As a Senior Software Engineer at Kloudfuse, my focus has been on optimizing our database and log ingestion layers. I have a strong background in big data analytics and distributed databases, which helps me keep our systems running at peak performance.

Today, I’m excited to share the details behind our latest release—Kloudfuse 3.0—which introduces support for ARM-based instances.

Why ARM?

The primary advantage of running ARM servers is their cost-effectiveness combined with efficient performance. On average, ARM servers can offer 20-40% lower costs than equivalent x86 servers, primarily due to their reduced power consumption and energy efficiency. These savings are particularly valuable for organizations with large-scale observability needs, allowing them to achieve operational savings while building high scale observability with Kloudfuse.

For Kloudfuse 3.0, I added support for ARM architecture, ensuring that our core logic and functionality remain consistent across different CPU architectures—without performance penalties or user experience degradation.

This capability addresses the need for flexibility in observability. By supporting ARM, we enable companies already deploying on this architecture—and with expertise in it—to seamlessly integrate Kloudfuse into their existing infrastructure.

Tackling the Challenge

The biggest challenge in building this capability was adapting our product, originally built solely for x86, to ARM’s unique architecture. For example, a single ARM vCPU is typically a dedicated core, unlike x86 cloud offerings where it may be a hyper-threaded or shared core. This dedicated core configuration provides improved performance in multithreaded scenarios. 

Additionally, the ARM compilation process required close attention; the compiler handles certain data types differently across architectures and generates distinct instructions. To ensure the ARM machine code met our performance standards, we had to carefully oversee the compilation process to make sure it was as optimized and efficient as the x86 code.

My Favorite Part: Getting Creative with ARM Optimization

My favorite aspect of building this capability has been the opportunity to write optimized code and leverage dedicated hardware features unique to the ARM architecture. Utilizing specific CPU 'tricks,' such as SIMD (Single Instruction, Multiple Data) instructions and advanced caching techniques, allows us to enhance performance and efficiency in ways that aren’t possible with more generic code. 

It’s incredibly rewarding to see how these optimizations can lead to significant improvements in processing speed and resource utilization, especially in high-demand environments. This hands-on approach not only deepens my understanding of the hardware but also allows me to deliver better results for our users.

Looking Ahead

The introduction of ARM support in Kloudfuse 3.0 represents a significant step forward in our commitment to providing flexible, efficient, and high-performance observability solutions. As we move forward with Kloudfuse 3.0, I’m eager to see how our customers leverage these ARM capabilities. We’re currently working with at least one customer already running on ARM, and I’m excited to verify their experience and gather feedback on this new architecture.