SoC Cluster Remote Memory Access via Fabric Interconnect

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Traditional shared nothing clusters lack the ability to provision memory based on workload, implement memcached functionality on a per-node basis, perform remote DMA transactions, and handle remote interrupts efficiently, limiting their performance and scalability.

Innovation Solution

The system enables flexible allocation and disassociation of computing resources by implementing remote memory access, memcached functionality, and remote interrupts across a cluster of nodes using a power-optimized interconnect fabric, allowing for scalable and efficient memory sharing and transaction handling.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If shared nothing architecture is used for cluster scalability, then node independence and scalability are improved, but memory provisioning flexibility and remote access capability deteriorate

Engineering Contradiction:
Improvememory provisioning flexibilityVSAvoidcluster architecture complexity
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent introduces a fabric interconnect as an intermediary between nodes, enabling remote memory access and DMA transactions without compromising node independence. This mediator allows nodes to access each other's memory resources while maintaining the shared-nothing architecture's scalability benefits.

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The fabric interconnect serves multiple functions: it enables remote memory access, supports DMA transactions, facilitates interrupt handling, and provides a unified communication infrastructure for various protocols. This multi-functionality resolves the contradiction by providing memory provisioning flexibility through a single architectural addition rather than multiple separate mechanisms.

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

2Productivity

If remote memory access capability is added to shared nothing clusters, then memory sharing and remote DMA functionality are improved, but network performance and power consumption worsen

Engineering Contradiction:
Improveremote data access efficiencyVSAvoidpower consumption
Core Design Contradiction:
ProductivityVSUse of energy by moving object

Solution Approach 1:

The patent replaces traditional network-based remote access mechanisms with fabric-based direct memory access. This substitution eliminates the need for network protocol processing and data copying, enabling efficient remote DMA transactions that consume less power while maintaining high productivity.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The patent extracts memory access operations from the network protocol stack and places them directly at the fabric level. By taking out the network processing overhead and handling memory access directly through the fabric interconnect, the system achieves high remote data access efficiency with reduced power consumption.

Inventive Principle:
Principle #2Taking out (Extraction)

3Adaptability or versatility

If memcached functionality is implemented on per-node basis, then distributed caching capability is improved, but system complexity and coordination overhead worsen

Engineering Contradiction:
Improvedistributed caching capabilityVSAvoidsystem coordination complexity
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent enables each node to independently provide memcached functionality by exposing its local memory resources through the fabric interconnect. Nodes self-serve as cache servers without requiring complex centralized coordination, achieving distributed caching capability while minimizing system complexity through autonomous node operation.

Inventive Principle:
Principle #25Self-service

Data Source

PatentUS11526304B2Memcached server functionality in a cluster of data processing nodes
Publication Date: 2022.12.13 III HOLDINGS 2 LLC
  • US11526304B2 patent drawing
  • US11526304B2 patent drawing
  • US11526304B2 patent drawing

AI summary

A method is performed by a first server on a chip (SoC) node that is one instance of a plurality of nodes within a cluster of nodes. An operation is performed for determine if a second one of the SoC nodes in the cluster has data stored thereon corresponding to a data identifier in response to receiving a data retrieval request including the data identifier. An operation is performed for determining if a remote memory access channel exists between the SoC node and the second one of the SoC nodes. An operation is performed for access the data from the second one of the SoC nodes using the remote memory access channel after determine that the second one of the SoC nodes has the data stored thereon and that the remote memory access channel exists between the SoC node and the second one of the SoC nodes.