Serverless computing’s second act

Serverless computing has had a tumultuous journey characterized by early excitement, skepticism, and challenges that some—including yours truly—believed might spell its end. However, this technology is now experiencing somewhat of a resurgence, or perhaps it is more of a forced renaissance. In other words, change or die. Perhaps it can accomplish both.

Let’s back up a bit. In the initial stages, serverless computing was embraced for its promise to revolutionize cloud application development. It allowed developers to focus on writing code without the concern of managing or provisioning servers. This streamlined operations and reduced costs.

Rocky road

Despite the enthusiasm, several challenges emerged. Among these were unpredictable pricing models, latency issues referred to as “cold starts,” and limited control over the underlying infrastructure. These challenges led to skepticism about the practicality of serverless in production environments.

Despite these issues, the hype train was at full speed. For example, a large provider took issue with me pointing out some of the shortcomings of this “new” serverless technology. Instead of addressing the problems, they called for my immediate firing due to blasphemous comments. I hit a nerve. Why was that? The cloud providers promoting serverless should have had more confidence in their technology. They knew the challenges. I was right about serverless then, right when I wrote about its decline. However, I’m always willing to reevaluate my position as technology evolves. I believe in redemption.

A technological comeback

Despite its early hurdles, serverless computing has bounced back, driven by a confluence of evolving developer needs and technological advancements. Major cloud providers such as AWS, Microsoft Azure, and Google Cloud have poured substantial resources into serverless technologies to provide enhancements that address earlier criticisms.

For instance, improvements in debugging tools, better handling of cold starts, and new monitoring capabilities are now part of the serverless ecosystem. Additionally, integrating artificial intelligence and machine learning promises to expand the possibilities of serverless applications, making them seem more innovative and responsive.

The serverless market is projected to grow from $7.6 billion in 2020 to $21.1 billion by 2025 at a compound annual growth rate of 22.7% Companies are leveraging its architecture for more efficient and scalable applications. Case in point: Coca-Cola reportedly achieved a 65% reduction in operational costs by migrating to AWS Lambda. Netflix has cited improved scalability and reduced complexity by adopting serverless for back-end processes.

In this renaissance period, serverless computing aligns better with modern application development practices emphasizing rapid iteration and scalability. This alignment may reinforce its adoption across industries, reshaping how developers approach building applications. Some are calling it a “new era for cloud computing.” Maybe.

Too little, too late?

One crucial question remains: Is this resurgence enough to secure the future of serverless computing, or is it simply an attempt by cloud providers to recoup their significant investments? At issue is the number of enterprises that have invested in serverless application development. As you know, this investment goes beyond just paying for the serverless technology. Localizing your applications using this tech and moving to other platforms is costly.

A temporary fix might not suffice in the long run. While the current trends and forecasts are promising, the final verdict will largely depend on how serverless can overcome past weaknesses and adapt to emerging technological landscapes and enterprise needs. A forced renaissance might be a strategic maneuver to reinvigorate serverless, but whether it will lead to enduring success or merely be a fleeting revival remains to be seen.

What should you do?

If you are already leveraging serverless on any public cloud, you’ll find that changing directions now is risky and costly, and I don’t recommend you do it just yet. However, I would not use serverless for any net-new development projects due to the risk I’ve seen.

I do suspect that serverless will get better—it has to if the public cloud providers want to avoid a very public meltdown as enterprises hit that “serverless wall” as I’ve seen so many times. Public cloud providers are good at controlling the narrative around specific technologies. Although they have not been able to intimidate me—yet—they have billions of marketing dollars to tell the market what to think about their technology, including serverless. Pundits like me don’t have access to those kinds of resources.

So, if you’re already running serverless, I suggest you continue and provide alternatives. But for now, and attempting to be fair to the serverless technology providers, it may be best to avoid the new stuff. Let’s see if this gets me canceled.