Most organizations agree that our future lies in the cloud. The disagreement starts when discussing the leading strategy of our time. Does the world run on hybrid clouds — or do multiclouds set the tone? A question with different answers, depending on the party you ask. In this article, we bring order to the chaos.
In 1980, the question to the most effective foundation for applications and data had a single answer: on-premises, king of infrastructure. Then, network coverage, speed and availability increased. Early successful SaaS concepts came to light in the ’00s. In 2006, AWS appeared on the horizon, the world’s first full-cloud alternative to on-premises infrastructure. Fast forward fifteen years, and you find yourself in today’s world. A world where data is churned out by applications in servers hundreds of miles away, arriving at offices at speeds similar to on-premises. And a world where the answer to the question of the most effective infrastructure is anything but one-sided.
Diversity complicates matters. Technology evolves fast enough to change the profile of an optimal IT strategy every three to five years. For years, cloud-only was proclaimed by AWS spokespeople as the undisputed road. Organizations were advised to abandon on-premises entirely and invest in full migrations to the cloud as soon as possible. The future of large global companies was said to lie in third-party data centers. Nonsense, we now know. The future is not black or white. The future has multiple colours. Call it multi — or hybrid.
From the scrap heap to the cloud
Today, most organizations strive to be able to house and manage applications and data in both company-owned offices and data centers, as well as third-party data centers. In other words: on-premises, in private clouds and public clouds. The combination is at odds with the predictions made by cloud-only fanatics in the 00s. Organizations are not choosing one or the other, but infrastructures with multiple, connected locations.
The explanation for choosing a private and public cloud is simple. Regional regulations, privacy and a personal grip on data and applications longstanding arguments. On the other hand, reasons for the inclusion of on-premises are relatively new. Once again, the continued development of network coverage and speed plays a leading role. Data can be exchanged over a larger area and at higher speeds than ten years ago. Back then, if you wanted to move existing on-premises applications to the cloud, the complete redevelopment of workloads was a common option. The server cabinets on which these applications ran ended up on the scrap heap or are still gathering dust to this day. Today, the applications and servers can stay where they are. Data is relatively easy to virtualize. Thanks to the advent of APIs and SD-WANs, organizations no longer need to redevelop applications.
The aforementioned pursuit of applications and data in on-premises, public clouds and private clouds suddenly becomes logical. Unlike a decade ago, networks are sufficient for synchronizing on-premises and cloud data. On-premises investments no longer need to be thrown away to accommodate legacy applications in a cloud model.
Most organizations are aware of this. Denodo Cloud Survey 2021 states that 35.8 percent of all global organizations deployed workloads in a ‘hybrid’ cloud infrastructure in the past year. Just over 24 percent did so in the private cloud, about 16 percent in the public cloud and only 9 percent in the multicloud.
The former and latter terms bring us to an important topic: the difference between the multicloud and hybrid cloud. As an attentive reader, you may have noticed that, up until this point, we have not explicitly labelled organizations that manoeuvre data and applications between on-premises, private and/or public clouds. Clearly, such organizations are in the majority. The way their approach is described, however, varies greatly.
The language problem
The confusion was underscored earlier this year by a Flexera report with a similar survey question to the Denodo Cloud Survey. While Denodo states that 9 percent of organizations have a multi-cloud strategy, Flexera says the proportion is 92 percent.
You would think that one report is examining apples — and the other pears. Not a bad guess, but in reality, the problem stems from definitions. To understand the phenomenon, a basic understanding of the terms multicloud and hybrid cloud is essential.
The term multicloud has traditionally been used to describe an organization using services from different public cloud vendors. For example, block storage via both Azure Blob and Amazon S3. Two separate licenses.
The reports mentioned above show that researchers agree that we only start speaking of a hybrid cloud when applications and data can move freely between multiple locations — private and public.
However, Flexera implies that every hybrid cloud is essentially a multicloud. The researcher speaks of a hybrid cloud from the moment a bridge is built between the cloud services; data moves freely in both environments, and applications can find each other. In short: every hybrid cloud is a multicloud, but not every multicloud is a hybrid cloud.
Denodo sees it differently. The researcher states that multiple, separate clouds form a multicloud. From the moment the clouds are connected, it speaks of a hybrid cloud. No multicloud is a hybrid cloud – and no hybrid cloud is a multicloud.
The researchers are not the only two organizations with different takes on definitions. HPE says that a hybrid cloud is just a tool for multiclouds: a name for the extension that connects multiple environments, but absolutely not a description of the overarching infrastructure. That would undoubtedly be the multicloud. On the other hand, you have VMware, which argues that all infrastructures with cloud services from multiple vendors fall under the umbrella of multicloud, regardless of whether those services are connected or not. According to VMware, the meaning of multicloud changes to a hybrid cloud once a private cloud is connected to the infrastructure.
Are you confused yet? Don’t worry. We are too. Fortunately, the problem lies in language. While this makes it difficult to use surveys and definitions to pinpoint the leading cloud strategy, in practice, hybrid and multiclouds have enough in common to identify a trend.
Cloud strategies with a focus on one or the other are dying out. Sometimes it is advantageous to connect two separate public cloud services and isolate a private cloud. In other cases, the addition of a private cloud achieves an even better result. That’s where most organizations agree. It is unimportant that HPE speaks of a hybrid cloud instead of a multicloud — and VMware says the exact opposite. Freedom of choice is key. The ability to leverage the resources that are available at a given time and place; the ability to move data and applications freely between clouds. Or not, depending on the approach that works best for a given issue at a given time.
The demand for independence from the templates prescribed by cloud providers has never been greater. Independence is, therefore, the future. The coming years are full of solutions that allow deployment and management of applications, data and workloads in any environment in a (proverbial) jiffy.
We will be discussing the latter in more detail in one of the next articles in this series. As soon as that article is published, we’ll add the link below.