Columnist Mike Vizard pointed out a fundamental flaw in today’s IT models in his recent article The Need for Transparency in IT. “When people don’t realize how much something truly costs, they generally don’t appreciate it.”
This results in overconsumption of tech resources and unpredictability in budgeting. If intelligent workload management (IWM) models are built to embrace the concept of visibility, however, we can break free from these challenges.
Sure, on the surface it may seem to business units that adopt cloud computing that they have access to infinite resources. But if those requesting resources have visibility into the costs of the VMs, they quickly realize these are not free. Having this knowledge aids managers in making more informed decisions. If, for example, I can see what the cost is of the performance level I desire, I can quickly determine whether it’s worth it to my business group to pay for that level of service, or perhaps I may choose lower tier of service. If I find that I need more, I simply request an upgrade, knowing the real costs associated with it.
Showback vs. chargeback
Some people believe that charging back business units for IT consumption creates unnecessary friction between IT and internal business units. Regardless of whether a business employs knowledge about actual IT costs for chargeback (i.e., charging business units for their own usage), or simply for “showback” (i.e., identifying costs to each business unit), this knowledge is imperative for reining in costs.
A return to chargeback
If you look back at the mainframe days, chargeback was prevalent because businesses had shared resources and needed to recoup their costs. Then with the shift to distributed computing models, individuals departments would purchase their own applications and servers. Each department might have paid a nominal fee for cooling and power, but they each owned their own servers and other associated assets. Now, with virtualization and cloud computing, we’re moving back to that shared resource model where more people share hardware and it’s treated as one big pool. However, that requires the right approach for chargeback.
Chargeback models
There are two primary types of chargeback: allocation-based and utilization-based.
With an allocation-based model, IT allocates a certain amount of compute resources, and charges the business unit the same price, no matter how much of those resources they consume. They may offer different costs tiers, similar to what we see with our cell phone minute plans. (For example, a 500-minute plan has one cost while a 1,000-minute plan has another).
With a utilization-based model, IT essentially “meters” service usage and bills the business unit based on actual consumption. However, organizations are struggling to determine how to actually allocate costs in virtual environments.
Initially, cloud management tools will build in the ability for allocated-based chargeback. As business users before more accustomed to the pay for usage scenario, we’re likely to see this evolve toward utilization-based chargeback models.
What model do you think works best for your business?
Posted
Aug 11 2010, 10:06 AM
by
JohnStetic