Network Clock Recalibration via Stored One-Way Delay
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
High-precision clocks in computer networks face challenges in accurately resetting their time after power failures, as existing timing transfer protocols assume symmetric paths, leading to significant errors due to asymmetric paths, which are impossible to measure with two-way time transfer techniques like NTP or PTP.
Innovation Solution
The solution involves measuring and storing a one-way delay from a trusted reference device, allowing the clock to be recalibrated by adding the stored one-way delay to the timestamp of a received timing message after power restoration, ensuring accurate time resetting without relying on symmetric path assumptions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If two-way time transfer protocols (NTP or PTP) are used to reset clock time after power failure, then the clock can be reset, but significant errors occur due to asymmetric network paths that cannot be measured
Solution Approach 1:
The system performs preliminary measurement and storage of one-way delay from the reference device to the network device before power failure occurs. This pre-measured delay value is stored and later used to accurately calculate the correct time after power restoration, eliminating the need to measure path delay again and avoiding errors from asymmetric path changes.
Solution Approach 2:
The patent introduces a reference device as an intermediary that provides timing messages with timestamps. This reference device serves as a trusted time source that enables the network device to accurately determine its time by combining the timestamp with the pre-measured one-way delay, rather than relying on two-way protocols that fail to account for asymmetric paths.
2Duration of action of stationary object
If backup battery supply is used to keep the clock running during power failure, then the clock maintains time, but the same concern applies if the power failure lasts longer than the backup battery can support
Solution Approach 1:
The system enables the clock to self-recalibrate after power failure by automatically receiving timing messages from the reference device and using the pre-stored one-way delay to calculate the correct time. This self-service mechanism eliminates dependency on backup batteries and manual intervention, allowing the clock to reliably recover regardless of power failure duration.
3Device complexity
If symmetric path assumptions are made in timing protocols, then the protocols are simpler to implement, but significant errors occur in asymmetric network conditions
Solution Approach 1:
The patent explicitly addresses asymmetric network paths by measuring and storing the one-way delay from the reference device to the network device. This approach acknowledges and compensates for path asymmetry rather than assuming symmetry, enabling accurate time calculation even when forward and reverse paths differ significantly in delay.
Data Source
AI summary
In one embodiment, a clock on a network device is initialized, and then a first timing message is received at the network device from a reference device having a first timestamp indicating when the first timing message was transmitted from the reference device. The network device may then determine and store a one-way delay from the first timestamp to a first time at which the first timing message was received at the network device. In response to restarting the clock, the network device may receive a second timing message from the reference device having a second timestamp indicating when the second timing message was transmitted from the reference device. The network device may then calibrate the clock such that a second time at which the network device received the second timing message is the second timestamp plus the stored one-way delay.