Today's surviving VS world consists mostly of single systems. In my time working with the VS since 1984, though, it was not uncommon to see companies that had 10, 20, 30 or more VS systems, often dispersed at different facilities but sometimes grouped at single sites.
The Hartford, an insurance giant, had something like 80 large VS systems at sites around the U.S.
Mellon Bank and its subsidiary Mellon Mortgage had a WSN network of 45 mostly large VS systems. I consulted for Mellon Mortgage, which had 18 of those systems at about six sites. The central site in Houston had Wangnet, the broadband campus networking system. CATV cabling ran up through the building and NETMUX devices supported peripheral clusters on each floor. The VS systems in the computer room used Wangnet to communicate with each other via CIUs. The VS systems ran a major PACE mortgage application but were also used as mainframe connection gateways and printed 900,000 pages of mainframe reports per month on two large Xerox laser printers connected to the VS systems via SPUR printer management subsystems.
The Arkansas Department of Health had (and still has) seven large VS sytems serving 2,200 users 24 x 7. This has been the largest RSF cluster in the world, all at one site. The Dept. also had mobile VS5000 systems that used to set up in places like supermarket parking lots to administer vaccinations. Whenever a phone connection was available the mobile systems uploaded the accumulated vaccination records to the central site.
Exxon USA had 13 VS300 systems in one computer room in Houston. I was told that their main function was to make mainframe reports available to executives via the DISPLAY utility, which was apparently much more functional than any utilities available through direct mainframe connection. Their building was equipped with a massive Wangnet broadband cabling system than ran up and down the building in a utility riser and branched out to encircle each floor.
Kent Electronics and its manufacturing subsidiary K*Tec Electronics ran many VS systems. Kent grew its central site to three VS12000-class systems in an RSF cluster serving about 700 users plus VS5000 systems in branch offices in a WSN network. K*Tec grew its central site to a pair of VS12000-class systems in an RSF cluster serving about 400 users plus a number of VS5000 systems in branch offices.
One of the largest VS networks was operated by the U.S. Department of State. Every State Department facility in the world had at least one VS system. I don't know how many VS systems were in that network but it had to be several hundred.
Many VS networks disappeared in the 1990s for two reasons: The growth of wide area intranets allowed desktop PCs and network printers to be supported through Lightspeed gateways relocated to the central site. This eliminated the need for small VS systems in local offices. The real tragedy, though, was that no one realized that WSN HDLC connections could be tunneled through IP networks. The network people didn't know that WSN used the HDLC protocol and the VS people didn't know enough about the details of WSN to explain it to the network people. In fact Cisco's Serial Tunneling (STUN) feature neatly provided for tunneling WSN connections through corporate IP networks. The false belief that WSN networks still required costly dedicated circuits led to critical examination and dismantling of WSN networks and elimination of networked VS systems without anyone realizing that the WSN traffic could have been shifted to the growing company WANs. It still might have made sense to relocate Lightspeed gateways to a central site and connect remote PCs via the new WANs, but the real driving force behind removing WSN networks was the false issue of expensive leased circuits for the WSN systems.
Another factor was the abandonment of Wang Office for email. One of the last large Wang Office sites was Mellon Bank, with 16,000 Wang Office users. As email moved to PC and server-based products such as Microsoft Mail and later Microsoft Exchange, the role of the VS as the mission critical email system lessened and eventually disappeared. This occurred despite the availability of good email gateway products from Wang and Lightspeed and possibly others. The VS was able to interoperate with Internet and intranet email. As happened with various aspects of the VS, many customers simply didn't know what the VS could do and Wang didn't do a very good job of keeping them informed. Even so, Mellon Bank's Wang Office community was still functioning into the late 1990s.
When we demonstrated the New VS technology to Getronics in 2004, everyone present agreed that it would be possible for customers to have a virtual VS on every desktop if they chose to do so, and that that would be powerfully advantageous. The luxury of having an entire VS serving a single user and running blazingly fast, which some of us only experienced when working at night or on weekends, would be possible thanks to the virtualization. Unfortunately the economics of making the virtual VS available to the VS community have not yet allowed pricing that makes this possible. The only way to pay the bills at Getronics (now Compucom) and TransVirtual is for the New VS to be priced comparably to the legacy VS, although we have managed to deliver significantly more performance for the dollar than was ever possible before.
One way to enjoy an abundance of New VS systems is to negotiate a site license. We have a couple of customers who have site licenses. They are relatively free to deploy as many New VS systems as they like within their own organizations. While we have strict requirements for the New VS platform in most situations, capable customers willing to take responsibility for the consequences may deploy the New VS on non-compliant platforms such as desktop PCs, particularly for non-critical functions such as software development and maintenance. Developers at Compucom and TransVirtual run the New VS on ordinary PCs. Our principal developer at TVS runs an ordinary 2.8 GHz Dell desktop that outperforms a VS18950. I run multiple VS instances in VMware Workstation on a server grade Dell Precision tower that has two Dual Core processors at 3.0 GHz and 8 GB of memory. We have a farm of IBM x330 dual 1.4 GHz 1U rackmount servers at our office for internal use and for running customer remote evaluations and demos.
The main reason we have strict requirements for most customer systems is the problem of supporting varied platforms. For us to directly support customer systems the platforms have to meet certain requirements and the platform features have to be completely known to us. They must also have remote control that allows us access to boot-time CMOS and setup functions. The Dell DRAC provides us with console access from power up all the way through Linux and virtual VS startup. In fact, most of our overseas systems have been remotely installed from Houston.
Anyway, the VS world at its height had many, many multiple VS sites. Some were RSF clusters and all were WSN networks. The largest WSN network was Wang's own, consisting of more than 900 systems in one network, all accessible for remote logon, file transfer and Wang Office distribution of email and packages.