Who is Open Programmable Infrastructure Project?
Open Programmable Infrastructure Project is an open collaboration initiative focused on standardizing and enabling programmable infrastructure for data centers, cloud environments, and networking platforms.
- Vendor-neutral community for programmable infrastructure architectures and frameworks.
- Focus on data processing units (DPUs), smartNICs, and related offload technologies (infrastructure acceleration).
- Development of common APIs, abstractions, and interfaces for offload-capable infrastructure devices (infrastructure programmability).
- Reference implementations, open-source components, and integration blueprints for enterprise and cloud environments.
- Collaboration among hardware vendors, software providers, and operators on open specifications for infrastructure offload and control planes.
Show more
More About Open Programmable Infrastructure Project
Open Programmable Infrastructure Project (OPI) focuses on a community and specification layer for programmable infrastructure devices that offload networking, security, and storage functions from host CPUs in data centers and cloud environments.
OPI targets data processing units (DPUs), infrastructure processing units (IPUs), smartNICs, and similar offload hardware (infrastructure acceleration), aiming to provide a consistent open framework for how these devices are programmed, managed, and integrated with existing infrastructure stacks.
The project promotes common abstractions and APIs (infrastructure programmability) so that operators, independent software vendors, and hardware vendors can align on interoperable control and data plane models rather than proprietary, device-specific integration approaches.
In enterprise and service provider environments, OPI’s work is relevant where organizations deploy offload devices to handle networking, observability, firewalling, storage virtualization, or crypto offload functions while seeking to retain a standard operational model across different hardware platforms.
OPI architectures and discussions typically address interaction with Linux-based host operating systems, Kubernetes-centric cloud-native stacks, and standard networking protocols, as well as integration with orchestration and lifecycle management tools that manage servers and accelerators as a combined infrastructure resource.
From an enterprise IT category perspective, OPI aligns with infrastructure acceleration, network and security offload, and cloud infrastructure management, where the goal is to treat DPUs and smartNICs as programmable infrastructure elements with clearly defined northbound and southbound interfaces.
Technical topics associated with OPI include gRPC or similar RPC mechanisms for control-plane communication, standard interfaces for telemetry and statistics export (observability), and models for representing network functions running on offload devices in relation to the host’s virtual machines, containers, or bare-metal workloads.
By defining shared models, APIs, and reference implementations, OPI provides a basis for ecosystem participants to build interoperable solutions across multiple silicon implementations and vendor platforms without tying enterprise operators to a single proprietary management or programmability stack.
Within a directory or marketplace taxonomy, Open Programmable Infrastructure Project is best placed under infrastructure acceleration, cloud and data center networking, and infrastructure programmability frameworks, with relevance to network offload, security offload, and DCO tooling that interacts with DPUs and smartNICs.
Our description of Open Programmable Infrastructure Project. Updated January 2026.
