From Virtualization to IT-as-a-Service: Bridging the gap takes new technology and processes

So you’ve virtualized nearly 100 percent of your organization’s IT infrastructure which means you’re ready to transform your IT operations to an IT-as-a-Service model, right? Well, not really. There is another transition you must make to bridge the gap between a highly virtualized environment and ITaaS.

EMC IT defines IT-as-a-Service (ITaaS) as Optimizing IT Production for Business Consumption. Many of our previous blogs have discussed the need to transform IT’s business and operating model to operate more like a service provider. That requires business model evolution and shifts in the skill of the people in the IT organization. These topics address the business consumption aspect of ITaaS. Optimizing IT production implies that there are additional opportunities to increase the utilization of your IT infrastructure beyond virtualization.

First, despite high percentages of virtualization, many IT organizations are still managing discrete groups of virtual machines that aren’t pooled together in a cloud ecosystem. The groups of virtual machines or clusters of servers running virtualization software may or may not be optimally configured and contain excess capacity needed to meet peak loads or anticipated growth.

So the next step in your journey to the cloud (and ITaaS) is addressing how to move workloads between those virtualized environments to fully balance utilization across your infrastructure. This migration from a clustered infrastructure to a cloud infrastructure allows you to leverage your virtualized environment further still to gain more savings and flexibility.

Not only is achieving that balance key to maximizing IT operational efficiency, it is also essential to ensure that ITaaS will work accordingly. After all, ITaaS requires that IT operate like a service provider, tailoring its service portfolio and service performance to meet the demands of its customers.

Second, if an ITaaS service offers clients a certain amount of capacity, your organization needs to make sure that it is readily available. You need to track consumption and manage the supply chain – how you fill the demand – for a consistent user experience. Third, you need to build a level of trust with those who consume your services and be cost competitive by maximizing the efficient use of your infrastructure without pockets of unused capacity.  Optimizing IT production begins with changing your measure of data center success from focusing on percentage virtualized to percentage of capacity utilized. Gauging consumption of memory, CPU, network and storage utilization across your data centers will measure overall efficiency of your new cloud infrastructure. The ability to aggregate infrastructure and move workloads to where the capacity is will allow further increases in utilization.

Fourth, to operate a global cloud infrastructure, you must become more disciplined in key processes like capacity management as well as take a fresh look at performance and change management. An infrastructure that is designed to be self-healing and dynamic challenges the notion of capturing a known state. It requires new processes and monitoring tools that can track the ever-changing relationship between applications and their underlying infrastructure.

Fifth, bridging the gap to ITaaS also requires next-generation applications that allow your operation to scale out and be more flexible.  Many traditional applications, while virtualized, still operate in a physical world and don’t effectively take advantage of cloud infrastructure. The ability for an application to spin up new virtual servers to increase capacity and performance need to be built into the application platform layer.  Offering and implementing platform standards on top of the cloud infrastructure will further increase your chances of achieving success.

At EMC, we are in the middle of making this transition, with capacity and performance management tools in place for parts of our environment and more traditional operations still in place in others. We feel that there is opportunity to drive higher degrees of efficiency in our environment and are implementing new metrics to gauge our success.

The ability to have an optimized cloud infrastructure will be different for every business and we are working to figure out our optimal state based on projected growth and demand. Capacity planning and demand forecasting are more critical than ever as you move to ITaaS. For instance, we need to determine how long it takes us to add infrastructure capacity and how that factors into our optimal utilization level. We are also still in the process of defining our ITaaS services and how they will be bundled and priced. Our utilization level will directly affect pricing.

In fact, utilization management of an IT infrastructure was previously important, now it is an absolute necessity with ITaaS. Users can now order a service on-demand, without the cumbersome governance requirements and, frankly, the delay tactics that IT previously used to slow down demand to allow for reaction time to add capacity. This immediacy also eliminates the six-week lead time I had before, so standardizing our infrastructure, utilizing new converge infrastructure capabilities like Vblocks, and partnering with service providers to build a hybrid cloud will enhance our ability to lower costs while providing the agility that our business applications require.

Think of it as treating your IT platform as a utility. You are using new tools to connect virtual islands together to create a true utility model. Just like an electrical power grid, you need to manage all those resources as one ecosystem, shifting workloads between various locations as needed to gain the most efficiency and partnering with other energy providers to deal with unexpected spikes in demand. So if one area is getting brown outs and one has excess power, an automated switching system can make adjustments before services are interrupted.

You have people providing services on top of grid, as well as people optimizing the grid. It’s the technology and the operations coming together to create a cloud infrastructure. Once you have achieved that, you are ready for ITaaS.

About the Author: Mike Norris