IT chargeback can be intimidating for the inexperienced, but it is one of the most useful tools in the fight against...
sprawl and excessive resource consumption in a private cloud infrastructure. Enterprises must keep track of usage so that resources are consumed fairly and are aligned with organizational priorities.
Implementing IT chargeback can be difficult for an organization with no history of accounting for resource consumption, because it requires justifying and inventorying each server and workload as it moves to the cloud. The upside to that process is it reduces waste, curtails sprawl and puts pressure on application and systems administrators to rightsize their virtual machines.
Ways to calculate IT chargeback
IT can use chargeback for computing and storage resources used, as well as for staff time. It is very common for chargeback to be done on a per-VM and per-gigabyte level, accounting for CPU and memory use along with storage.
More on IT chargeback
Which method is better for you: Chargeback or showback?
Using the chargeback system to demonstrate virtualization benefits
Mature IT organizations sometimes bill on a per-CPU cycle or per-byte-transferred level because mainframe chargeback used to be that way. But chargeback is overhead, and does not directly advance business goals. You should keep your chargeback as simple as possible while trying to reduce waste and increase accountability.
Many organizations start with chargeback billing as a flat fee and calculate the costs based on average sizes. For example, you might decide that a single VM costs $500 and just apply that to all VMs. It's a simple way to calculate things and very budget-friendly for people involved with budget planning.
In this practice, smaller customers subsidize the larger ones, and it does not encourage rightsizing of VMs or resources used. In that case, you might use a tiered system, which adds complexity to billing and budgeting, but it also encourages good sizing behaviors, and people may see it as more fair.
IT chargeback and storage in private cloud infrastructure
Storage chargeback can be as complicated as you want it to be. Many organizations have multiple tiers of storage and different billing rates for each. As storage auto-tiering becomes the standard, billing will get even more complex because individual blocks of each server will be stored on different tiers of storage.
It is easy to tell how much storage your marketing department is using on your central file server, but how can you tell how much is on solid-state drives, 15K or SATA disks? There are complex methods of determining that, but many organizations that have moved to auto-tiering storage have simply declared it a co-op and base charges on the total cost of the array divided by the usable storage. It's easy to budget for, and the automation saves staff time.
You might also choose to include IT staff time as part of the chargeback model. It's helpful to account for staff time used as part of the private cloud infrastructure, but it's not how IT should be funded in the long term. Standardization and automation can help identify what work should be charged back. For example, some IT shops have adopted a hybrid approach in which staff funding comes mainly from the organization itself, but staff time to work on requests that fall outside of standard configurations is paid for by the project or department making the request.
Take care to make the IT chargeback process as unobtrusive and low-overhead as possible. Enterprises that don't feel they can do chargeback right away often choose "showback," where reports are generated for management that show where resources are being consumed. Showback is an excellent first step to a real chargeback model and is useful in the initial stages of private clouds to help set budgets and expectations.
Many organizations that employ showback techniques treat it almost like chargeback, in that specific projects and departments are assigned a dollar amount -- except the bill is never sent to the "customer." It is a powerful way to track and conserve resources but can be completely foreign to developers, application administrators and staffers who have never needed to justify or account for their resource use. You will want to ease staff into these new models.