The big picture: At this point, virtually anyone who follows the tech industry in even the most casual way has probably heard, not only about the influence of cloud computing, but also the impact of what is commonly called “multi-cloud.” What many don’t know, however, is the specifics of how much companies are using these cloud computing resources, what types of workloads they’re running in the cloud, why they chose to use cloud services, and much more.
A new research study, initiated by TECHnalysis Research, dove into all those details. It began with a survey of 600 US-based businesses (200 medium-sized companies with 100-999 employees and 400 large enterprises with 1,000 employees or more) who were users of cloud computing services. The results show that today’s cloud computing environment is an incredibly dense, rich tapestry of different workloads at various maturity levels running in different locations on different underlying platforms for many different reasons.
The basic idea with multi-cloud is that companies use multiple different cloud computing options as part of their overall computing environment. In some cases, that could mean using multiple public cloud providers, such as Amazon’s AWS, Microsoft’s Azure, and Google’s Cloud Platform (GCP), or it could mean they’re using one public cloud provider and one or more “private” or “hybrid” clouds, or some combination of all the above.
Private cloud refers to computing environments that use the same basic types of flexible technologies and software platforms that public clouds offer but do so either within the company’s own data center or in what’s called a “hosted environment.” These hosted environments are external sites that house the physical resources (servers, storage, networking equipment, etc.) necessary to run computer workloads from multiple different companies simultaneously.
Typically, these locations—which are sometimes called co-located sites or “colos” for short—provide power, strong physical security, and most importantly, high-speed connections to large telecommunication networks or other network service providers. Unlike with public cloud companies, however, the physical assets (and the workloads) at these sites remain under the control of the company requesting the service.
Hybrid cloud refers to environments that mix some element of public cloud computing providers with private and/or managed/hosted providers either within the data center or at a co-located site.
What the study found is that for companies like those surveyed, who have been using cloud computing for several years now, approximately 30% of today’s workloads are being run in the public cloud, another 30% are legacy applications still being run in the corporate data center, and the remaining 40% are a combination of private and hybrid cloud workloads, as Fig. 1 illustrates.
Interestingly, when asked what companies expected the mix to look like in 18-24 months, the results weren’t significantly different, with about a 5% drop in legacy workloads and about a 2.5% increase each for public cloud and private/hybrid cloud workloads, suggesting the transition to new cloud-based workloads has slowed for many of these organizations.
In addition to this diversity of high-level workload types, the study showed a large number of options being used within each of those groups. On average, for example, survey respondents were using 3.1 different public cloud providers across both IaaS (Infrastructure as a Service—typically access to the raw computing resources of a public cloud provider) and PaaS (Platform as a Service—adding software and services on top of the raw hardware) offerings. Of the nearly 87% of respondents who said they were running a private cloud of some type, they averaged 1.6 different private cloud platforms.
When it came to specific workload counts, companies averaged 3.4 workloads per public cloud provider and 2.9 workloads for private and hybrid clouds. Doing the math, that means organizations like the ones who participated in the survey typically have over 15 cloud-based workloads that they’re using. On top of that, survey respondents deployed a number of SaaS (Software as a Service) cloud-based applications as well. These include Microsoft’s Office 365, Google’s G Suite, Salesforce, and many others, and the average per company worked out to 3.7. As a result, today’s US businesses are balancing nearly 19 cloud-based applications/workloads as part of their computing environment, as the table shows below.
The reasons for moving all these different workloads to the cloud vary quite a bit by the specific type of workload, but looking at the weighted totals across all the various types and locations provides some interesting, though not terribly surprising, insights into the rationale that organizations are using to make the move to migrate or rebuild existing applications, or create new ones in the cloud. (Speaking of which, companies said that approximately 1/3 of their cloud-based applications fit into each of these three categories: migrate, or “lift and shift,” rebuild, or “refactor,” and build new.)
The top reasons that survey respondents gave for migrating workloads to the cloud are to improve performance, to increase security, and because of the need to modernize applications. Cost savings actually came in fourth. Ironically, the top reasons those same companies cited for not moving some of their applications to the cloud were very similar: security concerns, performance challenges, regulatory requirements and costs. These dichotomies highlight the ongoing challenges and opposing forces that are a regular part of the modern cloud computing landscape.
There’s no doubt that cloud computing, in all its various forms, will continue to be a critical part of business computing environments for some time to come. Making sense of how experienced companies are approaching it can help vendors optimize their offerings and other businesses find their way through the often very confusing cloud computing world.