Cheaper to Buy vs Cheaper to Own: Hypervisor Economics Pt 2

From part 1, we found out that virtual machines (VMs) or hypervisors have become the building foundation of data centres these days. But it isn’t economically viable to strengthen the data centre’s performance and robustness by adding more and more virtual machines. 

Hitachi Data Storage (HDS) chief economist, David Merrill has observed, “It wouldn’t be uncommon for 40-percent of your IT budget to go towards storage-related costs. “ And it doesn’t seem to be said often enough in the industry, but from part one of this article it becomes clear that reducing total cost of ownership (TCO) of storage and VMs means having to address its costs of maintenance as well.


But that isn’t all, and in the ten years that HDS’ has been officially documenting methodologies, they have identified 24 types of costs that can influence the overall total of ownership of hypervisors.

HDS proposes that storage can ‘synergise’ with server to increase performance. A ‘Hypervisor Economics’ whitepaper by Merrill stated, “The combination of server and storage virtualisation technologies can make a compounded economic impact to IT environments that have not been previously available.”


The whitepaper also outlines the steps to hypervisor economics:

  • Identify your VM costs
  • Measure and track VM unit costs
  • Take specific actions to reduce VM unit costs

Identify your costs

These are divided into the Purchase Cost, Operational or Run-rate costs, Migration, Movement-related expenses and Risk-related costs. Of the total 24, there are 8 to 10 used most frequently for cost models such as hardware, operating system, maintenance, labour, environmental, network, migration data protection and outage.


Measure and Track VM unit costs

 There has to be a baseline measurement. An annual TCO measurement of $/VM is useful to show progress (or not) and set priorities and expectations moving forward.   For hypervisors, costs perI/O or TCO/VM/year have been found to be the most relevant.


Measurement and re-measurement will have to happen if there is to be continuous improvement. Needless to say, different organisations will choose different costs when defining economic metrics… comparing your TCO to other companies simply wouldn’t be giving an accurate picture!


Take specific actions

Of course cost reduction tactics have to be tangible and actionable. These actionable cost-reduction plans have to be aligned to the organisation’s roadmap and IT strategies; results could take as long as a year and indeed it is actually more valuable to measure TCO costs each year.  Something that HDS calls ‘levers’ are available to help reduce costs and there are two broad kinds: technology and organisation.


Technology levers have to do with hypervisor selection and things to consider include scalability, interoperability, operational efficiencies, ecosystem flexibility and hardware.


Storage and its architecture also play an important role and capacity, VM movement and overall performance should be figured into the selection process.


But making technology decisions and implementing it almost always will impact the organisation as well: cross-training server and storage teams so there can be fewer divisions of labour. Other things include implementing a chargeback system, using a services catalogue to communicate VM offerings to the requesting organisation and ensuring that procurement maximises enterprise agility and management.

Leave a Reply

Please Login to comment
Notify of