How can Arm chips like AWS Graviton or Apple M1/2 be faster and cheaper than x86 chips from Intel or AMD

Wed, 25 Oct 2023 00:00:00 +0000

The Arm price/performance paradox

In my cost optimization gigs I often see how many people don’t understand how Arm chips (such as AWS Graviton and Apple M1/M2) can be both cheaper and faster than x86 chips from Intel and AMD.

It’s counter-intuitive, so many think that since Graviton is ~20% cheaper, it must also be slower.

To explain this paradox, imagine two development teams:

Team x86 has 4 engineers and 6 managers
Team Arm has 6 engineers and 2 managers

That’s how you can get 50% more output at 20% less costs.

Why I recommended ECS instead of Kubernetes to my latest customer

Thu, 08 Jun 2023 00:00:00 +0000

And how a cost optimization exercise often leads to deeper modernization of cloud applications

June 08, 2023

It’s been a while since my last post here, and since then I’ve been pretty busy. I’ve been working with an innovative AI startup, helping them optimize their AWS cloud setup. At the same time I’ve been preparing an Udemy course on using ChatGPT for cloud-native software development with a DevOps focus.

In this post I’ll talk about my work with this startup, my thought process and what I recommended them, including the potentially controversial recommendation mentioned in the title.

Cost-Optimization on AutoSpotting

How can Arm chips like AWS Graviton or Apple M1/2 be faster and cheaper than x86 chips from Intel or AMD

The Arm price/performance paradox

Why I recommended ECS instead of Kubernetes to my latest customer

And how a cost optimization exercise often leads to deeper modernization of cloud applications