# US 2018/0322088 A1 · Host based non-volatile memory clustering mechanism using network mapped storage

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

- Number: US 2018/0322088 A1
- Document ID: US20180322088A1
- Status: Published application
- Kind: A1
- Application number: 16/033,121
- Filed: 2018-07-11
- Published: 2018-11-08
- Earliest claimed: 2015-01-27
- Assignees: International Business Machines Corporation
- Applicants: International Business Machines Corporation
- Inventors: Keshav G. Kamble; Vijoy A. Pandey; Atul A. Tambe
- Web page: https://patentgrep.com/patent/US20180322088A1

## CPC

- G06F 15/167: using a common memory, e.g. mailbox
- G06F 3/061: {Improving I/O performance}
- G06F 3/0632: {by initialisation or re-initialisation of storage systems}
- G06F 3/067: {Distributed or networked storage systems, e.g. storage area networks [SAN], network attached storage [NAS]}
- G06F 9/50: Allocation of resources, e.g. of the central processing unit [CPU]
- G06F 13/28: using burst mode transfer, e.g. direct memory access {DMA}, cycle steal (G06F13/32 takes precedence)
- H04L 65/611: for multicast or broadcast (systems for broadcast or conference H04L12/18; arrangements for broadcast or distribution combined with broadcast H04H20/00; arrangements for broadcast applications with a direct linkage to broadcast information or to broadcast space-time H04H60/00; selective distribution of broadcast services, e.g. multimedia broadcast multicast service [MBMS], H04W4/06)
- H04L 67/1097: for distributed storage of data in networks, e.g. transport arrangements for network file system [NFS], storage area networks [SAN] or network attached storage [NAS]
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- H04L: TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION (arrangements common to telegraphic and telephonic communication H04M)

## Abstract

In one embodiment, a method includes initiating cluster parameters that govern how a non-volatile memory (NVM) cluster functions and operates. Submission and completion queues of shared NVM on other nodes in the NVM cluster are mapped based on details of the shared NVM on the other nodes in the NVM cluster. The submission queue is configured to store commands to access the shared NVM according to a first-in-first-out (FIFO) scheme. The completion queue is configured to store completed commands after being processed through the submission queue. In another embodiment, a host-based data storage system includes NVM configured to store data. The host-based data storage system further includes a processor and logic integrated with and/or executable by the processor to perform the foregoing method.

## Claims

### Claim 1

A method, comprising:
initiating cluster parameters that govern how a non-volatile memory (NVM) cluster functions and operates; and
mapping submission and completion queues of shared NVM on other nodes in the NVM cluster based on details of the shared NVM on the other nodes in the NVM cluster,
wherein the submission queue is configured to store commands to access the shared NVM according to a first-in-first-out (FIFO) scheme, and
wherein the completion queue is configured to store completed commands after being processed through the submission queue.

### Claim 2 (depends on 1)

The method as recited in claim 1, wherein submission and completion queues of the shared NVM on the other nodes in the NVM cluster are mapped to a memory controller using remote direct memory access (RDMA).

### Claim 3 (depends on 2)

The method as recited in claim 2, wherein the details of the shared NVM on the other nodes in the NVM cluster include a RDMA over converged Ethernet (RoCE) network interface card (NIC) address of the other nodes in the NVM cluster that is used to access the shared NVM on the other nodes in the NVM cluster by other nodes in the NVM cluster.

### Claim 4 (depends on 1)

The method as recited in claim 1, wherein local shared NVM is accessible to the other nodes in the NVM cluster via at least one protocol selected from the group consisting of: coherent accelerator processor interface (CAPI) and NVM express (NVMe).

### Claim 5 (depends on 1)

The method as recited in claim 1, comprising:
sending data from local shared NVM to at least one of the other nodes in the NVM cluster via remote direct memory access (RDMA) using at least one protocol selected from the group consisting of: NVM express (NVMe) over Ethernet and coherent accelerator processor interface (CAPI) over Ethernet; and
performing operations on submission and completion queues of the shared NVM on the other nodes in the NVM cluster via RDMA over converged Ethernet (RoCE), the operations comprising at least one operation selected from the group consisting of: accessing data, accessing instructions, queuing data, and queuing instructions.

### Claim 6 (depends on 1)

The method as recited in claim 1, comprising multicasting the cluster parameters of the NVM cluster to the other nodes in the NVM cluster in response to initiating the cluster parameters.

### Claim 7

A host-based data storage system, comprising:
non-volatile memory (NVM) configured to store data;
a processor; and
logic integrated with and/or executable by the processor, the logic being configured to:
initiate cluster parameters that govern how a NVM cluster functions and operates; and
map submission and completion queues of shared NVM on other nodes in the NVM cluster based on details of the shared NVM on the other nodes in the NVM cluster,
wherein the submission queue is configured to store commands to access the shared NVM according to a first-in-first-out (FIFO) scheme, and
wherein the completion queue is configured to store completed commands after being processed through the submission queue.

### Claim 8 (depends on 7)

The host-based data storage system as recited in claim 7, wherein submission and completion queues of the shared NVM on the other nodes in the NVM cluster are mapped to a memory controller using remote direct memory access (RDMA).

### Claim 9 (depends on 8)

The host-based data storage system as recited in claim 8, wherein the details of the shared NVM on the other nodes in the NVM cluster include a RDMA over converged Ethernet (RoCE) network interface card (NIC) address of the other nodes in the NVM cluster that is used to access the shared NVM on the other nodes in the NVM cluster by any other node in the NVM cluster.

### Claim 10 (depends on 7)

The host-based data storage system as recited in claim 7, wherein shared NVM on the host-based data storage system is accessible to the other nodes in the NVM cluster via at least one protocol selected from the group consisting of: coherent accelerator processor interface (CAPI) and NVM express (NVMe).

### Claim 11 (depends on 7)

The host-based data storage system as recited in claim 7, wherein the logic is configured to:
send data from shared NVM on the host-based data storage system to at least one of the other nodes in the NVM cluster via remote direct memory access (RDMA) using at least one protocol selected from the group consisting of: NVM express (NVMe) over Ethernet and coherent accelerator processor interface (CAPI) over Ethernet; and
perform operations on submission and completion queues of the shared NVM on the other nodes in the NVM cluster via RDMA over converged Ethernet (RoCE), the operations comprising at least one operation selected from the group consisting of: accessing data, accessing instructions, queuing data, and queuing instructions.

### Claim 12 (depends on 7)

The host-based data storage system as recited in claim 7, wherein the logic is further configured to: multicast the cluster parameters of the NVM cluster to the other nodes in the NVM cluster in response to initiating the cluster parameters.

### Claim 13

A computer program product, the computer program product comprising a computer readable storage medium having program instructions embodied therewith, wherein the computer readable storage medium is not a transitory signal per se, the program instructions executable by a computer to cause the computer to perform a method comprising:
multicasting cluster parameters of a non-volatile memory (NVM) cluster to other nodes of a NVM cluster; and
mapping submission and completion queues of shared NVM on the other nodes in the NVM cluster based on details of the shared NVM on the other nodes in the NVM cluster,
wherein the submission queue is configured to store commands to access the shared NVM according to a first-in-first-out (FIFO) scheme, and
wherein the completion queue is configured to store completed commands after being processed through the submission queue.

### Claim 14 (depends on 13)

The computer program product as recited in claim 13, wherein submission and completion queues of the shared NVM on the other nodes in the NVM cluster are mapped to a memory controller using remote direct memory access (RDMA).

### Claim 15 (depends on 14)

The computer program product as recited in claim 14, wherein the details of the shared NVM on the other nodes in the NVM cluster include a RDMA over converged Ethernet (RoCE) network interface card (NIC) address of the other nodes in the NVM cluster that is used to access the shared NVM on the other nodes in the NVM cluster by other nodes in the NVM cluster.

### Claim 16 (depends on 13)

The computer program product as recited in claim 13, wherein local shared NVM is accessible to the other nodes in the NVM cluster via at least one protocol selected from the group consisting of: coherent accelerator processor interface (CAPI) and NVM express (NVMe).

### Claim 17 (depends on 13)

The computer program product as recited in claim 13, comprising program instructions for:
sending data from local shared NVM to at least one of the other nodes in the NVM cluster via remote direct memory access (RDMA) using at least one protocol selected from the group consisting of: NVM express (NVMe) over Ethernet and coherent accelerator processor interface (CAPI) over Ethernet; and
performing operations on submission and completion queues of the shared NVM on the other nodes in the NVM cluster via RDMA over converged Ethernet (RoCE), the operations comprising at least one operation selected from the group consisting of: accessing data, accessing instructions, queuing data, and queuing instructions.

### Claim 18 (depends on 13)

The computer program product as recited in claim 13, comprising program instructions for multicasting the cluster parameters of the NVM cluster to the other nodes in the NVM cluster in response to initiating the cluster parameters.
