The city of Dundee is unlikely to feature in most lists of UK technology hotspots. But that might be an oversight.
It is arguably the ATM capital of the world, leading NCR’s manufacturing and R&D activities around the banking machines; it is the largest biotech centre in the UK outside of Oxford and Cambridge; and it is the home of the notorious series of videogames, Grand Theft Auto, which was developed by local company Rockstar.
But another technology leader is boosting the area’s digital credentials further. Over the past 10 years, Dundee City Council has emerged as one of the UK’s foremost testbeds for advanced server-based computing as it has made extensive use of virtualisation technologies to move its employees’ PC services off the desktop and into the data centre.
Ged Bell
A 28-year veteran of public and private sector IT practice, Ged Bell is responsible for the design, architecture, procurement, deployment and support for all information and communications technology services across Dundee City's local authority and its schools. He has been an advoc-ate of server-based desktop service delivery for almost a decade.
Ged Bell, head of IT implementation at Dundee City Council, has been an advocate of server-based desktop delivery since 1997 when he started to introduce Citrix-based servers into different council departments. But in 2005, he abandoned Citrix in favour of an architecture based on central blade servers running “raw” Microsoft Terminal Services, with applications virtualisation provided by Softgrid (a software specialist since acquired by Microsoft). Now, of the council’s 4,000 PCs, 2,800 are thin-client terminals.
Aside from the cost savings on PCs, such a server-based architecture enables users to access their applications from anywhere, via an SSL VPN, with no local data storage. The set-up also helps with security and data retrieval, easing efforts to comply with privacy and freedom of information legislation.
The centralisation of this architecture was driven by demands for standardisation and consolidation of independent departmental “fiefdoms”, says Bell. “Thin client is now the default mode of delivery for us – the consolidation is taken right down to the printer and CD level. Standardisation is something we’re really driving for. Wherever you go in the department, you see the same desktop and a standard list of apps. Any app you need specifically to do your job will be delivered in a standardised way.”
However, standardisation can only go so far. “We have 300 applications that have very small numbers of users but are very important as far as those users are concerned,” says Bell. “These were not developed for a networked environment, which was a barrier to thin client penetration. We needed a way to run these apps securely and safely in a thin client on the server without impacting on everything else on that server.”
This was solved by Softgrid’s application virtualisation product, which allows any application to be run on a small footprint isolated from other applications. The 300 applications are all “sequenced” to prepare for streaming to the endpoint. This means running an installation routine on top of the Softgrid sequencing application, creating a large flat file which is then stored on the separate Softgrid server. Here it runs in a protected shell, so that nothing on the server is changed.
“As a by-product, it gives us better license management and control,” says Bell. Maximum concurrent user levels can be set through the server, which has brought cost savings in architecture and engineer departments, where expensive applications were being underused.
The tangible benefits are greatest in the overall hardware and software spend: the outlay on PCs and servers has fallen by at least £150,000 a year. “That is the number we can prove, but as we are also growing our infrastructure, we are close to adding another 50% to that figure,” says Bell. “Over time we have also been able to prove a 30% saving on support costs of thin clients over thick.”
These year-on-year savings have been reinvested in a disaster recovery site – something that is even more important now the infrastructure is centralised.