Packet Analysis Apparatus for Distinguishing Loss and Reordering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing packet analysis methods cannot accurately distinguish between packet loss and reordering in network communications, leading to incorrect failure identification and prolonged restoration times.
Innovation Solution
A packet analysis apparatus that acquires and compares identifier values from network layer packet headers to determine if packets are resent due to loss or reordered, using a combination of source and destination address information and sequence numbers to differentiate between these failure phenomena.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If packet sequence numbers are used to detect failures, then failure detection capability is improved, but the ability to distinguish between packet loss and reordering deteriorates
Solution Approach 1:
The patent segments the failure detection task into two independent detection mechanisms: one for packet loss (using sequence number gaps) and one for reordering (using identifier monotonicity). By separating these detection functions, the system can accurately identify each failure type without confusion, resolving the contradiction between general failure detection and specific failure type identification.
Solution Approach 2:
The patent introduces an intermediary identifier field in the packet header that serves as a mediator between the sending host and receiving host. This identifier monotonically increases with each packet sent and acts as a reference marker to detect reordering events. By using this intermediary identifier, the system can precisely distinguish reordering from packet loss, improving measurement precision while maintaining reliability.
2Ease of operation
If sequence numbers alone are used for packet analysis, then analysis simplicity is improved, but failure distinction accuracy deteriorates
Solution Approach 1:
The patent segments the packet header into two functional parts for analysis: the sequence number field for detecting packet loss and the identifier field for detecting reordering. This segmentation allows simple and accurate analysis by assigning specific detection tasks to specific fields, avoiding the complexity of analyzing all packet fields while achieving high precision failure distinction.
Solution Approach 2:
The patent changes the parameter used for reordering detection from sequence number to identifier. The identifier monotonically increases with each packet sent and provides a stable reference for detecting reordering events. This parameter change enables accurate failure distinction without complicating the analysis process, as the identifier follows a simple monotonic progression similar to sequence numbers.
3Productivity
If incorrect failure determination is made, then analysis speed is improved, but restoration time increases
Solution Approach 1:
The patent replaces the mechanical process of trial-and-error failure determination with a more efficient system using identifier monotonicity detection. By substituting the complex mechanical analysis of packet sequences with a simpler monotonic identifier comparison, the system achieves both fast analysis and accurate failure distinction, preventing time waste from incorrect determinations.
Solution Approach 2:
The patent implements feedback through continuous monitoring of identifier monotonicity in packet headers. The receiving host continuously compares incoming packet identifiers with expected values to detect reordering events. This feedback mechanism enables rapid and accurate failure detection, allowing for quick restoration without the delays associated with incorrect initial determinations.
Data Source
AI summary
A packet analysis apparatus analyzes content of communication obtained as a result of monitoring or capturing a packet passing through a network. The apparatus has a unit of acquiring source or destination address information from a network layer packet header. The apparatus has a unit of acquiring from the network layer packet header an identifier for which a value that increases monotonously with each sending for each source or destination address information is set. The apparatus has a unit of searching and acquiring an identifier corresponding to address information in a current packet from a storage part holding an identifier in a previous packet corresponding to source or destination address information. The apparatus has a unit of comparing the identifier in the previous packet acquired and the identifier in the current packet and determining that reordering occurs when the identifier in the current packet is smaller.