Can financing and management models survive IT change?
A soup of acronyms under pressure
Comment Whether or not you believe that IT is about to experience a revolution, it is clear that a number of the fundamentals of how the IT infrastructure is built are changing.
Many technological and operational developments are changing the way applications are developed and deployed. Everything from innovations in how applications are designed and coded, through to the methods used to deploy running applications are transforming at a pace rarely experienced in the world of computing.
On the application creation side it is now the case that the adoption of SOA principles allows functional components to be developed rather than an entire application. This architecture thereby potentially allows individual functions to be utilised by multiple systems rather than simply by the application for which they are coded as is the case in traditional application ‘silos’.
Another major factor influencing the deployment of IT systems is “virtualisation”. When deployed on virtual servers, individual applications, along with all the associated middleware and Databases, are no longer fixed to run on specific server hardware. This approach allows great flexibility to be incorporated into the running of systems as well as providing the opportunity for multiple applications to be run on a single server.
With so much potential in the technologies available to build and run IT applications, a major question then shifts to the support side of the stack, namely the way IT is financed and operated. In essence both SOA and virtualisation make it possible for systems to be built and run much more flexibly than in the past. They also force the concept of server, storage and application resource sharing as something to be considered on both operational and financial grounds.
But while operationally and even economically such developments make sense, there is evidence that they may not fit so comfortably with traditional IT budgeting models. The idea of sharing or pooling resources can also stir up internal political hackles. In the light of such high profile developments, can traditional budget models survive the revolution?
Until now much IT spending has been predicated on building/acquiring individual applications and their supporting server hardware, storage and middleware software. These costs have then been coupled with some estimation of the operational overheads to produce an IT budget. Frequently such a budget may have been met by the end user of the application / service being delivered.
This approach to budgeting for IT systems can make it difficult for shared systems or virtual server pools to be utilised for IT service delivery. Especially in scenarios where existing platforms are to be reused to run not just the original service but possibly to host other applications, especially if they are run for the benefit for another department.
Then there is the question of how to pay for a virtualised data centre where any server can be used to host any service and where storage pools exist housing all data. Is this a complete nirvana or a budget nightmare if users demand total transparency on costs and who is paying for what?
If an organisation can untangle the internal politics that comes as part and parcel of IT budget allocation, or the “cost blame game” as it is sometimes termed, then it will be in a much better position to ensure that IT resources are managed in a fashion to more accurately reflect business needs. Clearly there will be a requirement for more sophisticated monitoring tools as well as a demand to put in place clear reporting systems to ensure performance indicators are reflected.
Even more important will be the need to ensure that business needs drive resource allocation rather than leaving everything down to just the IT administrators. For the IT staff to start making such value judgements without the active participation of business managers will result in culpability wars without end. Such virtualised data centres and transparent IT resource pooling may well prove to be the catalyst for charge back models to take a much more prominent place in the IT budgeting processes.
Such changes to IT budgeting are needed for IT to operate as efficiently as possible. Should “Cloud” services become part of the equation then an even more dramatic change to IT budgeting will be required, especially should there be interest in utilising “pay per use” models for long term IT service delivery. Without significant changes to business/IT budget relationships it will be very difficult to optimise IT service delivery or to use external suppliers except in niche scenarios.
Enterprise IT budgets have existed in their current forms for a long time with few significant changes. It is very likely that the changes to IT service building and service delivery will create substantial pressure for IT budget change. Such changes will not be straightforward, even with considerable dialogue exchanges between business and IT managers. Indeed, they could well be fraught with difficulty but they are needed to ensure that business gets the IT services it needs and that IT be seen as the important supplier it is. Without such budgetary changes IT evolution will be slowed and, ultimately, business advantage will be reduced.
Tsk... tsk... tsk...
May I suggest for future articles if you wish to use a less common acronym, the first instance is followed by what it means in brackets?
I looked up SOA: Soldiers of Allah? State of Alert? Signs of Aging? Save Our Asian? (tv plot device, apparently) State of-the Art? I ended up at http://www.soa.org/ with the banner "Risk is Opportunity" (bleugh). Soa is even a Scottish island!
I finally found a reference to "Service-oriented architecture" which seems the most applicable candidate given the topic. Would it have killed you to have said that in the first place?
Smiley 'cos, don't take this too seriously. Just for future reference, okay?
Forgive my ignorance but...
...don't the Major Players in Virtualisation have the ability to log how many cycles/IOPS/etc. that a VM consumes? Then you can use that for billing, internally or otherwise? I honestly don’t know, and now I’m curious.
My IT department runs two separate budgets: Infrastructure and Apps. Since we are completely virtualised, one can grow independent of the other. When we upgrade our servers, we get way more capacity than we need, and the apps eventually grow in number or requirements to fit. Then we increase capacity again.
The ability to “ding” the department with the greediest app requirements, and back it up with solid numbers would be awesome.
Spot on - thanks for this article
This article is spot on.
I think that the business managers must continue to develop their understanding of IT and use it to be able to successfully take the transition between a business need to business project to ongoing IT support and maintenance. This ability is so far rare but ithat is changing.
On the subject of complex and often confusing budgets, at a very low level, I've seen this fighting over allocation of cost taking place. We recently went through a project whereby printing was rationlized -> rather than having a printer of everyone's desk MFD's were put into place and different departments charged for their individual printing including toner, paper, ink etc. The amount of disagreement of the proporation of the printing budget owed, even after a login code was used to allocate the printing to specific department, was huge and still goes on.
What's it going to be like when 7 different businesses occupy 7 different virtual servers in a data centre? How do costs get allocated then?