IBM and cloud computing: Back to the future

Back to the future. What goes around comes around. Circle of life. Déjà vu all over again. And so forth.

I spent two days this week at an analyst meeting listening to and talking with some very smart IT men and women, a group of marketing and development managers at The IBM (IBM) Company. The meeting’s title was “Next Generation Linux and Open Source.” But to these IBM managers and executives that phrase is redundant. Linux and adjunct open source software (OSS) is already the next-generation technology. And these guys also said that OSS is only a technology tactic, which I was happy to hear from an investment research point of view. For these IBM folks, OSS is a means for getting to the next generation of computing, cloud computing.

Coincidentally this week, I was looking for a 2-year-old white paper of mine that I knew was on the web but was not on the laptop I was carrying with me. While googling my own name, I came across a 13-year-old document, Moving Mainframe VM Users to a Distributed UNIX System. The article is available on the Association for Computing Machinery [ACM] portal--subscription required--and its author apparently used some of my 1995 research into client/server computing acceptance in the market to make his point. In the cloud computing era that IBM sees coming, large enterprises and organizations serving IT resources to small and medium enterprises—let’s call these organizations Service Bureaus for lack of a better name—can reverse the process described in the title of the 1995 ACM document. Already I was told at the IBM Linux/OSS analyst meeting, users can move distributed applications (be they UNIX, Linux, Windows or legacy) back to a VM mainframe, the latest IBM Zseries system.

But for me, that 13-year-old research at the ACM site is only one third of the way back to the future. As IBM described its view of the cloud computing, I flashed to the Multiplexed Information and Computing Service (Multics), one of the first IT products about which I ever did research. According to its alumni web site, Multics is a timesharing operating system begun at MIT in 1965, used in production (marketed by Honeywell and subsequently Bull) until 2000, and now available from MIT under open source terms and conditions.

But while Multics was groundbreaking for its virtual memory, IBM’s cloud features virtual everything, including administration. Multics was all about Utility Computing but had no “dynamo;” IBM’s cloud features content and context as well as the transmission lines. For Multics the last mile, at least originally, lead to a very small number of primitive terminal (initially a teletype device) for use by programmers. IBM’s cloud potentially interfaces billions of individuals, trillions of sensors and (whatever the next order word is) of bytes of data.

But from an IT investment research perspective, the question is “Can IBM put the PC genie back in the bottle after all these years?”

In 1965, before the Multics operating system itself was even written, two members of the Multics team wrote a sociological view of what utility/cloud computing would mean. Its findings/warnings ring pretty true today. Read its description of a theoretical automated tax-calculation program that might be written to be used on the Multics system. The implication was that the government would write and possess this Big Brother application. The sections in the 1965 paper about The Threat to Privacy and the Cult of Impersonality say it all 43 years later.

Of course, the complex government-owned application that the Multicians described in 1965 became available to the masses 20 years later on Wintel PCs from companies such as ChipSoft/Intuit and Meca. As a result many people now prefer to control their own IT destiny, and especially things like their own tax data. If cloud computing as IBM envisioned it is to become pervasive, these same concerns need to be considered vis a vis documents, search histories (I assume that train has already left the station?) and all other aspects of our lives. Businesses share the same concerns as individuals in terms of sending their crown jewels out into the ether but on a different scale.

Whatever happens to the IBM cloud computing vision, there is no doubt its technology will provide a very real payback for large enterprises, like IBM itself, with thousands of servers deployed. And it was great to see something that was just a concept so many years ago finally coming to life.

(Oh by the way, to close the circle with a double knot, prehistoric IT campfire storytelling tradition has it that the name UNIX derived in a backasswards way from Multics. Thompson and Ritchie lost access to the Multics system, on which they did some prehistoric gaming (as well as some real work) when Bell Labs dropped out of the Multics consortium in 1969. So supposedly they wrote a much simpler operating system to take its place, which colleagues called the un-Multics, or Unics for short. A more professional version of what happened, written by Dennis Ritchie himself, is available here.)