Ever since the cloud became a major force in enterprise computing, IT leaders have been confronted with a choice: maintain data, applications, systems, and networks in-house or move them to the cloud, with its promise of cost savings, scalability, and improved agility. The cloud, though, is not the answer for every application—especially mission-critical ones that for reasons related to performance, security, availability, or legal concerns are best handled in-house.
The cloud vs. on-premises solutions would therefore seem to be a tough choice. But the good news is, it’s a false choice.
On-premises IT consumption models that offer enterprises the best of both worlds are emerging: the advantages of the public cloud when that best suits the workload, combined with on-premises solutions that make it easier for companies to protect sensitive data and stay in compliance with local data regulations.
These models are often referred to as “pay as you go,” since they enable enterprises to maintain control of their crucial data and functions and only pay for added capacity when needed. You can think of it as a pay-as-you-grow plan.
With traditional enterprise computing models, financial, healthcare, and other kinds of businesses often need to grapple with costly procurement cycles and heavy capital expenditures. The result can often be overprovisioning—paying for capacity well before it’s used. Also, having the ability to quickly and cheaply ramp up capacity is important for any business that sees a sudden or unexpected spurt of growth, has a seasonal business, has acquired another company and needs to integrate it quickly, or is launching a new product and not sure of the sales trajectory.
How built-in analytics gives IT flexibility
Under the pay-as-you-grow model, customers work with their technology provider to determine what their computing capacity needs are today and in the future. That capacity for current demands is installed along with additional capacity for growth.
With HPE GreenLake Flex Capacity, for example, HPE Pointnext employs a usage portal for customers, a dashboard with built-in analytics and controls that reports on services used and capacity consumed. This gives the enterprise a clear, real-time picture of how it is using its capacity. The entire installation is metered constantly, using a wide choice of metrics, such as consumption per server, per gigabyte, per virtual machine, per container node, etc. Starting on a daily or weekly basis, we talk to the customer about what they're using and what we're seeing through metering. This metered usage provides for an effective form of capacity management—think of it as an ongoing process of “right-sizing.” This gives the enterprise a detailed picture of how much capacity it is using and who is using the most.
Additionally, you can integrate a kind of safety valve into your system, which we call a “de-risking buffer.” Together we determine when the IT systems will use up the buffer caused by, say, a major new project or product that’s about to be launched. Once there's a need for more capacity, or the metering reveals that the customer is running at, say, 90 percent utilization, new capacity can be brought in ahead of need. With this type of infrastructure, you should never run out of capacity.
Under this model, the enterprise doesn’t have to go through a cumbersome procurement process to get that new capacity, which in traditional systems can take months. Instead, you can stay ahead of the capacity demand with incremental amounts of equipment brought in ahead of need with a simple change order. In short, the customer’s deployed capacity evolves ahead of business demands.
This is crucial, because just as a company wastes money for maintaining capacity it doesn’t need, lacking computing power or storage when it is needed can result in missed business opportunities. One recent study found that at least 25 percent of IT projects are abandoned because of limited resources, and overall, projects get implemented an average of five months late.
The beauty of the flexible capacity model is that, like with the cloud, customers are paying only for what they're using, above a minimum commitment (see "When flexibility counts").
To be clear, the cloud can be the right place for certain functions. At the same time, not every application is right for the public cloud. In a Hybrid IT environment, what is needed is to apply features of the cloud on-premises. We are convinced that an on-premises pay-as-you-grow model is the future for many workloads—especially those that are growing, have faced challenges with slow procurement cycles, or have a high cost of capital.
Technology continues to change at a dizzying pace, and functions that were barely imaginable just a few years ago are now commonplace. So now, imagine a world where the capacity of your infrastructure can change instantly to meet your rapidly changing needs. Imagine that you only pay for what you use, that costs are aligned with your business activity, and that the infrastructure is simple to use.
In other words, imagine a world where your technology actually works at the speed of your business. With flexible capacity infrastructure, that world is suddenly within reach.
Case study: When flexibility counts
Like many companies, Capgemini Finland Infrastructure Services had outgrown its computing infrastructure and had some important decisions to make. To better serve its customers, the infrastructure hosting service provider needed greater flexibility in managing its server capacity as well as some help in improving its data center operations.
“Our main driver was the need to upgrade our data center infrastructure and processes while moving to more flexible and elastic data center services,” says Pauli Niemi, Capgemini Finland technology manager. “Because our customers were demanding usage-based pricing, we were looking for a more flexible way to procure computing capacity. We also wanted to reduce our data center capital investments.”
Capgemini Finland selected HPE Datacenter Care to help increase its data center efficiency. One feature the company liked was that HPE offered a single point of contact for support and service expertise. As part of the package, Capgemini opted for HPE GreenLake Flex Capacity, which provides the flexibility to manage data center resources more effectively.
HPE GreenLake Flex Capacity enables a “pay-as-you-grow” model that reduces upfront capital outlays and aligns expenses to revenue. It allows Capgemini to quickly expand server capacity at its two data centers without long procurement cycles. The company can now expand or shrink costs based on usage, with no large upfront capital expense. Crucially, its pays only for the data center resources that it actually uses.
“HPE GreenLake Flex Capacity makes us better prepared to acquire new customers and serve existing customers who need to scale their businesses,” says technology manager Jukka Asikainen. “While capacity expansion previously took at least four to six weeks, we can now put additional hardware in place within a couple of hours.”
In addition to cutting capital expenses, Capgemini Finland reduced the number of personnel devoted to data center maintenance and can now shift head count to support customer service and strategic initiatives.
Says Niemi, “We’re less involved in patching and upgrading our infrastructure and are able to focus more on supporting core business objectives, such as developing and testing new services and discovering new ways to better serve our customers.”
This article was written by Scott Ramsay,