Dario Lo Presti - Fotolia
For years, businesses have relied on established applications on-premises but the emergence of a hybrid cloud approach certainly has become an enticing option. The combination of an on-premises and an off-premises computing environment could help reduce costs, which is something IT professionals are often tasked with doing. Systems, networking and storage technology are changing rapidly, and many organizations simply cannot keep pace with these advancements. Power and cooling, meanwhile, are major cost factors in the data center, requiring significant staff expertise in the design, implementation and operations phases.
A hybrid cloud approach would help reduce the amount of power and cooling a company would have to use and pay for. But the idea that there will be computing environments off-premises that admins can't get their hands on – literally – has some companies hesitating.
The reluctance is real
Some businesses are reluctant to relinquish the hands-on control of workloads that comes with an on-premises environment. Many worry about uncontrolled costs, poor levels of performance and potential security problems that the cloud might introduce. Businesses in highly regulated industries or those in regions that have data location or governance laws fear they are going to have trouble complying with established laws.
Also, executives may be concerned about not having sufficient experience or knowledge to conduct a proper analysis of capital expenses and operational expenses, something that’s central to cloud computing.
These fears are certainly reasonable, but they can be addressed—and overcome. Educating decision-makers will help them understand what cloud computing is and how it can be accepted as simply another form of outsourcing.
A review of IT workloads will likely show that some workloads are best kept in-house and on-premises while others could be effectively and efficiently be hosted in a colocation facility or in the cloud.
Highly proprietary or tightly regulated workloads are the ones best suited for an on-premises environment. Generic workloads, such as collaborative computing, messaging or customer relationship management, are prime candidates for the cloud.
Using a service provider’s data center can lower costs for real estate, systems and software licenses while allowing a business to either redeploy or reduce its IT staff. Put simply, these cloud services provide users with necessary computing resources, while diverting the management responsibilities to others.
Companies in industries such as manufacturing, healthcare or finance, for instance, may decide it’s wise to focus on their core businesses rather than to invest time and resources in operating their own data centers. Wouldn’t it be better to simply pay for just the portion of a data center that the organization actually uses? It’s an important question to consider.
Careful examination of an organization’s needs and current IT assets are important steps during the planning/decision-making process. A business will often learn that it runs too many data centers, devotes too much floor space to data center functions, or operates facilities that should be redesigned or consolidated.
The strategic questions
Organizations too often overlook the importance of developing a sound strategy before selecting the appropriate data center architecture or technologies. This brings to mind the wisdom of Yogi Berra, who said: “If you don’t know where you are going, you will wind up somewhere else.”
IT planners have much to consider before developing a data center strategy. Here are some key questions to ask:
- What applications and workloads are required for business success? What is the expected lifecycle of each of these applications?
- Do these applications have performance, availability, reliability, transportability or regulatory requirements?
- Are these applications platform-neutral or do they require specific systems architecture, operating systems, application frameworks, databases, networking or storage technology?
- Would these applications execute properly in a VM or operating system partition?
- What growth is expected in the application’s use of processing, memory, storage or networking capacity in the next year? The next five years? The next decade?
Defining the correct IT strategy typically requires expertise in different system architectures, distributed computing, operating systems, development tools and environments, database management tools, virtualization software, management software, networking, storage and IT operations. Also, expertise in cooling, power supply, communications and other areas is essential in devising a reasonable, supportable plan.
Many businesses don’t have enough (or any) of the expertise needed to accomplish this type of thorough planning. This is where a potential partner with real-world information can provide decision makers with realistic and supportable projections, modeling and cost estimates.
Applications today are being designed to execute on multiple systems, each of which might be hosted in a different data center. Rather than relying on the power of a single computing resource, new applications are designed to scale up to support huge numbers of transactions or manage vast amounts of data.
Processors are faster, smaller, use much less power and often need less cooling. Storage, too, is rapidly evolving from using rotating media to using various types of solid-state technology. Importantly, smaller, faster storage requires less power and cooling.
Data centers must be designed to support today’s established systems as well as accommodate future changes. Organizations need to examine data centers from time to time to evaluate whether they are underprovisioned, overprovisioned or needed at all.
Upon careful examination of its IT assets, an organization is likely to come to one of the following conclusions:
- Its data centers are just fine and will be acceptable resources for a number of years to come.
- Its data facilities aren’t large enough to address foreseeable growth and must be enlarged or replaced.
- Current environments are too large for today’s requirements, and expected growth will not fully utilize that capacity. This may lead to efforts to consolidate data centers.
- An appropriate course would be to outsource some or all IT assets to the cloud.
Blending on-premises and cloud environments
Explore hybrid cloud computing's uses and best practices