The New Zealand Defence Force is embarking on a project to consolidate server infrastructure spread over 10 camps and bases to two main data centres.
The aim is to run thin client software on the NZDF’s 7500 workstations (desktops and laptops) linked to either of the data centres. A major challenge is integrating the IT of three distinct services.
They were interviewed about the Defence project by reporter Andrea Malcolm.
What is your definition of server-based computing?
Warwick Sullivan (pictured left): The provision of application services via a server where the connection between you and server is related to keystrokes, mouse movements and display changes.
What is the scope of the project?
Ron Hooton: We’ve been looking at our environment and the need for the organisation to act as one from a CIS [communications information systems] point of view. We see server-based computing as a great enabler. The scope is three-fold. Technology changes — we will consolidate what is done in 10 locations to two major data centres in Auckland and Wellington. These will be under one administration and will operate in duplicate. We’ll do half the work in one and half in the other and replicate across the two. With that comes some storage and server consolidation. The result will be a less complex and more standardised environment with consolidated applications.
Organisational changes — we will create a single infrastructure support team and helpdesk. We’ll still need people at camps and bases to look after central hardware but there will be changes to reporting lines and changes of role in some cases. There will be fairly significant organisational change and we have HR working on it.
Financial effects — server-based computing will reduce people and desktop costs. It is a mode of lowering the cost regime for delivering 7500 PCs and laptops.
How does it fit in with the overall IT strategy for the NZDF?
Hooton: We wouldn’t be doing it if it didn’t. We are integrating CIS functions across a whole range of things.
Are you consolidating all your servers?
Sullivan: Not completely. We’re not going as far as buying a large mainframe and compartmentalising it. Exchange messaging, domain controllers, DNS, and SQL servers will be consolidated along with non-core applications. Line of business applications won’t be consolidated. For example. the air logistics information system will remain on its Unix box. It’s a bit of a journey and we’ll do more rationalisation as we go further on.
What are the benefits you expect from this architecture?
Hooton: Speed to market. We can get applications and changes to applications rolled out very quickly because we’ll only be making changes at two data centres. Better security because data and application logic will be held in the data centres. Better availability and reliability. More up-time by spreading the load across multiple servers. There will be no single point of failure. If a desktop fails you can go to machine next door and use it without reconfiguration issues. This will improve business continuity and disaster recovery.
What is the current IT infrastructure?
Sullivan: We have a standard distributed environment although we have a centralised set of applications at our computing services bureau in Wellington. Camps and bases have their own LAN services including domain controllers and services in terms of applications and resource servers. Each has an enterprise services node, which has corporate resource services, message transfer agents, and domain controllers, which connect back to the data centres via a WAN. Each camp base holds its own data although we do backups across the network.
Do you need to buy new hardware/software to do this?
Hooton: This will drive the need for a significant number of servers because all processing will be done on the servers not on the PCs. We will continue to use existing PCs and laptops. Each desktop will have a Citrix 32-bit ICA client but we haven’t determined what operating system yet. We could retain the current Windows or get something else. We’re aiming for a very simple desktop environment — just the OS and the Citrix ICA client.
For the servers we will use the terminal server functionality of Windows 2000 and Citrix MetaFrame XPE. We will continue to connect to Unix-based applications through terminal emulation and a Windows-based programming application interface. For example, SAP, which handles a large part of our logistics operations and financials, is hosted on Unix.
We are rationalising applications. For example, we don’t need 14 different scanning systems, and we’re also putting applications on to a single server. There are quite a number of boxes supporting single SQL applications and we will consolidate to a smaller number of boxes. A practical benefit of this is that we’ll be able to get everyone onto the same versions of software; for example, the Office suite.
We are also doing a major upgrade of our communications and network services. Telecom Advanced Business Solutions is handling the technology for both this and the server-based computing project.
What is the challenge of doing this across the three services plus headquarters joint forces?
Hooton: The three services have distinct cultures. However, we have done a lot of talking and consultation with each one. The critical success factor and key challenge is to make sure that we keep the needs of the end users in mind. In some cases these needs will be unique to a service. We also have to make sure we keep their wants and needs separate. People have ideals and visions of what they may wish but we need to focus on the actual needs of the organisation.
We have a new order: three services, one force. We operate under one chief of defence. We have to make sure we keep in clear focus the needs of the single services but recognise we need to integrate.
What is more difficult — the technology changes or the organisational changes?
Sullivan: Organisational changes are always more challenging. You’re dealing with people and you invest more time in people than in technology.
Hooton: We’re constantly asking what does this organisation need to operate? What’s logical? We have to separate the emotion from the logic. I’ve been through changes many times and often emotion gets in the way. We have to ensure that changes are not unexpected. We don’t see that the technology component is a major risk. We’re not using bleeding edge stuff. There are a lot of precedents within New Zealand for server-based computing and use of Citrix technology.
How is the project broken up and what phase are you currently engaged in?
Hooton: First we did a six-week feasibility study to determine whether it would fly and the answer was yes. Then a six-month long discovery phase to find out true costs, savings, capacity and requirements. At the end of that we were still on the positive side of the ledger so we decided to proceed. Next is the design phase which, from a technology perspective, is defining the architecture. From an organisational aspect we have already launched into preparing people for change through formal consultation and communication processes. Around about the middle of the year we will go into a substantial pilot. We haven’t decided how many people will take part or how long it will last but it will be a limited number of users right across the organisation.
The final phase will be a rollout which will progressively bring more and more desktops into the server-based environment.
What steps are you taking to minimise risk?
Hooton: We are very risk-averse. We have a lot of internal governance and the chief of internal audits is on the steering committee. We have wrapped our governance structure around the entire project and from time to time we make sure we have fresh eyes and ears on the project to make sure we don’t miss anything.