Chargeback Challenges in State and Federal Government

In Corporate or Federal IT, customers are looking for more flexible datacenter solutions. End user organizations want to lower thier costs and only “pay for what they use” and at the same time have “unlimited capacity to provision what they need”.  How realistic is this?

IT would like to provide these services through “private clouds”. However, there is a cost involved in hosting this infrastrcuture. Tools to charge customers for what they use (metered) are available, but more importantly, how do you calculate and who pays for the unused capacity that must be maintained?

When you get to the granular level of metered chargeback in the x86 space, how can any customer accurately predict the level of usage out far enough to establish a “fair” pricing structure?

In the Provider space, like Amazon, Terremark, or even in the Unisys Secure Cloud offerings, the goal is to achieve a profit…so it’s a cost plus margin, where reducing costs and increaseing revenue is the goal. Transpose that to the Corporate or Public sector, where profits are seen as being just as bad as shortfalls, it becomes very difficult.

In a shared-service model, where more control is placed into the hands of the “user” and the “business” to provision or de-provision services, and a financial incenentive is included, IT is in a bind. Needing the capacity to serve thier customers when the demand rises requries an over-building of capacity to handle the peaks. But when the demand drops (perhaps on a monthly or quarterly basis…or someday even shorter), the cost to maintain the environment may NOT drop.

But with a finite set of customers, THEY are the ones that must carry the burden of the unused capacity. The loop here is that those same customers who are expecting to pay less because they are only paying for the services that they use are also the only revenue stream that they IT Shared Services team has to cover the costs of the infrastructure…even when they are not using it.

Frameworks can be developed to include the cost of unused capacity into the charges that are assessed to the customers when the ARE using the system, however, without an accurate forecast, there’s no fair way to determin if those “unused capacity charges” that are being added to the usage charges will actually cover the infrastructure costs….or will they be over charging the customers and have a surplus (bad) at the end of the year.

How can we work around this?

  1. Require super-accurate forecasting, and place all of the private cloud flexibility into the forecasting stage?
  2. Implement technical frameworks that can more accurately and “swiftly” increase AND decrease infrastructure capacity and costs at the same time?
  3. Build in the surplus and apply a refund at the end of the year?
  4. How about some creative financing / purchasing / procurement by partnering with your hardware and software suppliers and vendors?

Looking for other suggestions as this becomes a more important facet of private cloud consulting.