Twenty one years ago this month, ICL?s VME operating system was launched in the UK. The anniversary was an opportunity for the supplier and its watchers to mull over the future position of the system, and of traditional mainframe in general.
ICL was born from the union of ICT and English Electric, a marriage urged by the then-Labour government, and VME was expected to do well because it was British-made and no self respecting government agency would dare buy foreign goods.
Some observers suggest it only enjoyed major success in local authorities and government offices - the Inland Revenue is one of its first customers and today most of our tax details are held on VME systems - while others mourn the massive market ICL let slip to IBM. Its US rival proved smarter and bigger, despite VME being, in the eyes of many, technically superior.
The VME architecture is well respected in the industry and former chief designer, Brian Warboys, who is now professor of software engineering at the University of Manchester, believes that if today?s mainframes are stripped to the core, they would look like VME. ?In those days products lasted five years. I didn?t think it would still be around in 20-30 years. I was pleased with the architecture and I am still pleased with it technically. In modern developments you still see the VME approach reproduced,? he said.
He believes new environments such as client/server and networked computing echo what was possible in VME?s architecture, which was the result of three years? worth of development, started in 1969. ICL had a difficult decision to make - either continue with the many product lines that it inherited from the merger or build a system from scratch. The company chose the latter.
Philip Carnelley, analyst at Ovum, believes it was a grave mistake. A change of architecture meant major upheaval for users of George III, VME?s predecessor. ?When ICL launched its new operating system it was very avant-garde and the applications were not there. People thought that if they had to migrate to a new architecture then they may as well change to IBM. It was a very brave decision ICL made,? he said.
Warboys recalls his time at ICL. ?We were given the freedom to design a system without people moaning at us to write code - we were there to look at the design,? he said. ?That?s the difference between modern developments. We?d set out to do the whole job - design a complete system.?
The key was to design in flexibility. ?When you design a complicated system you need to lay down the architectural principles. It had to be flexible and simple because the system had to have minimal constraints. For example, if you want to build a cathedral there are certain requirements such as a steeple and arched windows - we laid down the mechanisms to make things conform to a pattern,? explained Warboys.
Because the engineering team had freedom and time, a luxury which today?s developers do not have, it was able to examine what was on the market, including learning from the technical constraints of IBM?s OS/360, which was built in 1964. Other major influences included Multis, an operating system designed with a layered architecture, and a control language from Burroughs.
The end result was a multi-layered architecture underpinned by a set of separable ?virtual machines?. Virtual machines run each program in their own ?space?. Each contains its own store, memory, operations and schedules and work independently of each other.
Brian Procter, chief architect at ICL?s high performance systems group, described the function of virtual machines by using the analogy of a meeting. ?If there are lots of people in a meeting it is difficult to make decisions quickly. You have to ask everyone and then take a vote, which takes time. It would be quicker if the group was smaller because only a few people are interested in the issue. That?s how a virtual machine works. It gives you just the things that you need to do the job.?
However the technical superiority of VME did not help ICL make its presence felt in the IBM community, which was sizable by the time VME was launched. Warboys admitted VME was not launched to kill IBM but as an alternative system to those used by customers of ICT and English Electric. ?By the time VME came out IBM was very big. Selling to the US was difficult if not illegal. If ICL?s home market was the same size as IBM?s, ICL could be as big as IBM,? said Warboys. ?ICL was not big enough [to compete with IBM] - there were no technical reasons - it was the amount of money ICL had.?
Because it made its money partly thanks to the UK?s buy-British policy of the time, ICL became complacent and avoided fighting on the world stage, said Warboys. This lack of marketing wit has hindered ICL more than anything. Thanks to VME?s virtual machine architecture it was relatively easy for the operating system to be made ?open? when the market was crying for open systems in the late 1980s. However, despite being one of the first mainframes in the world to be awarded XPG compliancy by the then X/Open Group (now merged with the Open Software Foundation to form the Open Group), this did little to help ICL sell VME in new sites or gain significant market share from IBM and other mainframe suppliers.
So what of the next 21 years? ICL is working on putting Intel processors in its next generation of VME mainframes. The aim is to bring the cost of mainframes down to the level of NT and Unix and enable users to buy a box that can house OpenVME, NT and SCO Unixware. ICL will not comment on when such a box would be available but the project is called the Millennium Programme.
Procter admits ICL does not expect the programme to attract new users but says it is a way to ?ensure that we can continue to supply systems to our customer base and that VME is still relevant to them", he said.
Ed Wilson, head of technology at the Inland Revenue, agreed. ?The millennium programme makes it more comfortable for us to use VME. Without a coherent strategy there would be a gradual erosion of the VME base and ICL would cease investing in it which would mean we would be left stranded. The programme gives us more confidence,? he said.
Wilson believes the project could give VME a new lease of life, particuarly by adopting a whole new range of third party applications software through running NT and Unix alongside VME and giving users of those systems access to VME technologies such as automatic back-up.
VME may have another 21 years in it, thanks to the growing interest in network computing and Intranets, which require heavy duty servers, but without a significant band of applications software and marketing dollars VME will be relegated to the history books.
J1043+2408 was observed for more than 10 years, and its radio light curve exhibited a periodic signal repeating in about 563 days
Success of Unity's test flight means Virgin Galactic is now close to taking its first paying tourist into space
V3 puts the pro-level football GPS tracker through its paces, and asks if it's more than a gimmick
Finding refutes many earlier studies that suggest that galaxies don't have much dark matter at the time of their birth