Peer-to-Peer Network Fault Tolerance via Distributed Node Replication
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing peer-to-peer networks deployed on processing equipment face inadequate fault tolerance, as contextual information becomes unavailable when a processing equipment malfunctions, despite conventional redundancy mechanisms, leading to potential system failures and erroneous responses.
Innovation Solution
A peer-to-peer network is organized in a circular form with each node having a single successor and predecessor, and data is duplicated on a second node located on a separate processing equipment, with synchronization information managing data availability and resynchronization during node failures.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If data is replicated on the successor node in a peer-to-peer network, then fault tolerance is improved, but the system becomes vulnerable to simultaneous failures when multiple nodes are deployed on the same processing equipment
Solution Approach 1:
The patent segments the replication strategy by introducing a third node distinct from both the primary node and its successor. This creates a distributed replication pattern where data is stored across multiple processing equipment, preventing the scenario where all replicas fail simultaneously. The circular structure with three nodes ensures that failures are isolated to specific nodes rather than affecting all replicas.
Solution Approach 2:
The patent transitions from a linear succession model (node i → node i+1) to a multi-dimensional replication model involving three nodes. This dimensional expansion in the replication graph allows data to be distributed across more independent units, increasing the system's resilience against coordinated or cascading failures.
2Ease of operation
If a centralized database is used to store contextual information, then data location and retrieval are simplified, but access time becomes penalizing and response times increase
Solution Approach 1:
The patent extracts the centralized database function and distributes it across multiple peer nodes in a peer-to-peer network. Each node stores a portion of the contextual information locally, eliminating the need to access a central database. This distribution maintains ease of operation through automatic data location while dramatically improving access speed by reducing the retrieval path length.
Solution Approach 2:
The patent transforms the single-point centralization model into a multi-node distributed model. This dimensional shift from one centralized location to multiple distributed locations enables parallel access paths, reducing the time required to locate and retrieve data while maintaining operational simplicity through the distributed hash table mechanism.
3Reliability
If redundancy is incorporated to distribute load and satisfy fault tolerance constraints, then system reliability is improved, but the system complexity increases and configuration becomes difficult
Solution Approach 1:
The patent implements a universal replication rule that applies consistently across all nodes in the network. Each node follows the same logic to determine its role (primary, secondary, or tertiary) based on its position in the circular structure. This universal approach simplifies configuration and management compared to heterogeneous redundancy schemes, as all nodes operate with identical rules regardless of their specific function.
Solution Approach 2:
The patent enables the system to automatically manage its own redundancy configuration. When nodes are added or removed, the circular structure automatically reconfigures itself without external intervention. The self-service mechanism handles failover and load distribution automatically, reducing the operational complexity of managing redundancy while maintaining high reliability.
Data Source
Figure 1~2
Figure 3
AI summary
The network (Np2p) has a set of nodes (N1-N5) distributed among a set of processing devices such that each node has a unique successor node. Each node has a memory to store data (D) associated with keys and providing data associated with the key on reception of a request (Req) containing a key. Data items stored in the memory of one node (N1) is duplicated in the memory of another node (N2), where the latter node is chosen from among the nodes deployed on the processing devices different from the processing device on which the former node is deployed.