Cloud computing promises to cut costs for businesses, but does switching to the service always save money? On the surface, the potential for savings is hard to argue, with costs like $99 (£61) per employee for SaaS applications and $0.08 (£0.05) per hour for virtual servers.
But, as HP’s Joe Weinman has stated, low unit costs alone are not what make public clouds cheaper. What matters most is the resource consumption pattern of your workload.
In reality, cloud costs add up quickly, as more people use the service or more elements are consumed. For example, high ongoing consumption costs are often hidden behind low costs per employee, which can ultimately create a large total bill.
Additionally, cloud prices don’t include the operational costs to use them, and the business continues to incur costs for managing, securing, monitoring, and backing up cloud deployments.
So when does an $0.08 per server per hour pay off?
When it’s not used all the time, and proactively manages the resource consumption of applications that the cloud bill remains low.
The key to cloud economics is using cloud computing optimally, understanding the behavior of the applications and services deployed by the business.
Although blindly buying cloud services won’t lead to substantial savings, understanding the business model can move cloud economics from a cost saver to a profit maker.
While business buyers may provision cloud services themselves, it takes developers and IT administrators to activate infrastructure and platform cloud services, and more importantly, a CIO to bridge understanding between IT and the business.
At Forrester, we’ve identified three stages of economic thinking when it comes to taking advantage of cloud platforms, which each deliver increasing benefits for the business. Progressing through these stages sequentially and applying the learnings from each stage will help businesses pursue economic value from the cloud. See the chart below.
Stage 1: Scale up — cloud is a no-brainer for transient and elastic applications
The first determination every enterprise should make is which applications and services lend themselves to cloud economics.
Applications deployed for less than 12 months should be put in the cloud so the infrastructure will not induce cost once the program’s life has ended.
Other ideal candidates include transient applications like test and development projects, temporary promotional web pages, or applications supporting seasonal activities.
Any application that scales out is also a good fit for cloud platforms because it can be deployed at a small level and scaled up as traffic increases.
The more an application’s resource needs fluctuate, the better it fits to cloud models.
Stage 2: Scale down — establish use thresholds to eliminate costs
Cloud economics pays off the most when the bill is brought down to zero. This can be done by
monitoring traffic patterns to minimize consumption and lower costs.
Setting strict thresholds for resource consumption and letting automation control application scale-back will help ensure applications are only being used as needed.
Because most applications don’t take this degree of automatic resource-throttling too well, some code and configuration refinement will be necessary to make this stage of economics work for the business.
Stage 3: Profit centre — add new revenue streams based on cloud economics
When CIOs master the first two stages, they can stop seeing cloud merely as a form of cost savings, and start looking at it as a profit-maker.
By working with the business to rebuild existing services or build new ones to leverage the advantages of cloud platforms, CIOs are enabling new revenue and profit opportunities for the company.
There are several steps that must be taken in order to move from one stage of cloud economics to another. Performance analysis is the key to moving from stage 1 to 2, as CIOs must learn to turn off consumption of data centre infrastructure whenever possible.
This requires new thinking across the business, allowing performance-monitoring tools to set scale-down thresholds.
In order to reach stage three and shift cloud economics from cost-saving to profit making, IT must develop an understanding of how the business makes money, and how those actions tie back to IT spend.
Start by tracking IT investments, asking enterprise architects to map these back to business needs, highlighting models that can leverage cloud services.
CIOs can enable this process by building a new class of applications that incur minimal costs, leverage hybrid deployments, and shift core functions to new cloud models.
While historically, every IT investment has been long-term, cloud economics turns IT spend into something that can be dialed up or down to match business needs.
To take advantage of this shift, CIOs need to gain knowledge and awareness of cloud services and empower employees to experiment with its use.
To balance this experimentation, it is also important to draw the line between where to use the cloud and where it is not advised.
Only through experience with the cloud can you learn where this line should be, and how to move it over time.
James Staten is VP and Principal Analyst at Forrester Research where he contributes to Forrester’s blog for Infrastructure & Operations Professionals. More information on the topic of Cloud and Cloud Economics will be provided at Forrester’s IT Forum EMEA 2011 (June 8 – 10, Barcelona).