There are plenty of reasons organizations shift to the cloud from a legacy IT environment where they’ve been relying on a traditional infrastructure with compute, storage and networking. Our recent Cloud Impact Study, which surveyed 400 senior IT professionals, found some of the main drivers of the change were to gain increased efficiency (72 per cent), increased business agility (56 per cent), increased security (51 per cent) and business continuity (48 per cent).
But it’s important to realize that not every application can be immediately and simply moved over to the cloud. Different workloads operate best on different infrastructures. As analyst firm Gartner noted in a recent article, organizations might want to consider shifting from a “cloud-first” approach to a “cloud-smart” approach, balancing cloud adoption with a company’s unique business goals. Cost, compliance and performance are common reasons why IT leaders should evaluate their individual workloads to find the infrastructure that’s best for each.
In this blog, we’ll look at four infrastructure choices – public cloud software-as-a-service (SaaS); public cloud infrastructure-as-a-service (IaaS) and platform-as-a-service (PaaS); private cloud; and on-premises – and the importance of having the right workloads in the right place for the right reason.
Public cloud SaaS
If you are considering moving an application or business process to the public cloud, you first need to ensure it can reasonably be converted to become a high-performing cloud-based application. If it can, then there are usually good reasons to make the shift to the public cloud. Public cloud eliminates the need to manage traditional infrastructure components, freeing you up to focus on improving business processes.
Public cloud SaaS implementations have become increasingly popular. Gartner recently forecasted that public cloud SaaS services would grow by 18 per cent in 2021, driven in part by the COVID-19 pandemic. Public cloud SaaS is ideally suited to a range of back-office applications including human resources management, customer relationship management, finance and e-mail.
While public cloud has no up-front capital costs, there are ongoing operational costs, so over time, the cost factor can be on par with a traditional software offering. But for most organizations, the benefits of SaaS such as flexibility, ease of management and low up-front costs outweigh the ongoing operational expense.
Public cloud IaaS and PaaS
Public cloud IaaS and PaaS offerings provide platforms to run and manage applications without investing in any physical infrastructure. Like all public cloud solutions, IaaS and PaaS have the benefits of no up-front costs and excellent scalability. PaaS delivers a framework developers can use to build customized applications without worrying about any kind of management responsibility, while IaaS offers customers virtualized cloud infrastructure, where they can manage and monitor applications, operating systems and data. If you need more resources to run your applications, you can add them with the click of a button.
In particular, public cloud IaaS and PaaS are ideal for test/development environments. For example, a company moving over to a service such as AWS Lambda doesn’t need to worry about any infrastructure at all. As a result, its programmers can focus on developing code, with the Lambda service scaling automatically.
On-premises
Moving workloads to the cloud isn’t always simple, or even realistic for some organizations. For example, a manufacturer running a third-party application on a traditional IT infrastructure for industrial automation can’t just ask its application provider to re-architect its software to run on Azure. If you licence your software from another company and you want to continue using it, you have to rely on the infrastructure for which that software is designed.
In an instance like this, it may be more cost-effective, at least in the short-term, to continue using a legacy on-premises implementation until the business processes tied to the legacy application can be re-architected to work in a cloud environment.
Private cloud
There are also situations where shifting workloads to the public cloud is possible but isn’t necessarily the optimal choice, or where expectations can’t live up to reality. In this case, a private cloud might be a good option, where an organization owns and manages the infrastructure, and provides virtualized resources to its users.
One example would be moving a database application from a legacy environment to the public cloud. Database applications with a high input/output (I/O) requirement, where there are large volumes of reads and writes, could require a high-performance cloud environment, which can get expensive. Companies can easily find themselves spending thousands of dollars more every month to operate their databases if they don’t optimize the infrastructure.
In fact, there have been cases where we have helped migrate companies back from the public cloud to a hybrid public/private cloud solution, so workloads with high I/O run in a private cloud where their costs can be controlled, and the remaining workloads reside in the public cloud. This allows organizations to take advantage of the benefits of public cloud, while staying within their budgets.
One of our customers, an advertising technology firm, operates a self-service platform that allows buyers to purchase ads in real-time. From the start, this customer ran a hybrid cloud model. A private cloud hosts their high-performance database, which needs to be constantly optimized and tuned, and the public cloud hosts their self-service ad buying platform that needs to be accessed by clients around the world in near real-time.
Private clouds are also ideal for compliance purposes. For example, if a company is required to hold data in a particular country because of regulatory requirements, a private cloud may be the best solution.
It’s critical to take the right approach
Cloud offers clear benefits for many workloads and business processes, while there are still instances where legacy infrastructure is a better option. The decision revolves around issues like data sovereignty, security, performance, cost and control.
Legacy infrastructure, public cloud, private cloud, or a hybrid approach – it’s all about ensuring the right workloads are in the right place for the right reason.
