As cloud usage continues to grow, it's important that IT managers understand where different types of applications are best suited to be deployed. Karthik Narain, Cloud First lead at Accenture, told ITPro Today that as organizations move up their cloud quotient from approximately 20% today to about 80% over the next three to five years, their IT landscape will contain applications running in public cloud, private cloud and edge environments — hence, they will operate in a public and hybrid cloud model.
Too many organizations get caught up trying to squeeze everything into a single model that is often either public or private cloud, according to Cooper Lutz, cloud architect at AHEAD. That said, by far the biggest challenge almost every organization faces is determining what “cloud” means to it, he added.
"As organizations begin to adopt public cloud services and shift to a hybrid cloud hosting model, determining where to place applications becomes a massive challenge for many organizations," Lutz told ITPro Today. "Many factors need to be taken into consideration, and each organization is different, but everything really should boil down to a concept of, what value do I gain out of hosting this application in the public cloud?"
Challenges of Hybrid Cloud Application Deployments
While some organizations have a broad directive to adopt the cloud, what is still missing is an application-focused, data-driven approach, according to Dave Knight, IBM Alliance Cloud lead at Deloitte Consulting.
"The decision of whether an app is migrated to the cloud, redeployed as a hybrid cloud app, left alone or retired should be based on the application and its value to the business," Knight told ITPro Today.
Hybrid cloud really shouldn't be viewed as a cost reduction play, he added. Rather, think of hybrid cloud as a way to extract the most business value from IT investments where application components run where they are most effective and drive business value — which isn’t necessarily the least expensive option.
Among the challenges and considerations for hybrid cloud application deployments cited by Knight are:
- Latency. Applications that require near real-time responses aren’t the best candidates for hybrid cloud deployments. Even if the on- and off-premises components are physically close, traversing public networks can cause some transactions to time out and fail.
- Regulations. Rules and regulations don’t typically prevent a hybrid cloud architecture, but they can limit what hybrid architecture and design options can be implemented. For example, data sovereignty requires that the data and application reside in the originating country, so an organization's hybrid cloud model would need to comply.
- Data gravity. Application architectures that involve moving large amounts of data aren’t typically thought of as good hybrid candidates. Shifting all this data around can result in significant charges — not to mention the amount of time it takes. The transport time on large volumes of data can reduce the time available for performing actual work once it arrives at the destination.
- Cost. In a hybrid cloud architecture, having part of the application on-premises and part off-premises happens by design. So by design, there will be continuing costs for existing infrastructure and new costs for the cloud portion of the hybrid model.
Identifying the Right Applications for Hybrid Cloud
For most organizations, choosing the right cloud strategy comes from approaching cloud as a holistic journey that considers the needs of innovation, business differentiation and value creation, along with cost and efficiency objectives, according to Accenture's Narain. When considering if hybrid cloud is a good option for a specific application, it's paramount to evaluate the appropriate application and data dispositions, he added.
"The decision is often more correct if it stems from an end-to-end cloud strategy and aligns with a client’s industry- and business-specific innovation journey," he said.
Narain suggests four key steps for organizations to take when determining a strategy of where applications should be placed.
- First, make sure there are well-defined application modernization and data strategies and road maps to help drive the cloud strategy. This should include making a bet on the organization’s preferred public or hybrid platform-as-a-service (PaaS) strategy for developing applications.
- Start developing a hybrid cloud strategy and road map aligned to the application and data strategy, with workload placement simultaneously optimized for performance and cost. The strategy should also integrate the infrastructure road map to ensure critical enabling capabilities are in place at the time needed.
- At the same time, start transforming operations to provide tighter integration between applications, data, cloud, infrastructure and security, as well as building on a cloud operating model powered by applied intelligence or AIOps (analytics, automation and AI).
- Look for technology and service partners with the skills and experience that can help the organization successfully navigate to the new platform and deliver real value to the business faster.
"Enterprise strategies must be application and data led, cloud and infrastructure enabled, secure and optimized for the operational run," Narain said. "For these reasons, a hybrid application and data placement will be the natural choice for most large enterprises."