One of the reasons that a lot of IT organizations are not all that comfortable with cloud computing is because they don’t have a lot of visibility into the true costs of cloud computing.
The biggest fear, of course, is paying for cloud resources that don’t wind up being used because the compute resources for the application have been over-provisioned. In an on-premise deployment, over-provisioning is a waste of fixed-cost resources. In the cloud, it’s an ongoing monthly expense that can really add up.
Pat O’Day, CTO for Bluelock, a cloud service provider focused on VMware environments, says customers should routinely have access to tools from cloud service providers that prevent this from happening. Bluelock, for example, gives customers access to a set of portfolio tools that allow them to easily figure out what the cost of running their application on a Bluelock cloud is going to be.
For example, algorithms are used to track current patterns of consumption and project a trajectory to the end of the month in order to provide an estimate of the pending bill for the month. Those reports also identify what issues might have caused that increase. In addition, the monitoring tool allows customers on their own to access a self-service portal based on VMware's vCloud Director software to cut back on resources in the Bluelock cloud or move CPU, RAM or storage around to other applications as necessary.
But longer term, O’Day says it’s obvious that customers will want to compare that information across different clouds. That should lead to the development of application programming interfaces (APIs), says O’Day, that roll up that information into third-party analytics tools that can be applied across multiple cloud service providers.
That’s going to be important because increasingly IT organizations are going to find themselves working with different classes of application workloads that will behave differently on different cloud platforms. In fact, IT organizations are going to need to have unprecedented levels of visibility into the cost of running IT in the cloud, not just on a yearly or quarterly basis, but on a daily, sometimes even hourly basis.
Given all that, it’s pretty clear that the next major challenge in the cloud will be figuring out just what the real cost of running any application workload at any given time truly is.