Byzantine Fault-Tolerant Distributed Processing with 1-Step Consensus
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing distributed processing systems fail to maintain data consistency when Byzantine failures occur, leading to improper processing and increased communication costs, as they do not adequately address device shutdowns or software bugs in servers like FPGA, which can cause invalid processing.
Innovation Solution
A distributed processing method that includes a first determination step to assess data consistency using a process-saving 1-step consensus algorithm and a second determination step using a 2-step consensus algorithm, with a combination step to ensure consistency, while setting the number of servers based on allowable failures and Byzantine failures, and employing a collision resolution value selection to prevent abnormal values from being output.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the number of servers is increased to maintain data consistency under Byzantine failures, then reliability is improved, but device complexity increases
Solution Approach 1:
The patent changes the parameter of consensus algorithm from traditional multi-step approaches to a 1-step consensus algorithm. This parameter change allows the system to achieve the same reliability guarantee with fewer servers, specifically requiring only n > 2f + b servers instead of more complex configurations, thereby reducing device complexity while maintaining data consistency under Byzantine failures
2Reliability
If the number of communication rounds is increased to ensure data consistency, then reliability is improved, but loss of time increases
Solution Approach 1:
The patent skips traditional multi-step consensus communication rounds by implementing a 1-step consensus algorithm. Instead of performing multiple sequential communication rounds as in traditional approaches, the system completes the consensus process in a single communication step, thereby reducing communication time while maintaining reliability guarantees for data consistency under Byzantine failures
3Reliability
If traditional consensus algorithms are used to handle Byzantine failures, then reliability is improved, but device complexity increases
Solution Approach 1:
The patent changes the algorithmic parameter from traditional multi-phase consensus algorithms to a streamlined 1-step consensus algorithm. This parameter change simplifies the processing logic while maintaining the ability to handle Byzantine failures, reducing device complexity by eliminating unnecessary communication phases and simplifying the consensus decision-making process
Data Source
AI summary
A distributed processing method to receive data by a plurality of servers each including a processor and a memory, and process the data by replicating, the method includes a first determination step in which the servers each receive the replicated data, and a first determination unit determines a degree of consistency of the received data and an output step in which the servers each receive a determination result of the degree of consistency of the data from the first determination unit, and if the determination result includes data that guarantees consistency, the server outputs the data that guarantees consistency. A first number of servers that are to receive the data is set in advance based on a prescribed allowable number of failures that defines the number of servers that can have failures, and an allowable number of byzantine failures that defines the number of servers that can have byzantine failures.