Introduction

Recent discussions at the AWS Community Day DACH event have sparked some thought-provoking ideas about the future of cloud computing and serverless architectures. This post aims to challenge our current paradigms and explore potential shifts in the industry.

The Power of Paradigm Shifts

In her book "Thinking in Systems: A Primer," Donella Meadows argues that the two most powerful ways to change a system are:

1. Transcending Paradigms: Recognizing that no paradigm is perfect.

2. Shifting Paradigms: Changing the mindset from which a system's goals, structure, rules, delays, and parameters arise.

With this framework in mind, let's consider what might come after serverless computing.

The Current Serverless Paradigm: Serverless computing offers numerous benefits: enhanced security, managed infrastructure, cost-effectiveness, and the ability to focus on business differentiators. However, it's not without its challenges. [I’m intentionally not going into details. I believe that Pay as You Go is the best model now.] Many AWS users, especially those new to cloud computing, have experienced the shock of unexpected bills. Personal experiences, like a $500 charge for AWS SiteWise Alarms PoC, highlight the potential pitfalls of serverless architectures. More pragmatic - you need to know when it is better to replace Serverless with Serverfull.

Emerging Trends Shaping the Future

Several technological advancements are poised to reshape our world:

• Renewable energy sources

• Mini nuclear reactors

• Local power grids and power walls

• Generative AI and AI Agents

• Advanced database systems (e.g., PostgreSQL with expanded capabilities)

• Ubiquitous high-speed internet (fiber, 6G, low-orbit satellites)

A New Paradigm

On-Premises Private "Cloud"

Imagine a world where the "managed" portion of cloud Total Cost of Ownership (TCO) is eliminated. AI Agents could manage Kubernetes clusters with operators 24/7, potentially bringing compute power back on-premises. This scenario doesn't eliminate the need for human expertise but shifts the focus to creative and strategic tasks.

How close are we to this potential future? With the EU and USA at around 20% renewable energy adoption, AI Agents already in use, and Kubernetes becoming the standard for container orchestration (84% of respondents using it in production according to recent surveys), this future may be closer than we think.

The Benefits and Challenges

While this shift might seem like a return to CAPEX-heavy models, it's more nuanced. Organizations would still have OPEX costs for subscriptions to hardened OS, infrastructure management tools, and AI Agents. The primary benefits include:

1. More predictable OPEX costs

2. Improved data processing capabilities, especially in terms of latency

3. Better support for edge computing and IoT applications

The Transition

For those just beginning their cloud journey, don't be discouraged. The skills and mindset developed through working with serverless architectures and Kubernetes – such as event-driven thinking and domain-driven design – will remain valuable in future iterations of technology.

I don’t know how long it will take for AI Agents to be trusted to manage your local private clouds—without this key ingredient, Public clouds will dominate.

Conclusion

As we consider the potential return of compute power to on-premises environments, it's crucial to remember that people remain our greatest asset. The ability to adapt to new paradigms and technologies will be key to success in this ever-evolving landscape.

What are your thoughts? Do you believe we'll see a shift back to on-premises computing? How might this transition benefit or challenge organizations?

Cover image generated by generative AI.