With applications running in the cloud, it’s easy to focus on the cloud services without ever thinking of the underlying operating system — or even needing to think about it. It almost seems quaint, to think about standard operating environments, base images, and compatibility like it’s 2010.
When all of your innovation and future projects are in the cloud, does the operating system really matter anymore?
What’s running in the cloud today?
Right now, according to respondents surveyed for Flexera’s State of the Cloud report, half of their workloads are running in the cloud, with another 7% shifting over within a year. Red Hat’s own internal research indicated that as many as 70% of Linux servers are deployed in the cloud.
This is part of the evolution that puts development and innovation at the center of IT departments —rather than delivering support services, IT and engineering are crucial for business strategy and execution. It’s the old mantra that “the developer is king.” The need for rapid scale for production loads and to spin up new proof-of-concept environments means that infrastructures are shifting from physical servers and datacenters to self-service, cloud-based environments (and usually multiple cloud environments).
It matters how you adopt cloud
Spoiler alert – yes, the operating system (probably) still matters to you and your IT teams, but the ways that it matters, and how visible those issues are, is different than it was when datacenters were the heart of the infrastructure.
To answer this question, first you have to take a step back and evaluate how organizations are actually adopting cloud.
The high-level numbers make it look like there is a general, amorphous shift to cloud, with a majority of “workloads” being in “the cloud.” But when you break it down, you start to realize that there is a distinct pattern to what kinds of workloads are deployed and where.
So, start with the promises of public cloud. The cloud is supposed to deliver:
- Dynamic scalability, on demand.
- Ease of access and self-service.
- Lower costs and more efficient provisioning.
- Flexible technology choices.
The idea is that everything would naturally gravitate toward an environment with scalability, ease, etc.—but the reality is that some workloads are simply more suited for those environments, while other workloads have different priorities like security or user management.
While not a strict rule, cloud workloads tend to be more transient and likely to change than on-premise workloads so they tend to be used for rapidly-iterating applications like development, testing, and proof of concept (POC).
But ultimately, workloads are not tied to any one location. According to the same Red Hat research, workloads are pretty consistently moving in and out of clouds as the needs of that specific application or other data or user considerations change:
- 58% of respondents moved workloads from on-premise to cloud.
- 30% moved cloud workloads to on-premise.
- 27% repatriated workloads from cloud back to on-premise.
And therein lies the hidden importance of the operating system as a strategic choice for cloud infrastructure. Cloud (and on-premise) workloads are fluid. They will move. Organizations need a way to be able to migrate those workloads effectively—and the number one strategy that they use (40% of respondents) is a standard operating environment.
Be intentional with your infrastructure
Relying on the operating system as part of your SOE strategy is only part of what an operating system contributes to your infrastructure, from foundational security settings and access to user and service accounts to management and automation.
When the operating system is so critical as part of your infrastructure foundation, how was it forgotten?
Much of it probably comes down to familiarity causing it to be overlooked. Modernization efforts, new innovations, trying to onboard new technologies, focusing on new development—the push forward can make it easy to lose some of your former best practices and priorities.
As digital transformation efforts have only accelerated in the past year, some of the pitfalls of poor cloud planning and execution are becoming apparent even faster. Some analyst research has claimed that 40-50% of physical infrastructure capacity is wasted, because of over-purchasing to account for demand spikes or to increase access to resources. And according to StormForge research, about the same amount (48%) of cloud instances are wasted, because of a lack of structure or planning in how cloud resources are used.
As an IT leader, you have the opportunity to make the most out of your infrastructure choices, whether in the datacenter or in the cloud. Go back to the basics of good IT strategy:
- Define a standard operating environment. The operating systems you use for both development and product are a huge part of that, but not all of it. Identify other key services, like library versions or required services.
- Set guidelines for how and when to request new cloud resources—and when to remove them. A lot of cloud instances are for temporary workloads like testing or POC environments. Define both best practices for requesting a new instance as well as policies around removing them to minimize cloud waste. Management tools like Red Hat Insights can even be used to identify idle resources (and consistent management tools are another benefit to an SOE).
- Define a migration strategy. Lifecycle management—for both the application and the operating system—is crucial to manage resources. Set requirements around storage, memory, bandwidth, or data access that indicate when an application needs to move between on-premise or cloud environments. And define how to handle technology changes, whether to redeploy or upgrade and how to handle transitions.
- Plan for security throughout your entire process flow. Security starts with your operating systems, with service permissions and user access and a lot of other critical settings, and it flows from the foundation of the IT stack upward. Evaluate the interactions between environments, data needs, user needs, and resource requirements, and plan each with an understanding of the native security and security requirements at each point.
Defining the foundation
Your operating system is the foundation of the rest of your IT strategy. Shift your approach from a project-level perspective to a more holistic view of the entire infrastructure and its purposes.