As organizations become increasingly comfortable with putting business-critical apps in the cloud, many – particularly Internet-centric startups – have completely bypassed making in-house infrastructure investments.
However, while virtual public cloud services can enable business agility with lower upfront costs, no capital spending and pay-as-you-go pricing, these organizations often encounter performance and cost challenges as they begin to scale out their applications.
As infrastructure and operations teams face these more complex, large-scale cloud deployments, the question is not whether a public cloud can support performance-sensitive applications but how to ensure they choose and deploy the right cloud solution. To this end, here are three suggestions:
Investigate bare-metal public cloud as an option
Applications and workloads that require more performance and processing power can put stress on traditional, virtual clouds. Performance degradation stems from the introduction of a hypervisor layer and the multi-tenant nature of virtualized public cloud platforms. While the hypervisor enables the visibility, flexibility and management capabilities required to run multiple virtual machines on a single box, it also creates additional processing overhead that can significantly affect performance. Also, when too many virtual machines compete for server resources, they become “noisy neighbors,” restricting I/O for data-intensive workloads, and resulting in an inefficient use of physical resources.
Bare-metal cloud has emerged as a way to complement virtualized services with a dedicated server environment that eliminates the performance constraints of virtualization and the hypervisor without sacrificing the elasticity benefits associated with the cloud. A true bare-metal cloud combines the high-performance of dedicated, single-tenant servers with instant scalability, on-demand self-service and programmability via an API.
Bare-metal and virtualized clouds are not competitors. They are simply different flavors of IaaS technology that allow customers to meet a wide range of workload and application requirements.
In fact, establishing a mixed cloud environment is often an ideal approach. With this setup, companies can choose how to best support each of their core applications and services on an individual basis, thereby reducing capital costs, maximizing operational efficiency and establishing a foundation for innovation through adaptable hosting models.
Take a hands-on approach to comparing IaaS offerings
When choosing a cloud solution for a high performance, business critical application, don’t write an RFI (News - Alert)/RFP/RFQ – these are all methods for learning about what the post-sale experience will be like. Instead, spin up some infrastructure with a few providers, and turn it off when you’re done.
Taking this hands-on approach will allow you to gain a more in-depth understanding of what the offerings provide in terms of performance, availability, customer service and ease-of-use, among other criteria. Your staff will get invaluable hands-on time, and you’ll likely spend less time and money than you would in a protracted sales process. This type of evaluation will give you a real-world view into how well a provider’s infrastructure will support your specific applications.
Set the right expectations for your first high performance app deployment
When you’re ready to begin your public cloud deployment for a performance-sensitive or data-intensive application, set realistic expectations.
Deploying your first application will require training. Don’t forget incentives during this process, including your own. Establish achievable metrics and objectives, but don’t create an artificial roadblock by placing too much emphasis on setting “perfect” goals at this point, or you’ll never get started. Most importantly, don’t be shortsighted – stay focused on long- term goals and results.
Inevitably, there will be issues that arise, and everyone will be quick to blame the infrastructure, so focus on an application that is robust and has a supportive user community. The real key is to start the process and prove to your organization that it can be done, which sets the stage for a more comprehensive application migration strategy.
Edited by Maurice Nagle