Running Your On-Premise Cloud Like a Business: 10 Key Steps to Take

by Chris Preimesberger

Before you can justify any on-premises cloud initiative, you’ll need to first accurately determine your existing IT costs and consumption metrics. By benchmarking versus third-party cloud vendors and outsourcing options, you’ll establish a baseline and uncover where to focus initial efforts. In addition, analytics geared toward “what-ifs” and scenario planning can help tease out additional concerns regarding security, internal costs and investments, open-source technology needs, and so on. Ultimately, you must be able to cost out internal IT products, identify areas where cloud could potentially help and then eventually measure the ROI of the transition to ensure projected cost savings become reality.

An on-premises cloud allows IT organizations to leverage existing investments in hardware, construct hybrid clouds and enable end users within the business to self-provision. Start by identifying those applications and IT services ready to move into the on-premises cloud. Consider Web services, big data applications, and dev/test or dev/ops. It is critical to define on-premises cloud services in terms the business understands. To do so, put a price on your new on-premises cloud services. Once users are paying a price, they’ll better understand how the cost of IT impacts their business functions. As a result, they’ll help you minimize waste and gain more insight into how you can improve quality and meet demand.

After you have identified, priced and associated service-level agreements (SLAs) with your IT services, you can implement true self-service IT that’s fully capable of automating the provisioning and fulfillment of IT services requested by the business. With self-service systems, IT managers have the ability to offer IT services, pricing and measurable SLAs to the business in a way that can be measured and enforced. The delivery of IT services can be customized to meet your customer requirements. At minimum, your on-premises cloud will need to be agile. An orchestration layer capable of automating standardized processes around compute resources ensures requested IT services can be provisioned quickly on-demand.

The successful transition to an on-premises cloud requires you to know, on a monthly basis, how much IT is consumed and what that amount of consumption costs. Consumption-based billing enables a fair and defensible IT chargeback or showback process by delivering a summarized “bill of IT” to your on-premises cloud customers on the cost, quality and value of the IT services each customer consumes. This also improves transparency, provides cost, quality and value levers to the business, and ultimately builds trust between IT and the business. Customers can use their “bills” to decide whether they still value the services being provided, want to upgrade to higher levels of service and options offered, or prefer to evaluate other providers to make sure they are getting the best deal.

Agility requires responsiveness to the demands of the business, and you’ll need to know how the system functions from both a supply and demand perspective. As demand increases, will you have the capacity to respond? Do you need cloud-bursting capabilities to ensure business continuity and flexibility? Homing in on the appropriate demand-management equation allows you to adapt as demand scales monthly and seasonally with consumer behavior, and it leads to the agility now required to stay competitive in today’s volatile business environment. An on-premises cloud offers flexibility, agility and automation so you can optimize utilization and reduce costs.

If demand exceeds capacity, you’ll need to enable your on-premises cloud to “burst” out to hybrid clouds to ensure business continuity and agility. Fortunately, because on-premises and public clouds share a common API, transitions like this can occur seamlessly via tools you have always used, with little or no modifications. In other words, when you need more power for a particular application, you can move workloads between clouds with no change to business processes you already have in place. Scaling in and out improves efficiencies, cuts costs and balances workflow.

Once you understand the cost structure of your on-premises cloud, aim for a quick win so you can move on to adding more services and more compute. Your goal is to deliver the right service at the right price, and ideally, to deliver IT as a service (similar to a utility). Set targets and timeframes for realizing ROI and measure results objectively so you know projected costs savings are achieved. Then, plan for process improvements and additional cost efficiencies to follow. More and more IT organizations realize they must leverage infrastructure as a service and use cloud providers for surge capacity, shifting IT into a much more variable cost structure that can be planned and managed.

Measure existing workloads to continuously fill a pipeline of migration candidates for your private cloud and, when ready, public and hybrid options. Factors may include variable vs. fixed costs total costs, usage profiles (bursty? seasonal? waning?), platform, security, and lifecycle status of underpinning assets. As automation improvements and economies of scale drive down unit costs and improve service levels (such as time to provision), more and more workloads will become candidates for cloud delivery models.

Your customers will inevitably try to compare your on-demand offerings to Amazon, Azure, Rackspace and others. Why not help them do it right with the right perspective? What is the cost of your comparable services? How do they differ from public cloud options? Does your offering implicitly benefit from different security services and risk? Can you deliver performance monitoring, portability of data and workload, rate structures for startup, shutdown as well as unit consumption? Will you offer spot versus reserved instance terms and pricing? Yes, there are a lot of questions to answer.

If you build it, they may not come. Explain the business benefits versus alternatives to stakeholders in their language and perspective. (Is the cost lower for them or only for IT? Better performance, quality or security? How much time is needed to provision a new service? Will there be lower switching costs down the road?) Then, listen to stakeholders’ questions and concerns, and tune not only your message, but your services as well.