Organizations are rapidly consolidating multiple physical servers to virtual environments. There is no debating the cost savings and operational benefits of virtualization. The issues that are now arising as virtualization continues to proliferate are rooted in management and how organizations can gain cost visibility and recover the costs of their virtualization projects. As a result, corporate finance and IT functionalities are moving ever closer together.
In many organizations, the IT department has always been and will continue to be a cost center and an ongoing expense on the balance sheet. Most IT departments simply cannot justify making massive investments in new virtualized data centers without recovering costs from its business units - a reason why chargeback is one of the hottest virtualization topics today.
What organizations quickly realize is that virtualization projects are not inexpensive endeavors. In fact, they can be quite costly, including new server hardware, software, network storage systems, people, processes, and training. However, because of increased IT efficiencies, the ease of adding new virtual machines, and the fact that departments no longer need to purchase their own physical servers, there is a perception that virtualization does not cost them anything and adding new virtual machines is free. Adding a new virtual machine is not free, but how does the IT group convey that to the business units it serves?
Based on process savings, physical space reduction, hardware consolidation, and energy efficiencies, a server virtualization project can produce a manageable return on investment (ROI) if the value of the project can be correctly measured. To gain cost visibility and implement a proper chargeback process in a virtual server environment, IT must be able to accurately determine which business groups are using which shared resources (such as CPU, memory, storage and network) and be able to associate a dollar value to that usage. Driving to a measurable ROI is the success factor for IT.
Chargeback Is Changing
The concept of chargeback is not new. It was pioneered in the "old" mainframe computing data center when organizations were spending millions of dollars on mainframe systems that were shared by multiple departments within an organization. In a client/server environment, there have also been processes in place based on traditional tools for implementing chargeback procedures in the physical data center. With virtualization, simply stating that the virtual infrastructure costs a specific dollar amount is insufficient as server resources are shared. Other major challenges are the dynamic nature of a virtual infrastructure, constant configuration changes, and resource allocation adjustments. Traditional general-ledger-like accounting approaches were not architected to work in a virtual environment.
While most organizations recognize the need for chargeback, they simply do not know where to start. There is a lot of confusion in the marketplace as traditional tool vendors attempt to market their wares as being virtualization capable. In addition, there are more and more methodologies for doing chargeback in a virtual infrastructure being rolled out - leading to even more confusion.
In a dynamic virtual environment where virtual machines automatically move (VMware VMotion) from one host to another, a physical server is no longer a relevant boundary. Resource pools and clusters determine where and how resources will be used. Thus, organizations must look at chargeback methodologies that measure resource consumption, regardless of where the host resources are drawn.
One method for chargeback for virtual data centers that is garnering interest is the fixed-plus-overage model. Most organizations have calculated what a virtual machine costs as they continue to virtualize. While this information alone is not enough to properly chargeback business groups, it is an important first step. To continually ensure the proper allocation of resources so IT can recover the associated costs, IT must be able to measure actual resource (CPU, memory, storage, and network) usage. This usage information coupled with costs to add a new virtual machine enables IT to accurately charge business groups for the actual costs of their virtual machines.
How Does the Fixed-Plus-Overage Model Work?
Every host (the physical server) represents a fixed cost that can be divided out per virtual machine (10 virtual machines divided by the cost of the server equals $X). There may be other fixed costs tied to each virtual machine as well, such as the cost of the storage attached network (SAN), licenses, support and maintenance, and environmental and space expenditures. So, the easy part is establishing the fixed costs per virtual machine. These are known commodities and most companies have a pretty good estimate of that number. The tricky part is continually measuring the unknown or the actual resource consumption of all shared resources. This is a process that pragmatically cannot be tackled manually.
When multiple business units have virtual machines deployed on a single host, resources for that host are shared. For example, the marketing, finance and HR might have virtual machines on the same host. How does the IT group properly measure the usage by department for each shared resource? It is not fair to simply divide the allocated resources by three as one department may use more resources than the others.
In the fixed-plus-overage model, IT can set the fixed price and, using the measured resource usage (MRU) approach, calculate the amount of resource overage (or underage) per department. Think of it as utility meters on a house measuring how much gas, electric, and water a residence is consuming. Each virtual machine is essentially fitted with a utility meter to automatically and accurately measure shared resource usage per virtual machine. IT can see exactly the resources being consumed by each department and more effectively chargeback for the usage. If the finance department is using more resources than allocated, it will be charged differently than the marketing and human resources departments.
More Than Just Cost Recovery
While cost recovery is usually the major driver for implementing a chargeback process, it is not the only issue that can be solved. Having cost data enables IT to gain cost visibility and properly allocate the resources required to service each business unit based on the resource consumption. This is important when further planning the virtual infrastructure as IT can make more informed decisions on additional purchases and upgrades to help optimize the infrastructure. IT also now has the fiscal information to present to corporate finance to justify these decisions.
The ultimate goal is to lower and control costs with intelligent planning to drive a quicker time to ROI. Success of a virtualization project will be based on that time to ROI. With a chargeback process in place, departments will learn there is a cost associated with adding virtual machines. This should help address the virtual machine sprawl "epidemic" that is spreading like wildfire. While virtual machines aren't free, inter-organizational departments will see significant cost and time-savings benefits from virtualized environments.