The OPI project aims to standardize DPUs and IPUs for industrial adoption

In recent years, silicon vendors have developed new types of computing architectures that go beyond CPUs and GPUs – welcome to the world of data processing units (DPUs) and infrastructure processing units (IPUs).

The goal of DPUs and IPUs is to enable organizations to offload certain data and cryptography and artificial intelligence/machine language (AI/ML) tasks to dedicated hardware to speed up operations. To date, there have been few standards around DPUs and IPUs to enable interoperability or industry standardization for deployment, management, and scheduling, but that’s about to change.

Today, the Linux Foundation announced the launch of the Open Programmable Infrastructure Project, which aims to gather open source efforts around DPUs and IPUs and organize vendors to drive adoption for businesses of all sizes. Founding members of the Open Programmable Infrastructure (OPI) project include Intel, Nvidia, Marvell, F5, Red Hat, Dell and Keysight Technologies.

The first projects to become part of the OPI include the IPDK (Infrastructure Programmer Development Kit) being developed by Intel and the Diamond Bluff project being created by Red Hat and F5.

“Red Hat and F5 and a few other companies worked on the Diamond Bluff project and we [Intel] started talking to them and we soon realized that we had similar goals and should work together,” Kyle Mestery, senior principal engineer at Intel, told VentureBeat. “The goal of the Open Programmable Infrastructure Project is to foster a community of open standards and open source projects around these next-generation architectures, which include DPUs and IPUs.”

The market for DPUs and IPUs is new but growing

Both IPUs and DPUs have found their way into the architectures and deployments of hyperscalers and cloud providers in recent years.

The opportunity for the hyperscalers and cloud providers was to provide a more granular level of services for users to consume. There’s also potential for IPUs and DPUs to help enterprise users as well, which is one of the goals Mestery said the OPI will achieve over the coming months and years.

Intel has in recent years developed Infrastructure Processing Units (IPUs) as a form of silicon hardware technology utilizing Intel’s Field Programmable Gate Arrays (FPGA) and Application Specific Integrated Circuits (ASIC) IPU technology.

Nvidia has also been active in this space, expanding its BlueField DPUs, which focus on accelerating analytics and AI workloads. Marvell is also in the market with its Octeon DPU technology, which can be used to support 5G data workloads.

How to manage DPUs and IPUs at scale

Both the IPDK and Diamond Bluff projects involve examining various aspects of how an organization can deploy IPUs and DPUs in a data center deployment.

The OPI project also talks about the API (Application Programming Interface) layer for IPU and DPU management to help organizations take a unified approach to visibility and control. However, Mestery emphasized that the OPI does not seek to reinvent technologies and will collaborate with other Linux Foundation initiatives, including the LF Networking project and the Cloud Native Computing Foundation (CNCF).

As an example, Mestery cited CNCF’s OpenTelemetry project. The aim is to collect observability data on ongoing operations. He noted that when an OPI open source project needed telemetry data, it made sense to use OpenTelemetry rather than create something new.

As the OPI ramps up, Mestery said he’s confident the open-source effort will be able to demonstrate the benefits of DPU and IPU technology. As a goal, he wants the project to make it easier for organizations to deploy DPU and IPU technologies in their own private data centers and use them for a range of purposes, be it edge computing, private cloud or whatever you want to do with it .

“I hope that we can provide a framework that makes it easier for organizations to use these technologies,” said Mestery.