Transport Abstraction API for 5G-NR Resource Interoperability

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Creating interoperability between disaggregated computing resources in 5G-NR architecture is time-consuming and resource-intensive.

Innovation Solution

Implementing an application programming interface (API) that facilitates data transfer between 5G-NR computing resources by abstracting transport protocols, allowing applications to communicate with hardware accelerators without knowledge of their specific transport configurations, using transport abstraction layer APIs to manage buffer allocation and transfer.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If traditional methods are used to create interoperability between disaggregated computing resources, then connectivity can be established, but significant time and computing resources are consumed

Engineering Contradiction:
Improveinteroperability establishment speedVSAvoidtime to create interoperability
Core Design Contradiction:
ProductivityVSLoss of time

Solution Approach 1:

The patent introduces a transport abstraction layer as an intermediary between applications and hardware accelerators. This layer provides standardized API interfaces that translate application requests into hardware-specific operations, eliminating the need for direct integration between each application and each hardware accelerator. The abstraction layer handles protocol conversion, buffer management, and resource allocation automatically, dramatically reducing the time and resources required to establish interoperability.

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The transport abstraction layer implements universal API interfaces that can work with multiple types of hardware accelerators (GPUs, FPGAs, ASICs, DPUs) through a common interface. This multi-functionality allows a single application to communicate with different hardware accelerators without modification, and allows hardware accelerators to be swapped or added without changing the application code, thereby accelerating interoperability establishment.

Inventive Principle:
Principle #6Universality (Multi-functionality)

2Productivity

If direct integration is used between applications and hardware accelerators, then specific transport configurations can be optimized, but device complexity increases

Engineering Contradiction:
Improvedata transfer efficiencyVSAvoidintegration complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent segments the integration complexity into two separate layers: the transport abstraction layer handles the complex protocol translation, buffer management, and resource allocation, while the application layer focuses on high-level data processing logic. This segmentation isolates the complexity in a manageable intermediary layer, allowing applications to maintain simplicity while still achieving optimized data transfer through the abstraction layer's efficient implementation.

Inventive Principle:
Principle #1Segmentation

3Reliability

If custom integration is performed for each hardware accelerator, then specific transport protocols can be optimized, but ease of operation decreases

Engineering Contradiction:
Improvetransport protocol optimizationVSAvoidease of use
Core Design Contradiction:
ReliabilityVSEase of operation

Solution Approach 1:

The transport abstraction layer acts as an intelligent intermediary that translates standardized application requests into hardware-specific optimized operations. Applications interact with simple, uniform API interfaces while the abstraction layer automatically applies protocol-specific optimizations for different hardware accelerators. This maintains ease of operation for applications while preserving the benefits of custom protocol optimization at the implementation level.

Inventive Principle:
Principle #24Intermediary (Mediator)

Data Source

PatentUS12520232B2Application programming interface to prevent deselection of storage
Publication Date: 2026.01.06 NVIDIA CORP
  • US12520232B2 patent drawing
  • US12520232B2 patent drawing
  • US12520232B2 patent drawing

AI summary

Apparatuses, systems, and techniques to perform one or more APIs. In at least one embodiment, a processor is to perform an API to prevent deselection of storage to be used to transfer information between a plurality of fifth generation new radio (5G-NR) computing using different transport protocols.