Service Mesh for Unified Cross-Cluster Resource Access

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing technologies face challenges in accessing resource instances across different clusters in a unified manner due to customized input/output interface access patterns, leading to difficulties in communication between functional units in applications.

Innovation Solution

A method and apparatus for performing resource scheduling in a cluster by creating a network service instance, establishing a network link using a network address and allocated port, updating the network state, and launching the resource instance to communicate via the established link.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of manufacture

If customized input/output interface access patterns are used in each cluster, then each cluster can provide optimized resource access, but it becomes difficult to access resource instances in a unified manner across different clusters

Engineering Contradiction:
Improvecluster resource access optimizationVSAvoidcross-cluster resource access unity
Core Design Contradiction:
Ease of manufactureVSEase of operation

Solution Approach 1:

The patent introduces a service mesh architecture as an intermediary layer between applications and resource instances. This service mesh provides a unified access interface that abstracts away the customized I/O patterns of individual clusters, allowing applications to access resources across different clusters through a consistent interface while each cluster maintains its optimized access patterns internally

Inventive Principle:
Principle #24Intermediary (Mediator)

2Adaptability or versatility

If each cluster develops its own cluster technology based on general cluster architecture, then each cluster can be optimized for specific functions, but communication between functional units in applications becomes difficult

Engineering Contradiction:
Improvecluster technology customizationVSAvoidcommunication complexity between functional units
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent implements a universal service mesh that can operate across clusters with different technologies and architectures. This service mesh provides common communication protocols, service discovery mechanisms, and load balancing capabilities that work uniformly across diverse cluster environments, enabling functional units in applications to communicate seamlessly regardless of which cluster they reside in

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

3Productivity

If resource instances in different clusters have customized access patterns, then each resource instance can be accessed optimally within its cluster, but applications cannot access resource instances in a unified manner from outside the cluster

Engineering Contradiction:
Improveresource instance access efficiencyVSAvoidunified access from outside cluster
Core Design Contradiction:
ProductivityVSEase of operation

Solution Approach 1:

The patent adds an additional abstraction dimension by introducing the service mesh layer between the application layer and the resource instance layer. This intermediate dimension provides unified access patterns at the application level while preserving optimized access patterns at the resource level, effectively solving the contradiction by operating at multiple levels of the system hierarchy

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

Data Source

PatentUS20250080473A1Method, apparatus, device and medium for performing resource schedule in cluster
Publication Date: 2025.03.06 SHANGHAI SUIXUNTONG ELECTRONICS TECH CO LTD
  • US20250080473A1 patent drawing
  • US20250080473A1 patent drawing
  • US20250080473A1 patent drawing

AI summary

There are provided methods, apparatuses, devices, and media for performing resource schedule in a cluster. In a method, a network service instance is created for managing a network service of the resource instance. A network link is established for accessing the resource instance using a network address of the resource instance and a network port that is allocated to the resource instance. A network state in the network service instance is updated based on the network link. In response to detecting that the network state indicates that the network link has been established, the resource instance is launched to communicate with the launched resource instance via the network link. With example implementations of the present disclosure, there is provided an access manner that does not rely on dedicated capabilities with any particular cluster. In this way, resource instance access under multiple clusters is managed in a general and unified manner.