Scale-to-zero and pay-per-use have been the main drivers for many organizations adopting serverless. Given the premise, the motivation behind everyone going serverless is to reduce the cost of operation. Almost every serverless success story revolves around the cost factor.
Misguided Cost Comparisons
There was a time when I used to compare the cost of EC2 instances against Lambda functions. It was interesting at the beginning.

Soon I realized it was misleading. It just didn’t feel right to me. I thought I wasn’t doing justice either to the servers or to serverless.
- Am I comparing the bare metals?
- Have I considered the license fee of those applications running on EC2?
- How about the cost of EC2 instances in the second Availability Zone?
- Serverless is not just Lambda functions. Have I added the cost of other services?
- How do I explain the cost in a hybrid situation?
Questions like the above compelled me to stop doing the numbers game, for good.
Per Feature Cost Estimation
By now, I have developed the serverless-first mindset. Rather than wasting time comparing apples and oranges, I turned my attention towards estimating the cost of application features based on their high-level architecture. Though not accurate, this gave me insights into the cost of those features that we were building.
As an engineer, it certainly helped me to understand the influencers of serverless cost. Importantly, it allowed me to share those numbers with the business stakeholders. Having the cost visibility proved valuable as we were moving ahead with serverless.
This approach also had its challenges.
- Estimation became difficult with complex architectures
- Accommodating the impact of unpredictable traffic spikes proved difficult
- It was tricky to add up indirect costs such as CloudWatch metrics
Nevertheless, it raised the serverless cost awareness within the team, which is vital for everyone developing serverless solutions. Even with some margin for error, the numbers aligned with the low-cost expectation everyone had on serverless.
How low is that low-cost would then become one of the mysteries of serverless!
Balancing Cost And Performance
Cost reduction is an important exercise. Often, it is driven from the top and trickles down. When it lands on the serverless ecosystem, it encounters an ambiguous activity known as optimization. It exists in two different forms.
- Cost optimization, and
- Performance optimization
In serverless, both cost and performance intertwine. The problem is, when we look at ways of reducing cost in serverless, these two can pull us in opposite direction.
Clash between cost and performance
As we all know, every application has many parts – microservices, APIs, storage, features, etc. It depends on how we identify and group them.
One of the benefits of a serverless application is that we can visualize and optimize each part at a granular level. It allows us to perform both cost and performance optimization at a fine-grained level. However, the complexity is that both cost and performance don’t always complement each other. They often cause cost imbalance!
- Cost optimization does not guarantee to be performant, and
- Performance optimization is not always cost-effective
Performance Over Cost
The low-cost perception of serverless can dramatically change when we optimize for performance. Here are a few cases that assert this claim.
- Setting Provisioned Concurrency for Lambda functions can be expensive. However, for better performance and end-user experience, in cases where needed, we accept to pay the premium

- To maintain high availability, we use cross-region data replication using DynamoDB Global Tables. It adds a higher price to our DynamoDB cost

- Engineering secure solutions that have better monitoring and observability is a business priority. It can incur a higher cost in the use of CloudWatch Metrics

Security, data protection, low latency, high availability, resiliency, and observability are a few core aspects of a serverless application that take precedence over cost optimization. End-user experience and satisfaction is another area where no business will be willing to sacrifice for the sake of cutting costs.
In short, better engineered serverless solutions become the priority.
Towards DevOps
Building serverless applications indeed demand a different way of thinking from how we conceptualized applications in the past. Building secure and performant solutions are as important as implementing the best business logic using the best compute model.
- Gone are the days when engineers spent months to produce a detailed design specification and handed it over to an off-shore team to code.
- Gone are the days when engineers wrote the code, compiled it, and threw it over the fence to other teams for the rest.
- Gone are the days when companies kept a demotivated maintenance team for fixing production bugs.
- Gone are also the days when single-purpose software engineers only knew how to code in a particular programming language.
Welcome to DevOps!
In contrast to the past, engineers of this age, while writing single-purpose Lambda functions, have evolved to be multi-resourceful multifaceted monoliths, capable of doing many things!
Bringing Business and Technology Closer
The DevOps movement is certainly narrowing the gap between engineering and operations. It results in smaller all-encompassing teams owning features from requirements to production and beyond.
These teams follow agile and lean principles to become efficient and cost-effective. It helps to build stakeholder trust in the engineering teams. Trust brings confidence in the teams to deliver features faster to satisfy the ever-changing customer demands.
New world new demands
Building solutions for our modern world has many demands.
- Applications expected to withstand brutal cyber-attacks.
- Businesses are under extreme pressure to comply with regulatory and industry standards to protect personal data.
- A customer’s focus changes every few seconds. Keeping them engaged and serving them faster is crucial.
Across the enterprise world, traditional IT departments get dispersed in favor of autonomous product teams. It allows technical and non-technical colleagues to sit together and build better solutions. In turn, it helps to develop respect and strengthen mutual understanding.
Serverless the catalyst
What serverless brings to this is the agility and the convenience to satisfy such demands. Features get designed, developed, deployed, and improved faster. The software feature life cycle is spinning faster than ever before, every day. Serverless is one technology that makes it easier to keep up.
This business and technology collaboration increases business value. Customer focussed, fast-paced, and feature-driven development makes the business thrive.
Thanks to serverless, the acceleration just gets accelerated.
The Pandemic Effect
Year 2020!
While it remains in the history books for the pandemic reason, it has taught harsh lessons and injected new ideas into the entrepreneurial think tanks.
- Many organizations changed their mode of operation to survive.
- Many caught in the confusion and only had a glimmer of hope and a limited time to react.
- For others, it was a choice between keep-alive or go-burst.
Things that we thought impossible became possible. Business priorities changed every other day. Those who swiftly adapted withstood. Those who had their business built on scalable infrastructure quickly adjusted as per the unpredictable customer demand.
Business growth during locked-down
A recent survey in the UK by the BBC shows a significant increase in the launch of new business during the pandemic. It may surprise a few, but from the modern cloud technology perspective, this is a possibility.

Amidst the chaos, heartache, and confusion, everyone needed a technology that they could use to build solutions faster and manage from their living rooms. For the majority who succeeded, their cloud and serverless adoption must have played a crucial role.
This business model and the value differentiation with serverless will continue to strengthen in 2021 and beyond.
The convenience of building with serverless will start to break free from the clutches of its cost thinking.
The Emergence of Convenience Computing
AWS Lambda is the core of serverless computing. Lambda has changed a lot since its birth. It is still evolving. It has undoubtedly become sophisticated with longer execution duration, more memory, higher processors, heavier workloads, etc. Yet, it helps to write simple functions and comfortably build serverless solutions. If serverless is all about low-cost, then Lambda service wouldn’t have evolved this far.
Being the cheapest is not always the prime reason behind many technology adoptions. Requirements change with time. These days,
- Businesses are looking for the easiest way to run their compute workloads
- Product owners are looking for the fastest way to release a feature
- Engineers are looking for the safest way to handle customer data
- Customers are looking for the simplest way to purchase a product
The list grows. Teams don’t always place cost as a blocker while trying to meet such demands. More often than not, they choose the convenient way to operate to be competitive and successful. When it comes to a technology that offers that level of comfort, Serverless is the clear winner.
The BBC survey I referred to earlier also highlighted the following.
- Running a business from home made sense during the pandemic
- Companies did well that focused on online
- Not all of these businesses will be permanent
Traditional business models are continuously getting rewritten. Scaling up and down as required has become the normal mode of operation.
Temporary and stopgap businesses are common these days. Many online businesses spring into action or go into hibernation automatically based on trading conditions and business constraints. Modern cloud and serverless technologies make such things possible.
We can say this is the new Convenience Cloud Model. With serverless, this is guaranteed from the start.
Conclusion
Serverless has successfully won many businesses over within its relatively shorter existence. As a technology, several use cases have proven the versatility of serverless. A plethora of new use cases is being built with confidence and comfort every day. Its cost centered typecasting is changing.
The challenge, however, is to maintain its simplicity amidst all the sophistication. The simplicity is what makes it convenient for everyone to build solutions faster.
Perhaps it is time for a new tagline for serverless. Let Convenience Computing be that one as serverless confidently marches ahead in the right direction into the future!