As businesses large and small start to explore the cloud, you need to understand not only the benefits, but also the changes brought by moving to a cloud delivery model.
Here are our top tips for building a cloud-computing strategy.
Cloud computing is on the rise. It isn’t showing signs of slowing anytime soon: According to a recent Gartner study, spending on public cloud services reached $111 billion in 2012, and total spending over the next three years will be $677 billion.
Your ideal cloud—public or private—should be a self-service environment for delivering the most critical applications to the users that need them most. It should be elastic—able to grow and shrink as the demands on it ebb and flow.
Many businesses are looking to build private clouds as their first step. It highlights a fundamental change in how enterprise IT operates:
- Step one in designing a private cloud strategy should be an assessment of the workloads that are most important to the business.
- Step two should be choosing a cloud management solution that embraces all workloads to provide the most cloud benefit to the business today and over time.
- Step three is to understand that not all cloud implementations are alike.
Step One: Workloads Drive The Cloud
The vast majority of enterprise workloads do not yet embrace “cloud-era” architectural principles.
Much of the early private-cloud design was modeled after that of Amazon- and Google-style commodity public clouds. However, these clouds can overlook a critical architectural factor that enterprises must consider: the diversity of their computing workloads.
If builders of enterprise private clouds are to see significant benefits from a private cloud strategy, they must do more than accommodate the few cloud-era applications that exist today. So, to truly impact agility and costs, enterprises must choose a cloud strategy that accommodates traditional applications like SAP, Oracle, and the other applications running throughout the enterprise.
Even though these applications may not run perfectly on cloud platforms, and though they may lack the ability to elastically scale, they can still benefit from other cloud features—such as automated orchestration, provisioning, and self-service delivery.
[Need help choosing the first applications you should run in your cloud? Follow Cesar’s 7-step template.]
Step Two: Designing A Cloud For All Workloads
Even though you’ll probably initially choose one or two tactical applications to move to the cloud, your strategy should look to the future.
A cloud strategy should consider all workloads: It should include a management layer that enables automated orchestration and provisioning of any infrastructure—commodity or specialized.
Take a lesson from large public providers and segment your cloud-ready applications from your traditional enterprise applications. This enables each to best use the infrastructure components that they benefit from the most, while still delivering a new level of agility to the entire datacenter.
This type of design not only delivers benefits to enterprise IT today, but also enables a longer-term strategy of migrating legacy enterprise applications to commodity infrastructure over time.
As cloud architectures infiltrate enterprise datacenters, traditional vendors such as SAP and Oracle will rewrite their applications to operate efficiently in this new infrastructure. Enterprises that embrace a multi-workload cloud strategy today will be best positioned to benefit from those changes as they occur.
Step Three: Understanding The Economics
Vendors typically think of clouds as pools of “good-enough” commodity hardware, designed for low-cost and massive scale. Those infrastructures are planned with the expectation that hardware will fail.
The challenge for the enterprise is this: Inexpensive infrastructure that’s expected to fail is the exact opposite of how we’re used to designing enterprise datacenters. The typical datacenter is designed to avoid failure, relying on largely static racks of high-priced, fully redundant, specialized hardware designed for 99.999% uptime.
That’s a very different cost model to a cloud infrastructure. Does this mean that enterprises shouldn’t be building clouds? Absolutely not.
But it does means that enterprises need to carefully consider the benefits of cloud, understanding that Amazon-like cloud economics may not be a reality for them—at least not for all applications.
The Bottom Line
Every organization is different. Your strategy for moving to a cloud model will be unique to your circumstances.
Start planning now. Or be left behind—at a disadvantage to your competitors.
Sameer Dholakia is group vice president and general manager of the Cloud Platforms Group at Citrix, driving the company’s product strategy for cloud infrastructure and server virtualization. He joined the company in 2010 when Citrix acquired VMLogix, where he served as CEO.
Now Read This (more from NetAppVoice)
• Are Cloud Data Security Fears Overblown? A Sensible View.
• 100 Words Into The Future: Is The Cloud Safe?
• Steve Wozniak: Apple, Cloud, Flash, Luck, And Open Doors
• Read more from our talented writers