IP Library Granted Patent US 12,483,623
Granted Patent B2
US 12,483,623 · App. 18/589,307 · Granted Nov 25, 2025

Remote direct memory access for real-time control applications

Inventors: Markus Jochim (Troy, MI); Khaja Shazzad (Windsor, CA); Sudhakaran Maydiga (Troy, MI)
Assignee: GM Global Technology Operations LLC
H04L67/1097H04L47/22
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 12,483,623
App. No.
18/589,307
Granted
Nov 25, 2025
Kind
B2
Abstract

In an aspect, systems and methods of connected nodes, including electronic control units (ECUs) in a vehicle, integrate remote direct memory access (RDMA) capabilities with time-sensitive networking (TSN) traffic shaper configurations in a manner that guarantees lossless, bounded-latency critical traffic (CT) streams, while also allowing for different classes of traffic to flow through the connected nodes. In another aspect, mechanisms are introduced to enable an automotive software framework (e.g., classic AUTOSAR) to support RDMA communications.

Claims (43)

1 . A vehicle, comprising:

a frame including a body, the body and frame defining a cabin and areas for placement of functional vehicle components; and

a network of electronic control units (ECUs) in the areas and selectively coupled together via conducting elements, at least some of the conducting elements coupled to respective switches for controlling data flow between identified ECUs of the plurality;

wherein the ECUs comprise respective network interface cards (NICs) having remote direct memory access (RDMA) hardware for generating messages, the NICs being further configured to execute traffic shapers to ensure that specified critical transmissions (CT) comprise lossless RDMA data exchanges characterized by bounded latencies.

2 . The vehicle of claim 1 , wherein the ECUs are configured to perform RDMA operations using one or more of a plurality of classes of data.

3 . The vehicle of claim 2 , wherein the plurality of classes of data include CT and best effort (BE) traffic.

4 . The vehicle of claim 3 , wherein the NICs further comprise ingress and egress ports, the ingress and egress ports having respective ingress and egress queues for each of the plurality of classes of data.

5 . The vehicle of claim 1 , wherein the NICs include at least one egress port, the egress port having the egress queues for transmitting the CT and the BE traffic.

6 . The vehicle of claim 5 , wherein respective switches comprise ingress and egress ports, each of the egress ports having the egress queues.

7 . The vehicle of claim 6 , wherein, for at least one of a plurality of transmission paths of the network:

the traffic shapers comprise asynchronous traffic shapers (ATSs);

a processing system in the NICs is operable to configure the egress queues of the corresponding NICs and the at least one switch in the at least one transmission path through which the CT flows to be managed by the ATSs;

the processing system is operable to configure the egress queues through which the BE traffic flows in the at least one transmission path of the plurality of transmission paths as strict priority without traffic shaping; and

the processing system is operable to calculate, for each egress queue in the at least one transmission path, a committed information rate (CIR) and a committed burst size (CBS), and to use the CIR and CBS with the asynchronous traffic shapers to guarantee the lossless RDMA operations and respective bounded latencies for the transmission path.

8 . The vehicle of claim 7 , wherein the processing system is further operable to:

configure the egress queues in the egress ports of the NICs for the plurality of transmission paths of the network through which the CT flows; and

determine each of the respective CIR and CBS values for each of the plurality of transmission paths to determine a bounded WCL.

9 . The vehicle of claim 8 , wherein the processing system is configured to determine the CIR and CBS values during initial system design.

10 . A system, comprising:

a network comprising nodes, the nodes having respective network interface cards (NICs), at least some of the nodes being configured with remote direct access memory (RDMA), each NIC having an egress port, each of the egress ports having a plurality of egress queues, each of the egress queues corresponding to a class of traffic;

a processing system in the NICs to identify RDMA transmission paths which include a periodic flow of critical traffic (CT) across intervals;

for each of the RDMA transmission paths, the processing system is operable to:

identify all nodes from or through which the CT is transmitted;

configure, for each egress port, queues carrying the CT to be traffic-shaped to guarantee a lossless bounded latency for each successive interval; and

calculate committed information rate (CIR) and committed burst rate (CBR) values relevant to an egress port of a sink node in the transmission path to ensure a lossless, bounded-latency RDMA CT transmission at the sink node.

11 . The system of claim 10 , wherein the processing system is further operable to:

configure, for each egress port, any egress queues carrying lower priority data streams.

12 . The system of claim 11 , wherein the processing system is configured to calculate the worst case latencies (WCLs) of each frame of the RDMA CTs in each of the transmission paths to identify the bounded latency for the CT frame.

13 . The system of claim 12 , wherein the network further comprises switches having egress ports, each of the egress ports having egress queues.

14 . The system of claim 13 , wherein the processing system is configured to:

determine the WCLs for a periodic frame of CTs transmitted along all applicable transmission paths to the sink node based at least in part on a switch delay, a transmission size, a number of transmissions per interval, a long frame transmission time, and a maximum size frame of non-CT traffic preceding the frame of the CT; and

identify a bounded WCL for the CT using the WCLs for the periodic frames sent to the sink node across each of the transmission paths.

15 . The system of claim 10 , wherein the network includes transmission of best effort (BE) traffic.

16 . The system of claim 15 , wherein the NICs includes at least one egress port, the egress port having at least two egress queues for transmitting the CT and the BE traffic, respectively.

17 . A system for enabling an automotive software framework to support remote direct memory access (RDMA) communication, comprising:

a plurality of networked electronic control units (ECUs), each ECU comprising:

an application layer comprising a plurality of control algorithms;

an ECU hardware layer comprising a network interface card (NIC) and one or more microcontroller units (MCUs), wherein the NIC comprises one or more traffic shapers and a physical memory for sending an RDMA message to a second ECU,

wherein the data is stored in the physical memory of the second ECU; and

wherein the data is received at one of the control algorithms using an RDMA data channel on the second ECU, the RDMA data channel coupled respectively to the RDMA write and read modules on the second ECU.

18 . The system of claim 17 , where the ECUs are configured to guarantee a lossless bounded latency for a critical transmission (CT).

19 . The system of claim 17 , wherein the automotive software framework comprises AUTOSAR.

20 . The system of claim 17 , wherein the one or more traffic shapers eliminate a need for an RDMA over Converged Ethernet (ROCE) NIC protocol in the NICs of the ECUs.

Assignments (1)
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Feb 28, 2024
From: JOCHIM, MARKUS; SHAZZAD, KHAJA; MAYDIGA, SUDHAKARAN
To: GM GLOBAL TECHNOLOGY OPERATIONS LLC
Reel/Frame 066584/0438 →
Continuity (1)
Related Publication 20250274512A1 · Aug 28, 2025
References Cited (18)
US 7912641B2 · Osentoski · 2011 [cited by examiner]
US 10678243B2 · Luo · 2020 [cited by examiner]
US 11038802B2 · Yamasaki · 2021 [cited by examiner]
US 11107097B2 · Kawashima · 2021 [cited by examiner]
US 11156462B2 · Zhang · 2021 [cited by examiner]
US 11238160B2 · Kallenberg · 2022 [cited by examiner]
US 11499830B2 · Zhang · 2022 [cited by examiner]
US 11505114B2 · Takori · 2022 [cited by examiner]
US 20070294033A1 · Osentoski · 2007 [cited by examiner]
US 20130268165A1 · Hashima · 2013 [cited by examiner]
US 20190089636A1 · Yamasaki · 2019 [cited by examiner]
US 20200378765A1 · Zhang · 2020 [cited by examiner]
US 20200394668A1 · Zhang · 2020 [cited by examiner]
US 20210065224A1 · Kawashima · 2021 [cited by examiner]
US 20210293545A1 · Zhang · 2021 [cited by examiner]
US 20210300235A1 · Takori · 2021 [cited by examiner]
US 20210335060A1 · Bauer · 2021 [cited by examiner]
US 20230096468A1 · Ong et al. · 2023 [cited by applicant]