SHARE and the History of Mainframe Virtualization

With contributions from Jim Elliott, Reg Harbeck, and Ray Mullins

This year marks an exciting milestone in the history of mainframe virtualization. In January 1967, production use began on the CP-40, a research system on which MIT and IBM researchers first tested the concept of virtual memory. The CP-40 laid important groundwork for the development of future mainframe virtualization technology, and 50 years later the z/VM can trace its heritage back to that system.

In a recent presentation, SHARE Project Manager and mainframe consultant Jim Elliott tracked the development of mainframe virtualization from those first academic research projects through to the early ’90s. Elliott had a first-hand look at the technology’s evolution along the way. He spent 43 years at IBM, working on various iterations of virtualization technology from the pivotal CP-67 right up until today’s z/VM platform. Along the way, the objective has remained to help businesses do more with less.

Early Research Projects

MIT’s John McCarthy first proposed the idea of a “time-sharing computer system” in 1961, and an early prototype allowed up to three users to work on the same computer system – a revolutionary concept at the time.

Over the next several years, McCarthy led projects using systems from both IBM and General Electric. In parallel, researchers and mainframe vendors made progress on early research and commercial solutions, one of the most notable successes being the University of Michigan’s Michigan Terminal System (MTS), Elliott said.

Some of this work would introduce important mainframe concepts. These included virtual storage, the ability to run multiple concurrent users in the same environment, and IBM’s early time-sharing operating system (TSS) – an offering that failed to gain ground within the company and that was ultimately discontinued, Elliott said. But, the systems that would more closely match our current view of mainframe virtualization would come soon after.

The CP-40 and CMS

In conjunction with MIT, researchers at the IBM Research Cambridge Scientific Center began digging further into the concept of time-sharing in a project in the mid-60s. By modifying an S/360-40 system, they developed the two key software components:

  • The CP-40, or “Control Program,” which supported 12 VMs
  • The CMS, or Cambridge (later Conversational) Monitor System

The CP-40 was ultimately the first system to demonstrate virtualization – the concept of a virtual environment that simulates a real physical machine – as we know it today. CMS included simplified commands and technologies that would ease the adoption of virtualization technology.

In 1967, CP-40 evolved into the CP-67, which could support up to 24 VMs and a variety of guest operating systems. It included memory and a virtual CPU, an operator console, and support for a range of devices, including a printer, punch card reader, minidisks, tapes, a paging device, networking controller and a display console. CP-67 would gain widespread adoption and directly influence future systems, from the VM/370 in 1972 right up to today’s z/VM solutions.

And although the CP-67 and its predecessors were created 50-plus years ago, they relied on the same original design principles that still serve as the DNA of modern mainframe virtualization solutions. From the notion of a hypervisor, to interfaces that interact with the hypervisor, to VM accounting, performance monitoring and security management, these principles informed systems that continue to shape how businesses work today.

The Constant Through It All

While mainframe virtualization has evolved over the last 50 years, one thing has remained steadfast – the powerhouse network of SHARE members. This is a community that collaborates with intent, learns from each other daily, and allows for such advancements on the mainframe to come to fruition.

Over time, SHARE members have had tremendous influence on the progress many celebrate today. “These are the deeply intelligent, responsible, hard-working, honorable technologists of the same ilk that founded SHARE and built the mainframe. These world-class people are quietly, almost invisibly, keeping the world economy going by supporting the greatest computing platform in history,” said Reg Harbeck, SHARE Program Officer. “As I learn the depth, history, strategy and latest innovations on that platform, I learn it from the most amazing people… These are people I might meet one or two of in an entire career if I didn’t attend SHARE.”

It’s these unique connections people like Reg have formed that make not only the last 50 years of progress worth reflecting on, but the opportunities that lie ahead, as well. “SHARE Members can continue to shape the future direction [of mainframe technology] by sharing their idea and opinions with IBM and fellow members, through defined mechanisms such as the RFE process, formal (ExecuForum) and informal (Birds Of A Feather, hallway encounters) discussions, attending the SHARE Technology Exchange… and, best of all, volunteering for SHARE,” explained Ray Mullins, SHARE Project Manager. “[This] is something I've always wanted to do—help this industry that has been an important part of my life for almost four decades.”


Recent Stories
Using IPCS to Solve CICS Failures

Tips and Tricks: The Transition to Remote Work

Master the Mainframe 2005: Anna Clayton's Springboard to Success