A virtual machine (VM) that is running in a public cloud, which augments or extends traditional on-premise VM computing roles, is known as an Infrastructure as a Service (IaaS) class VM. Individual business cases are the economic drivers that move IaaS class VMs into the public cloud. There are many motivations, such as seeking high availability, global access, and elasticity where large public cloud providers can uniquely deliver value.
For example, a public cloud provider's content delivery network (CDN) might be able to synchronize IaaS document libraries across regions much simpler and cheaper than previously possible. Other IaaS implementations may not have as clear a pay-back or return on investment (ROI), so the question arises; How much is an IaaS VM worth?
Apples to apples, then some math
Various public cloud and application providers offer business case calculators, but each is probably somewhat slanted towards the technology of the calculator vendor. A good place to start when calculating ROI on cloud IaaS migrations is to know what it would cost to refresh existing infrastructure on new but similarly configured hardware and software platforms.
Let's say you have an application that features two load-balanced servers, that if deployed on two industry-standard servers (ISS) with 2-GB RAM, a dual-core CPU, a local hard disk, and a three-year warranty represents at least a $5,000 capital expenditure. Operating system and management licenses can increase this by over $3,000. A conservative $500 annual per server power and facilities charge means it would cost about $11,000 to host the two-server application for three years on conventional infrastructure (excluding networking).
So when looking to migrate that workload, anything less than $11,000 total 3-year expenditure-for equivalent or superior service--will be attractive. This amounts to about $306 per month using conventional infrastructure, or about $153 per month per server. You can go shopping for public cloud IaaS VM providers knowing what the relative costs are to do-it-yourself.
Compare rates, but consider the big picture
At Amazon Web Services (AWS), a Medium Utilization Reserved Small Instance VM on three-year term has a total 3-year cost of $1,261, or $36 per month per server. Microsoft's Windows Azure Virtual Machines (Preview) service, Small Instance VM is $57 per month with no multi-year obligation. Rackspace Cloud Servers lists a similar computer with a $117 per month cost. Obviously all three public cloud options are cheaper ($36, $57, and $117 per month) over the three-year period than doing it yourself, with traditional on-premise servers ($153 per month).
At first glance, AWS has the far cheaper option, but here, the comparisons across public cloud providers become more nuanced. For example, the AWS price requires a three year commitment, and there are higher prices for shorter terms. The Azure offering uniquely includes geo-replicated hard disks in the base price. Another difference is the Rackspace offering includes bandwidth while the other providers charge extra for networking. Each provider's blend of services adds complexity in the decision making, but provider diversity can be leveraged to use the highest-value features of each public cloud provider. (See Thoran Rodrigues' comparison of IaaS providers for more details on the considerations for each vendor.)
Rarely will the base monthly run rate of the IaaS VM be the decision maker. Do the math to calculate what it would cost to do it in-house; then verify that the public cloud solutions cost less out of the gate. If your idea passes the first test, move to analyzing the details of the cloud provider offerings. Here are some valid considerations for selecting a public cloud provider, after quantifying the ROI in your particular business case:
- Your near-term public cloud plans are part of an overarching cloud strategy, which should be the primary decision maker when it comes to cloud providers.
- A particular provider may or may not permit virtual private network (VPN) connections between the public cloud and either an on-premise network, or another public or private cloud.
- Your operating system or vendor application stack has tools that permit easy moving of workloads into a particular cloud provider.
- Some other factor, such as the cloud provider's CDN or networking charges are particularly suited to the business needs of the workloads.
- You are staging workloads across public cloud providers on purpose; effectively removing a particular public cloud as a single point of failure for your enterprise application.
The example in this article was for two modest capacity application computers. You situation might involve complex workloads such as database and transaction processing that require more robust, and more expensive IaaS-class VMs. Regardless of the VM category and selected public cloud service provider, small scale pilot projects to gather precise cost/performance metrics are a must to validate cost and ROI models in public cloud projects.
John Joyner, MCSE, CMSP, MVP Cloud and Datacenter Management, is senior architect at ClearPointe, a cloud provider of systems management services. He is co-author of the "System Center Operations Manager: Unleashed" book series from Sams Publishing, and is developing cloud-based management solutions based on the Microsoft System Center 2012 suite. John is a retired U.S. Navy Lt. Commander 'Surface Warfare Officer', with the subspeciality 'Computer Scientist, Proven'. His tours of duty included Chief of Network Operations for NATO's southern region and network administrator aboard the aircraft carrier USS CARL VINSON (CVN-70).