Fast forward four years. It’s 2009. Remember that you’re Sara, the CIO of a mid-to-large retail company. You have many accomplishments that the business leadership team (and Board of Directors) has recognized: a streamlined supply-chain management system that reduces costs and out-of-stock conditions in stores, expanded your ecommerce presence to Europe, and improved the service levels of your major mission critical applications substantially as well as most of your “tier 2” applications, to name a few.

Internally, you recognized that to have accomplished these goals for the business, changes within your IT group had to be made:

• You examined the IT processes that historically were either poorly performing or not defined and created programs and training for staff to implement IT management processes based on ITIL
• You virtualized most of your data center except for your mission critical applications and large database systems
• You implemented an IT service management management function that resulted in a robust but practical and usable configuration management database (CMDB)
• You created an SLA enforcement mechanism for your critical applications that has the effect of notifying your systems operations center when mission critical application components fail

As a result, sales are up, costs are noticeably lower, and service quality is significantly improved. Life is sunny.

Until the clouds come (sorry, I just had to put that in!)

The implementation of your online presence in Europe has increased the traffic to your site infrastructure by 20%. Moreover, this move results in many holiday specialty items. While your ecommerce infrastructure (web site, fulfillment, inventory, and supply chain) has generally kept up with demand, it has struggled with performance and scaling issues during the holiday season. Also, the annual operational cost of the data center is quite high due to the need to support seasonal demands. This new expansion is stressing your servers even more at peak and, though you’ve grown capacity to meet the demand, your CFO is putting pressure on you to cut operational costs. As you look at your utilization reports in February and March, noting the significantly low traffic that your servers are getting, you know a change needs to be made.

To solve the performance issues and the overall seasonality challenge that’s causing significant annual operating costs, you form a task force to investigate the implementation of a private cloud environment. While you believe this will potentially solve the performance and scale issues, you’ve come up with more questions than answers. On the plus side, cloud technologies promise significant scalability benefits (elasticity) to be able to more effectively handle the seasonality concerns, including scaling back services when demand is low. This plan could also result in improvements in your ability to deliver services on demand by deploying ready-made virtual machines for web and application servers. However, many questions persist:

• What model do I choose: PaaS, IaaS, or some combination?

• If I choose infrastructure as a service, how do I guarantee the quality of the resulting applications?
• If I go with platform as a service, what standard services do I put in place that will promote agility without sacrificing quality? How do I enable flexibility and not compromise compliance?

• Will it be cost-effective? I have a charge-back model today for the services I provide to various lines of business. How will I need to redefine that model so I can keep costs in line?
• How will it affect my service delivery? Right now now I have a pretty well-defined service management system that helps me keep my critical apps up. Will there be significant VM sprawl as a result of the elasticity I should be getting? If so, how will that affect my ability to manage this sprawl for my critical applications. Heck, I just finished getting a handle on my tier 2 apps. Will I lose control of those?

If you haven’t gathered by now, there are themes here that the War on Cost team has been exploring and discussing for some time across a variety of scenarios. They are the enduring business value pillars of reducing costs, improving agility, improving quality of service, improving governance and compliance, and managing risk. All of these value dimensions are in play with the introduction of a cloud paradigm. In fact, they have the potential to be amplified:

• Agility – Possibly the value pillar with the highest potential amplitude. The cloud paradigm (whether public, private, or something in between) has the potential for extreme agility. Automatically provisioning development, test, and production environments in a matter of minutes or hours instead of days or weeks has enormous potential
• Costs – Costs have the potential to be dramatically reduced by providing a commoditized set of services that can be automatically commissioned and decommissioned at will. Also, server utilization can be significantly improved through intelligent load balancing as demands for the service change, particularly due to seasonal demands from stores and online traffic
• Quality of Service – QoS in all its forms (reliability, scalability, performance, availability, etc.) can be greatly improved by intelligently and dynamically load balancing workloads across (potentially large) arrays of servers running as VMs on a farm of servers that are well-utilized
• Governance, Compliance, and Risk Management (GRC) – These three are related values and the cloud offers huge potential benefits here too. As services are defined (whether its infrastructure, platform, or software), they are standardized (one of the main tenets of cloud computing). Standardization has the natural implication of introducing governance and compliance regulations, which, in turn, has a direct net benefit to managing risk

But, and it’s a really big but, what about complexity?

You’re thinking about the sheer number of VMs that can get spun up in your data center. Where once you had a hundred or so servers, you now have somewhere between five hundred and a thousand VMs (you’re not really sure). Will implementing cloud exacerbate that VM growth? How can I contain it? How much should I contain it?

So, therein lies the challenge before you: if you’re going to move to a private cloud because you and your business leadership team see the potential for huge benefits, how will you manage the complexity introduced by this cloud to be able to maintain and enhance your IT team’s ability to deliver for the business?

What do you think? Do these ideas ring true with you? Are there other major considerations to weigh? Let us know what your experience has been. Have you answered these questions for your organization? Don’t be shy. We want to hear from you.

All the best,

Erik, Strategist, War on Cost