Red Hat and RISC-V: To the far edge and beyond

Red Hat has always been an advocate of growth at the intersection of open source and computing solutions–which is exactly where RISC-V can be found. RISC-V is one of those technologies where the future is both evident and inevitable. By integrating open source concepts with the hardware development process, it’s not hyperbole to say that RISC-V is disrupting the hardware industry.

Our excitement around the unique value RISC-V brings to the hardware ecosystem as an open and collaborative instruction set architecture (ISA) is nothing new. Red Hat has been providing Fedora on RISC-V for several years, collaborating with universities on interesting RISC-V based research projects, writing about RISC-V since at least 2018, working within RISC-V International particularly in the platforms & profiles working group and most recently joined the RISE Project as a founding member. We continue to identify new opportunities for RISC-V, especially within edge, artificial intelligence and next-generation infrastructure. As for the business value of RISC-V for enterprises at large, I wanted to share some takeaways from a keynote I gave at the recent RISC-V Summit in Santa Clara.

Far edge

Moving data from remote locations to a centralized compute fabric for processing can be prohibitively expensive and technically challenging. This becomes even more of a challenge if the data’s insights are vital to providing services back at the remote location, such as improving mean time to insight or reducing cost and latency. These challenges encourage enterprises to process data as close as possible to its source. A major key challenge involves developing the capabilities to deploy and maintain applications on use-case optimized hardware and far edge devices.

Additionally, the ability to orchestrate data collection, storage and analysis is becoming more complex. Increasingly, these capabilities are scattered across a wide variety of connected, partially connected and disconnected environments, with varying requirements to own and/or operate the related infrastructure. This is further compounded by having to meet geographic, regulatory and digital sovereignty constraints which preclude traditional approaches such as centralized data lakes.

We believe that RISC-V is very well suited to address these myriad challenges, from hardware flexibility and performance per watt to supporting a growing software ecosystem, both as accelerators in heterogeneous systems and as apps processors driving the workloads.

Artificial intelligence

Given global macroeconomic conditions, there is industry-wide interest in leveraging machine learning (ML) models to augment human developers and ops engineers. Layering these models into existing workflows are aimed at improving overall productivity, increasing operational reliability and faster time to development. As the role for ML expands, there is a need to deploy, maintain and enhance the security of those models in conjunction with the application development pipelines that use them.

RISC-V is emerging as a primary source for AI/ML acceleration in a number of heterogenous and homogenous solutions. While Red Hat is famously a Linux platform company rather than a developer of individual solutions, we see very strong value in understanding the solutions that our customers implement and work to support them in that journey. Thus, we feel it’s imperative that we develop support for RISC-V as accelerators as well as primary CPUs to meet our customers’ current and potential future needs.

Next-generation infrastructure

We’re also watching a revolution in next generation infrastructure, which is already a very quickly evolving area of technology. While we’re just now seeing ARM architectures taking a starring role in datacenters, we believe that RISC-V is not too far behind–and with far edge and many devices in the middle, RISC-V is already here in the marketplace.

Increasingly, the computational needs of applications require more than industry standard servers, whether in the data center or on the far edge. To address this dynamic, we’re integrating support into our platforms to more seamlessly integrate and run workloads across emerging, software defined, heterogeneous compute infrastructures, such as CXL, DPUs, and of course RISC-V.

To meet all of these opportunities, Red Hat is proud to be a founding member of the RISE Project supporting the growing RISC-V software ecosystem. RISE is a great way for organizations like Red Hat to help the broader ecosystem work upstream in the important projects and communities that will need to support RISC-V as it grows.

Red Hat has always been at the forefront supporting open technologies, and it is hardly possible for a computing architecture to be more open than RISC-V. We look forward to continuing our trajectory with this innovative architecture and the ecosystem around it.