We thought everyone finally had cloud terminology all cleared up. You’ve certainly seen the countless blogs about IaaS, PaaS, and SaaS; not to mention the ever-proliferating surveys and reports on hybrid cloud being the deployment flavor du jour.
But things aren’t as clear as we might want them to be. For example, tell me what you think of when you hear the term “public cloud.”
Is it a hyperscale provider like AWS, Azure, or Google? It is, isn’t it? If not, you probably work with or for an organization similar to Green House Data, which has a public cloud offering with some major differences from the hyperscale players.
So how can we clear up the cloud? Has public become synonymous with hyperscale and self-provisioning? Has private cloud fallen by the wayside? And what should your business focus on, anyway?
Previous explanations of public, private, and hybrid clouds focused on whether the underlying hardware hosted multiple tenants (public cloud) or was a dedicated, virtualized environment for a single client (private cloud). Hybrid, meanwhile, continues to be very popular, as it is loosely defined as any combination of cloud resources and dedicated infrastructure, such as bursting from on-site data centers to the cloud as demand necessitates, or adding new cloud-native apps alongside currently deployed colocated infrastructure.
But read any cloud-focused blog these days and it would appear that the perception of public has shifted. A public cloud server is no longer any old virtual machine that is hosted on shared hardware. Rather, it is assumed that a public cloud server can be instantly provisioned via web portal, and that it comes from a hyperscale provider with multiple availability zones. There is no distinction made between a hosted public VMware environment and Azure or AWS workloads, because the hyperscale players control over half of the market.
One study found that 68% of cloud related IaaS or PaaS services were held by some 24 hyperscale providers.
However, that leaves a significant portion of users who are not actually on a hyperscale cloud. Are they exclusively using private cloud services? Highly unlikely, as only 38% of all workloads run in the private cloud.
The market will more or less decide things on its own — but we might see more marketing around “hosted VMware” instead of “public cloud VMware” in the near future. That being said, there is little functional difference between a hyperscaler and a provider like Green House Data. The biggest change would be that you have to go through a contract and deployment process that could last up to a few days when deploying your initial cloud resources with a smaller provider, rather than simply entering your credit card information online and instantly receiving a server.
Of course, customer service levels are also vastly different when working with an enormous corporation like Microsoft when compared to a boutique provider like Green House Data. Both options offer engineering resources and rapid response times to tickets, but you may have to hit a minimum spend to get the same level of service with a hyperscale provider.
With an outsourced private cloud, equipment is provided without the costly capital investment.
While year-over-year private cloud deployments have gone down, strong demand also remains for these types of environments. Often times, a company will report hybrid cloud use in a survey, because they have both public and private components to their cloud infrastructure.
At Green House Data, private cloud deployments are frequently requested and make up a significant portion of our overall cloud composition. Government organizations, healthcare providers, companies that store financial data, and other compliance-mandated or security-oriented deployments will regularly use a private cloud (even if cloud hosted servers are generally more secure than on-premise infrastructure, anyway).
You might stick with private cloud if you have those privacy or security concerns, or if you require very high performance and wish to limit the impact of fellow tenant activity. Specific cloud modules, applications, or custom hardware deployed alongside your cloud could also necessitate a private deployment.
For the majority of workloads, hybrid will remain the best fit for years to come, as companies continue to maximize their existing investment while expanding into the cloud and redesigning to be cloud-native. Those hybrid workloads may even straddle hyperscale and smaller providers alike, using one for primary workloads and one for location-specific or disaster recovery resources, for example; or taking advantage of the engineering team at a local provider for a specific project that requires supplemental staff.
No two cloud workloads are ultimately alike. Finding a collaborative provider like Green House Data can be the key to designing the most cost- and performance-effective cloud infrastructure for your business goals.