Is this a bad example of Private Cloud, in the end you can easily argue this is what was being attempted in this CRN Article? Does this illustrate that the larger and more complex you make an environment the more expensive it becomes to manage? Or do you just want to put it down to mis-management and/or government bureaucracy?
If you note the article, a major issue is the heavy customisation and the demands of many varied departments - sounds like many corporates to me.
Or is it the need for public service organisations to employ layers of consultants to cover their backs?
My exposure to government IT technical resources has mostly been very positive, often they are some of the best technically I have met.
So my questions are:
Is this because of the un avoidable layers of bureaucracy?
Is there a level of complexity where you get a diminishing return on centralised infrastrucuture?
I have always promoted the benefits of centralisation but would be keen to hear any comments on this.
My experience with ERP systems leads me to believe that any attempt to centralise multiple business entities, whether they be govt or private sector, onto a single ERP platform is going to fail. It's a flawed concept.
ReplyDeleteWhile I agree with Phil on the value of centralisation mine is a qualified agreement. It is only of value where there is sufficient common ground to justify it.
In my company, Circatec, we break large developments into multiple small ones and when developing software we identify the functional areas that are going to be entity specific, designing those functions to be parameter driven or configurable. This is the only way a package software developer can provide agility and flexibility without extensive modifications.
There are many papers written on the demon of complexity and the point at which size and complexity GUARANTEES FAILURE. This Oracle project appears to be one of those.
As a footnote the management of modifications is critical. One of the core underlying problems is that the software vendor's implementation team think the users are the client - they aren't the corporation or govt entity is. Responsibility rests with both the vendor and the client to insist that no modificationsa be made without challenging why. I'm not talking about formal cost benefit analysis, that comes later, but rather questioning whether this change is being made to accomodate an unwillingness on the part of the user(s) to change process to suit the software or because of a genuine shortcoming in the software functionality. Do this and usually well over half the modification requests don't even get on the issues register!