Too many cooks

With such a disparate mix of software, networks and equipment, managing the IT services was becoming complicated and it was threatening to hamper development. Peter Burnham, assistant director of the Computer Centre, explains: "We felt that the University would benefit from a single platform for students. We needed to rationalise our services, bringing them up-to-date and provide a solid platform for future development. The NetWare services were essentially coming to the end of their life. Although they were running a very good service, users were demanding a more modern interface, new applications software would not run under Windows 3.1 and we were starting to see problems with software drivers for new PC hardware."

Microsoft’s round of Win2k presentations coincided with the IT department’s internal search for a new solution. The choice was between Microsoft and NetWare; the University already had a major commitment to Microsoft Office and felt that Novell was not making NetWare’s future path clear at the time. "We didn’t consider Windows 9x as sufficiently secure for student usage and were not happy with its handling of roaming users," says Burnham. "The nature of our site meant that conversion would be at least a two year exercise. If we went with NT 4.0, we would have finished conversion at about the time Microsoft would be releasing its replacement."

Burnham had confidence in Microsoft’s strategic direction but was not totally confident in the time scale that the University had imposed on itself. When the plans were originally drawn up, NT 5.0 (as it still was then) was expected to have been released. It was an option that carried a certain degree of risk, but the alternatives were not seen as attractive as Burnham explains. "Continuing to run Windows 3.1/NetWare 3.12 would have been unpopular; we would have been unable to run new software and had problems connecting new users. Moving clients to Windows 9x would have given us a less secure service and manageability problems and a move to NT 4.0 would have necessitated a redesign and rebuild within two years."

Never ending story

As Microsoft was betting its future on NT 5.0, it seemed to be the obvious choice for Burnham and his colleagues. But how have the delays to the release date affected the project so far? "We’ve ended up rolling out a significant service on a pre-released product. This was not our intention when we first proposed the idea in 1997, and can be seen to carry a greater risk. At the end of 1988, we redefined our rollout to reduce the service commitment during 1999." The university’s current Win2K service has over 4,000 registered users, sharing almost 900 Win2K clients and concurrent logins exceed 500 users. Some growth is expected during the year and the next big change should be in the summer of 2000. By autumn, there will be around 15,000 registered users sharing maybe 2,500 clients with concurrent logins exceeding 1,500 users.

The university became part of Microsoft’s Rapid Deployment Programme which provided a level of support that it could not otherwise have expected, providing assistance in terms of information provision and fault reporting. Although Burnham has not yet got SMS working with Win2K, everything else that he was expecting to find in the new operating system was there or sorted out due to the Rapid Deployment Programme. "These relate to security of PCs in our Open Access Areas," explains Burnham. "We needed to have certain policies associated with machines rather than users and we had some problems with programmable disk quota settings."

Most problems were dealt with through upgrades to the OS. The early beta releases were incomplete, causing some stumbling blocks, but most of these issues pre-dated the user service which started with Beta 3 which, according to Burnham, is a very impressive and reliable product: "The quality of Beta 3 gave us confidence to roll out our initial service on pre-released software." New hardware was necessary for servers and client devices, but this was handled as part of a replacement cycle and therefore did not come as a blow.

Getting better all the time

Burnham feels that the development of the service that his department supplies will be facilitated by this implementation. He has already installed an impressive list of software and intends to get the most out of subject specific software as it is installed. He needs to provide an improved disk quota mechanism, which might be based on the Computer Centre’s own code around the Microsoft quota mechanism. He is also looking at third party offerings. The Windows 2000 implementation will give the university a sound base for at least the next five years, ten with more modest upgrades. "I don’t think that there is any doubt that it will be an industry standard for several years and we should take advantage of our early involvement and experience," says Burnham. And how would he respond to the detractors who say that the risk was too big, taking on an as yet unreleased operating system? "You have to judge the risks from your own perspective, we were not in a ‘risk-free’ situation. Whilst our initial plans did not anticipate starting student rollout on pre-released software, in retrospect I’m glad we have. The software has worked well and has benefited from our relationship with Microsoft. This places us well for the conversion of the bulk of our users in summer 2000. After that we reap the main benefit of a common platform for all our students."

[an error occurred while processing this directive]