Introduction

The DOCA Pipeline Language Runtime Service (DPL RT Service) is an NVIDIA® BlueField® service that implements the backend functionality to manage and program the DPU datapath.

This DPL RT Service includes a P4Runtime gRPC based server, that binds the P4Runtime protobuf interface to the underlying hardware driver APIs. The P4Runtime server listens on TCP port 9559, which is the port that has been allocated by IANA for the P4Runtime service. The server allows a P4 Controller to connect over gRPC so that it can set the ForwardingPipelineConfig, which installs and loads into hardware the compiled DPL program output and the associated P4Info metadata. Furthermore, the controller can query the target for the ForwardingPipelineConfig to retrieve the device config and the P4Info, as well as performing P4 table maintenance (which were defined in the P4 program source code).

Installation and Getting Started

All requirements and preparation steps to enable the DPL RT Service is explained in the DOCA Pipeline Language Services Guide.

High Level Steps include:

1. DPU Requirements to enable the DPL RT Service:

• Enable BlueField DPU Mode.

2. DPL RT Service Container Deployment:

• Downloading the service Resource file.
• Running DPL RT Service DPU preparation script.
• Spawning DPL RT Service Container.
• Verifying DPL RT Service Container.

These are covered under "Container Deployment" section of Service Guide.

Documentation

The DPL RT Service Guide is available here.
The DPL RT Service Release Notes is available here.

DPL RT Service uses a configuration file for defining DPL Device Ports. Details and configurations examples are provided under "Service Configuration" section of Service Guide.

License & EULA

DOCA is licensed under the NVIDIA DOCA License. By pulling and using the container, you accept the terms and conditions of this license.

Technical Support

Use the NVIDIA Developers forum for questions regarding this software.