Adoption of serverless strategies is on the rise. In fact, over 75 percent of organizations surveyed
report that they have either implemented a serverless strategy or are planning to do so in the next
two years (451 Research). Current customers are also doing more with serverless technologies—
Amazon Web Services (AWS) users ran workloads on AWS Lambda 3.5 times more in 2021 than they
did in 2019 (Datadog).
The popularity of a serverless strategy is growing because it provides the opportunity for faster
time to market by dynamically and automatically allocating compute and memory based on user
requests. For most use cases, serverless allows customers to focus on business logic and writing code
that provides customer value, as opposed to spending time on managing infrastructure, application
scaling, building service integrations, and so on. It also provides cost savings through hands-off
infrastructure management, which enables organizations to redirect IT budgets and development
resources from operations to innovation. The pay-for-use model with serverless technologies leads
to a shift from large capital expenditure lockup to flexible, on-demand consumption, allowing users
to scale, customize, and provision computing resources dynamically to meet their exact needs. This,
in turn, increases business agility.
But predicting costs in a pay-for-use model can be difficult if the inputs are variable. And
prospective customers want to optimize costs by comparing server- or virtual machine (VM)–based
computing with serverless options. In 2019, we introduced a framework for comparing the total cost
of ownership (TCO) for both serverless and server-based applications, factoring in infrastructure,
development, and maintenance costs. In that analysis, we concluded that while infrastructure
costs, often referred to as the "cost to run" the application workloads, may seem higher with a
serverless approach, the TCO is significantly lower due to savings in infrastructure, development,
and maintenance costs.
Deloitte
Industry Insight
2
