Multi-tier Computing Node Architecture for Distributed Database Link Management
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Distributed databases face difficulties in maintaining a large number of back-end links as the number of storage nodes and concurrent tasks increases, leading to link overload and maintenance challenges.
Innovation Solution
Implementing a method where service request messages are sent through multiple tiers of computing nodes, with more nodes in tiers closer to the storage unit, reducing the load on individual computing nodes and allowing for flexible distribution policies.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If computing nodes create links with all storage nodes to handle concurrent tasks, then the system can process more service requests, but the number of back-end links becomes too large and difficult to maintain
Solution Approach 1:
The patent introduces a multi-tier computing node architecture where computing nodes are divided into different tiers (first tier, second tier, third tier, etc.). Each tier handles a subset of storage nodes, segmenting the overall system into manageable chunks. This segmentation reduces the number of links each computing node must maintain while preserving the system's overall processing capability through hierarchical distribution of requests.
Solution Approach 2:
The patent introduces intermediate computing nodes (second tier and beyond) that act as mediators between the first tier computing nodes and storage nodes. These intermediary nodes relay service request messages, reducing the direct link requirements between first tier nodes and storage nodes. The intermediary nodes buffer and forward requests, effectively reducing the total number of direct links needed in the system.
2Adaptability or versatility
If the number of storage nodes increases to handle more service data, then the system capacity increases, but the number of back-end links that computing nodes must maintain becomes unthinkable
Solution Approach 1:
The multi-tier architecture segments the system so that each computing node only needs to maintain links with a limited number of storage nodes or intermediate nodes. When storage nodes increase, the system can add more tiers or nodes within existing tiers without proportionally increasing links at each individual node, thus maintaining scalability without exponential link growth.
Solution Approach 2:
The patent transitions from a flat one-to-many connection model to a hierarchical multi-dimensional structure. Instead of computing nodes directly connecting to all storage nodes in a single dimension, the system adds temporal and hierarchical dimensions through multiple tiers, allowing the system to scale storage nodes without linearly increasing link complexity at each node.
3Productivity
If computing nodes maintain a large number of back-end links to handle concurrent tasks, then more service requests can be processed, but link recovery and recreation increase time delay
Solution Approach 1:
By segmenting the system into multiple tiers, each computing node maintains fewer links, reducing the impact of link failures. When links need recovery, fewer links mean less time for detection and recreation. The hierarchical structure allows localized link management, isolating failures to specific tiers rather than affecting the entire system.
Solution Approach 2:
Intermediary computing nodes act as buffers that can absorb and redirect traffic when direct links fail. Instead of immediately recreating failed links between first tier nodes and storage nodes, the system can route requests through alternative intermediate nodes, reducing the time penalty for link failures and enabling smoother link recovery processes.
Data Source
AI summary
A service request message sending method, a distributed database architecture and a computer readable storage medium are disclosed. The method includes: receiving (S101) a service request message, and sending (S102) the service request message to a corresponding storage unit through N tiers of computing nodes, where N is an integer greater than or equal to 2, and the closer the tier is to the storage unit, the larger the number of computing nodes in that tier is.