Decentralized RPC Node Discovery via Peer-to-Peer List Synchronization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The centralized discovery mechanism in micro-service architectures for cloud services increases system complexity and operation and maintenance risks, as it relies on a single 'service registry' component that can cause the entire system to become unavailable if it fails.
Innovation Solution
A decentralized RPC micro-service dynamic discovery mechanism where nodes in an RPC cluster push and merge their local member lists through peer-to-peer communication, synchronizing information without a central component, ensuring continued operation even in network failures.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a centralized service registry component is introduced to acquire RPC member information, then the system can dynamically discover and register RPC nodes, but the system complexity and operation and maintenance complexity increase
Solution Approach 1:
The patent extracts and removes the centralized service registry component from the system. Instead of having a separate registry service, each RPC node maintains its own local member list and performs peer-to-peer synchronization, eliminating the need for the centralized discovery mechanism while preserving dynamic service discovery capability.
Solution Approach 2:
The patent segments the centralized service registry functionality and distributes it to each individual RPC node. Each node independently maintains and synchronizes member information with other nodes through peer-to-peer communication, transforming a monolithic centralized system into a distributed segmented architecture.
2Adaptability or versatility
If a centralized service registry component is introduced to acquire RPC member information, then dynamic service discovery is enabled, but the operation and maintenance risks increase due to single-point failure
Solution Approach 1:
The patent segments the service registry functionality across multiple independent nodes rather than relying on a single centralized registry. This distribution eliminates the single-point failure risk, as the failure of any individual node does not prevent other nodes from discovering and registering RPC members through their local member lists and peer-to-peer synchronization.
3Device complexity
If a centralized service registry is used for RPC node registration, then centralized control is achieved, but the operation and maintenance complexity increases
Solution Approach 1:
The patent removes the centralized service registry component entirely, replacing it with a distributed peer-to-peer model where each node independently manages its own member list. This eliminates the operational overhead and maintenance complexity associated with managing a centralized registry service while maintaining the ability to track and discover RPC nodes.
Data Source
Figure 1~2
Figure 3~4
AI summary
The present disclosure provides a method and an apparatus for acquiring remote-procedure-call (RPC) member information, an electronic device and a storage medium, and relates to the field of micro-services technologies in cloud services. The method includes: pushing, by nodes of a RPC cluster, respective local member lists of the nodes of the RPC cluster to each other, and merging, by each node of the RPC cluster, a received member list with a current local member list while performing deduplication until member lists of the nodes of the RPC cluster are synchronized. According to the present disclosure, the member list of each node of the RPC cluster may be converged and synchronized eventually, and each node may acquire relevant information of all RPC members eventually. Even if a network failure occurs, the member lists of the nodes may be converged and synchronized eventually after the network failure is eliminated. The embodiment of the present disclosure may be applied to a cloud platform and the Internet of Things. The normal and stable operation of an entire cloud service system may be ensured.