In IT these days, many position cloud as a panacea to fix all the ills of a modern business-focused IT infrastructure. This is especially true when it comes to trying to reduce overall costs. While upfront capital expenses (CapEx) can be reduced with public cloud, the resulting operational expenses (OpEx) could eliminate any savings from reducing CapEx. On the other hand, private cloud implementations tend to be heavier than normal upfront CapEx costs, with the promise of reduced OpEx costs. In the end, costs could range widely depending on the model chosen and how it is implemented and managed.
When considering the public cloud option, it’s important to realize that any existing business should not expect to simply swipe a credit card and pay the monthly bills because there are still plenty of upfront costs that should be expected. Assuming your business already has an existing IT infrastructure, it will take time to get the new infrastructure setup and configured before even migrating the existing data and workloads. The data migration is no small task either, and likely will require some amount of downtime on top of the costs of data migration services that may be required to manage the process.
In order to secure the best overall cost for a public cloud infrastructure, paying for the year – or multiple years — upfront will probably be required. This makes a lot of sense when you consider all the time and effort that goes into setting up the new environment, migrating to it, and training the staff to use it. Who wants to invest that much into a platform that will only be used for a few months. Given the additional costs to move out of the public clouds (including both labor and potential “exit” fees), businesses are incentivized even more to commit to a longer term public cloud implementation. This puts the IT budget into a similar cycle as building an on-premises infrastructure: evaluate and reevaluate thoroughly, heavy negotiation on price and services, process redefinition, and staff training since you’ll have to make do with the selection for one or more years.
Once on the other side of a public cloud migration, even with lower CapEx costs, the business may end up with even higher OpEx costs. Similar to early iPhone users on AT&T, there are many stories about businesses that were shocked to see their usage-based bills. With runaway costs caused by VM sprawl, unexpected workload bursting, and zombie workloads, the monthly costs may not be as predictable or beneficial as initially expected.
For companies considering private cloud implementations, the CapEx costs are likely to be higher than business as usual — potentially much higher. But since not all applications perform well enough or are cost-effective in the public cloud, private cloud adoption is growing, and while customers are considering public cloud, they’re considering it in the context of a hybrid cloud model. Obviously, operating a private cloud still requires up-front costs for acquiring and implementing hardware, software, and a cloud-management platform, but these owned and on-premises components, still need to provide the scalability, extensibility, manageability, and flexibility the business is looking for.
This requires a focus that keeps in mind that “cloud” is comprised of two different models. Characterized by automated, self-service provisioning, elastic on-demand capacity, metered usage and chargebacks, and a single, shared resource pool to support all workloads, the first cloud model is the model the business needs to get out of the infrastructure, including flexibility, seamless scalability, and the ability to iterate very quickly to provide new capabilities and services. The second model is the construction and management of the infrastructure itself to ensure ease of scalability, service extensibility, and the ability to manage the environment via software — and consequently automation.
One interesting result is that when building out a private cloud, it will most likely be a silo with dedicated hardware, a dedicated physical or logical network space, and potentially a dedicated support staff. This approach makes sense since it makes it easier to achieve predictable results — a goal all new infrastructure should have — and because the project is likely to be a sizable project with its own dedicated budget to include the additional hardware, software, and implementation services required to bring it online. This means the private cloud infrastructure is generally implemented as a greenfield deployment, which invites the consideration of new technologies that can make the implementation quicker to build and more efficient to operate. This is where hyperconvergence can have a huge impact in the delivery of cloud infrastructure.
Born out of the software defined storage movement, hyperconvergence has its feet planted firmly in the software defined everything space, making it a natural platform for a cloud infrastructure. With hyperconverged infrastructure there is less equipment to acquire upfront, which results in not only saving on equipment costs, but also in investigation time, training costs, more predictable scalability costs, and a quicker time to value during the implementation.
As an infrastructure approach that combines multiple IT infrastructure components — some more than others — onto a single platform with simplified management, hyperconverged infrastructure can help reduce the cost of private cloud even further by also driving down OpEx costs. By simplifying the management of the infrastructure that previously would have required multiple management tools, a hyperconverged infrastructure can not only improve the one-time manual tasks that are completed via GUI tools, but also can make the automated tasks simpler by reducing the number of API calls that are required to accomplish a task.
With all of these advantages to a cloud infrastructure, many service providers are starting to look seriously at hyperconverged infrastructure to underlie their cloud infrastructure. With global management capabilities available in some hyperconverged infrastructure platforms, utilizing the same vendor for both the private and public cloud segments can introduce some powerful enhancements to a hybrid cloud infrastructure.
By providing the lowered upfront CapEx costs and predictable scalability of a public cloud infrastructure, and the workload locality and tight control offered by an on-premises infrastructure, hyperconverged infrastructure can deliver the CapEx and OpEx benefits to companies to build out a highly automated private or hybrid cloud infrastructure.