system or a particular company. With relocations and reorganizations and new companies and new technologies, some of us have to keep on finding new things to work on and new ways to pay the bills.
– Dave Tuttle