Packet Service Availability Determination Using Sliding Window Extension
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for determining service availability in Ethernet networks, such as those described in MEF 10.3, face inaccuracies due to ignored time intervals at the end of Maintenance Intervals (MI) and High Loss Intervals (HLI), leading to incorrect state transitions and SLA compliance issues.
Innovation Solution
A method that uses frame loss measurements in short intervals Δt, with a sliding window of size n, to determine availability, and an extension period after the MI to ensure all Δt's are marked as available or unavailable, preventing the ignoring of frames at the end of intervals and improving accuracy by ensuring all time intervals are accounted for.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If a sliding window algorithm is used to mark Δt as available or unavailable based on frame loss measurements, then service availability can be determined systematically, but the last Δt at the end of each Maintenance Interval is ignored leading to measurement inaccuracy
Solution Approach 1:
The patent applies preliminary action by extending the sliding window evaluation into the repetition time period before the next Maintenance Interval begins. This allows the algorithm to pre-determine the state of the last Δt of the previous MI using frame loss measurements collected during the repetition time, ensuring no Δt is ignored and maintaining measurement accuracy while preserving systematic evaluation efficiency.
2Reliability
If High Loss Intervals are present in a Maintenance Interval causing the entire MI to be marked as unavailable, then service unavailability is detected, but the inaccuracy worsens due to ignored Δt at the end of the previous MI
Solution Approach 1:
The patent uses preliminary action by collecting frame loss measurements during the repetition time period and applying the sliding window algorithm in advance to determine the state of the last Δt before the next MI starts. This ensures that even when HLIs are present, the measurement is accurate because the evaluation is completed using all available data including the repetition time, preventing propagation of inaccuracy from ignored Δt.
Solution Approach 2:
The patent maintains continuity of useful action by utilizing the repetition time period continuously for frame loss measurements and sliding window evaluation. This continuous measurement approach ensures that the determination of Δt states is based on complete data without gaps, improving both reliability of unavailability detection and precision of measurement by eliminating ignored intervals.
3Productivity
If the sliding window algorithm marks Δt states within Maintenance Intervals, then availability performance can be measured, but inaccurate transitions of service availability state occur due to ignored Δt
Solution Approach 1:
The patent applies preliminary action by performing the sliding window evaluation during the repetition time period to determine the state of the last Δt before the next Maintenance Interval begins. This preliminary determination ensures accurate state transitions because the algorithm has access to complete frame loss data without ignoring any Δt, while maintaining the productivity of systematic availability measurement.
Data Source
AI summary
Service availability determination systems and methods include determining availability of a packet service in a Maintenance Interval (MI) based on frame loss measurements in short intervals Δt and marking each Δt as available or unavailable based on the frame loss measurements and an associated Frame Loss Ratio (FLR) threshold, wherein each Δt is a High Loss interval (HLI) when exceeding the FLR threshold; utilizing a sliding window of size n, n being an integer, to determine whether the packet service is available or unavailable; and utilizing an extension period after an end of the MI with the sliding window to ensure all Δt's in the MI are marked as available or unavailable.