# US 2025/0156340 A1 · Techniques for reducing security risks when implementing remote direct memory access through a proxy node

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2025/0156340 A1
- Document ID: US20250156340A1
- Status: Published application
- Kind: A1
- Application number: 18/943,141
- Filed: 2024-11-11
- Published: 2025-05-15
- Earliest claimed: 2023-11-11
- Assignees: NVIDIA Corporation
- Applicants: Nvidia Corporation
- Inventors: Chris J. NEWBURN; Eliav BAR-ILAN; Oren DUER; Gal SHALOM; Vishwanath VENKATESAN; Aviad YEHEZKEL; Yorey ZACK; Kiran Kumar MODUKURI
- Web page: https://patentgrep.com/patent/US20250156340A1

## CPC

- G06F 3/0611: {in relation to response time}
- G06F 3/0659: {Command handling arrangements, e.g. command buffers, queues, command scheduling}
- G06F 3/067: {Distributed or networked storage systems, e.g. storage area networks [SAN], network attached storage [NAS]}
- G06F 12/1475: {in a virtual system, e.g. with translation means}
- G06F 15/17331: {Distributed shared memory [DSM], e.g. remote direct memory access [RDMA]}
- G06F 2212/1052: Security improvement
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)

## Abstract

In various embodiments, a memory key service is used to remotely access one or more portions of a physical memory included in a host node. The memory key service generates a first memory key that maps a host address space from a first protection domain associated with the host node to a second protection domain associated with a proxy node. The memory key service generates a second memory key based on the first memory key and a first address range associated with the host node. The memory key service transmits the second memory key to a software component executing on the proxy node. The software component causes a shared storage system to access a first portion of the physical memory based on the second memory key.

## Claims

### Claim 1

A computer-implemented method for remotely accessing one or more portions of a physical memory included in a host node, the method comprising:
generating a first memory key that maps a host address space from a first protection domain associated with the host node to a second protection domain associated with a proxy node;
generating a second memory key based on the first memory key and a first address range associated with the host node; and
transmitting the second memory key to a software component executing on the proxy node, wherein the software component causes a shared storage system to access a first portion of the physical memory based on the second memory key.

### Claim 2 (depends on 1)

The computer-implemented method of claim 1, further comprising:
generating a third memory key based on the first memory key and a second address range associated with the host node; and
transmitting the third memory key to the software component, wherein the software component causes the shared storage system to access a second portion of the physical memory based on the third memory key.

### Claim 3 (depends on 1)

The computer-implemented method of claim 1, wherein the first portion of the physical memory comprises a first host buffer that corresponds to the first address range.

### Claim 4 (depends on 1)

The computer-implemented method of claim 1, wherein the first memory key comprises a cross guest virtual machine identifier memory key.

### Claim 5 (depends on 1)

The computer-implemented method of claim 1, further comprising storing the second memory key in a memory key cache based on the first address range prior to transmitting the second memory key to the software component.

### Claim 6 (depends on 1)

The computer-implemented method of claim 1, further comprising failing to retrieve the second memory key from a memory key cache based on the first address range prior to generating the second memory key.

### Claim 7 (depends on 1)

The computer-implemented method of claim 1, wherein the software component causes the shared storage system to access the first portion of the physical memory by causing a storage driver executing on the proxy node to transmit to the shared storage system a remote direct memory access request that specifies the first address range and the second memory key.

### Claim 8 (depends on 1)

The computer-implemented method of claim 1, wherein the software component comprises a software application that resides in a user space.

### Claim 9 (depends on 1)

The computer-implemented method of claim 1, wherein the second memory key includes one or more memory access attributes that enable at least one of remote write access or remote read access for the first portion of the physical memory.

### Claim 10 (depends on 1)

The computer-implemented method of claim 1, wherein the shared storage system comprises at least one of shared file storage, shared block storage, or object storage.

### Claim 11 (depends on 1)

The computer-implemented method of claim 1, wherein at least the generating the second memory key is implemented via a service.

### Claim 12

One or more non-transitory computer readable media including instructions that, when executed by one or more processors, cause the one or more processors to remotely access one or more portions of a physical memory included in a host node by performing the steps of:
generating a first memory key that maps a host address space from a first protection domain associated with the host node to a second protection domain associated with a data processing unit (DPU);
generating a second memory key based on the first memory key and a first address range associated with the host node; and
transmitting the second memory key to a software component executing on the DPU, wherein the software component causes a shared storage system to access a first portion of the physical memory based on the second memory key.

### Claim 13 (depends on 12)

The one or more non-transitory computer readable media of claim 12, further comprising:
generating a third memory key based on the first memory key and one or more address ranges associated with the host node; and
transmitting the third memory key to the software component, wherein the software component causes the shared storage system to access the one or more portions of the physical memory based on the third memory key.

### Claim 14 (depends on 12)

The one or more non-transitory computer readable media of claim 12, wherein the second memory key comprises a remote key that is subordinate to the first memory key.

### Claim 15 (depends on 12)

The one or more non-transitory computer readable media of claim 12, wherein the first memory key comprises a cross guest virtual machine identifier memory key.

### Claim 16 (depends on 12)

The one or more non-transitory computer readable media of claim 12, further comprising storing the second memory key in a memory key cache based on the first address range prior to transmitting the second memory key to the software component.

### Claim 17 (depends on 12)

The one or more non-transitory computer readable media of claim 12, further comprising failing to retrieve the second memory key from a memory key cache based on the first address range prior to generating the second memory key.

### Claim 18 (depends on 12)

The one or more non-transitory computer readable media of claim 12, wherein the software component causes the shared storage system to access the first portion of the physical memory by causing a remote direct memory access data transfer operation to be invoked between the shared storage system and the host node.

### Claim 19 (depends on 12)

The one or more non-transitory computer readable media of claim 12, wherein the software component comprises a software application that resides in a user space.

### Claim 20 (depends on 12)

The one or more non-transitory computer readable media of claim 12, wherein the second memory key includes an attribute indicating that the second memory key is subordinate to the first memory key.

### Claim 21

A system comprising:
one or more memories storing instructions; and
one or more processors coupled to the one or more memories that, when executing the instructions, perform the steps of:
generating a first memory key that maps a host address space from a first protection domain associated with a host node to a second protection domain associated with a data processing unit (DPU);
generating a second memory key based on the first memory key and a first address range associated with the host node; and
transmitting the second memory key to a software component executing on the DPU, wherein the software component causes a shared storage system to access a first portion of a physical memory included in the host node based on the second memory key.
