There are myriad technical considerations when deciding how to architect and deploy your cloud infrastructure, but your business structure, size, strategy, and industry are also significant factors.
You don’t need to take a deep dive into technical evaluations of each workload to choose between public cloud and a hybrid or private infrastructure. It’s possible that your business practices will make that decision for you before you ever get to individual app/system analysis.
Here are the business traits to consider as you weigh your cloud options.
The size and/or age of your organization may be enough to dictate your cloud architecture. Startups are ideally suited to the public cloud’s flexibility and lack of asset ownership. They have very little investment in existing infrastructure and may need to adjust on the fly. A monolithic, large company with a multitude of server assets is more likely to look at a hybrid or private cloud approach.
It can be difficult for CFOs to adjust to the cloud. You’ve probably seen numerous articles discussing the shift to operations expenses (OpEx) when using a subscription computing service rather than capital expenses (CapEx) when purchasing your own equipment. Preference for OpEx or CapEx from your purchasing department might be enough to make the decision for public or private already.
However, having a big war chest available might also make a hybrid deployment attractive, since you have more control over the infrastructure and security. Similarly, bandwidth costs can factor in, especially when migrating to the cloud. A high data volume might mean on-premise components are preferable.
This category could go either way. Generally, if you don’t have many staff dedicated to helpdesk and support, and limited cloud experience, the public cloud is a good option. If the opposite is true and you have a strong, cloud-focused IT staff, then you’re ready to manage your own cloud infrastructure on-premise or in a hybrid configuration.
It’s not always so simple. Mastering a public cloud platform can be difficult and many of the large providers do not offer responsive support. Your limited staff could be learning on the fly, and if a configuration issue occurs during off hours, you might be scrambling. You need to spend a significant amount with hyperscale public cloud providers, or pay for an additional support contract, in order to gain dedicated support resources.
An on-premise virtualized environment could be a better option for those with limited resources then, as long as your staff knows the platform. Or you could use a smaller, “boutique” provider or MSP who has wide experience with cloud platforms and technologies. They can help you with a public or a hybrid/private infrastructure, filling your staff gap regardless of the architecture.
Talk to an infrastructure consultant today.
If you need to reach far-flung global operatives and require low latencies for your end users, then a hyperscale cloud may be your only option. Setting up private clouds across the world is no small task. Even if you have offices across your country, or a large cohort of remote workers, a distributed cloud environment in a public model is often the way to go, especially if you have performance and latency concerns.
If latency is not an issue, or the majority of your staff is centrally located, a private cloud could be a good option, and you can always expand via hybrid. Similarly, large scale organizations with multiple, centralized sites could build individual data centers near their staff.
Data sovereignty, compliance, and security considerations come into play with the location of your users as well.
Data sovereignty is becoming more and more of a global issue (read some of the latest developments on European Union and United States data privacy and exchange laws). If your country has restrictions on exchanging data across borders, you may need multiple private clouds, or otherwise segregated public clouds from different providers or different availability zones.
Compliance and security concerns also drive many organizations towards a private or hybrid cloud, where they can keep some data segregated (a la with payment information, health information, or national security information) and less sensitive data in a public environment. If you have few regulations in your industry, a public cloud is likely the better option.
Risk acceptance is another factor. Simply put, you do have more control (for better or for worse) over your own infrastructure. A private cloud allows that control, so you can build in your own redundancies and maintain your own network. While the majority of cloud providers offer a 100% SLA, compensation in the case of downtime can be limited consolation. You must accept the risk that your public cloud may not always be available, often due to factors outside of your control like internet or service-provider outages.
No cloud is one-size-fits all, and these rules of thumb may or may not help you reach your future cloud goals. But they can at least help guide your organization towards an initial cloud model. We’ll take a look at technical and other miscellaneous factors that also inform your cloud design in a future post or white paper, so be sure to check back. In the meantime, feel free to engage with one of our cloud infrastructure experts for hands-on guidance.