Decentralized Time Synchronization Using Monotonic Counters
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In decentralized or distributed telephony systems, maintaining time synchronization and data coherence across multiple terminal sets without a centralized equipment is challenging, as existing methods rely on timestamps which can be unreliable due to clock drift and non-linear time adjustments, and may require complex and costly dedicated time signals.
Innovation Solution
Each network device maintains a counter representing time passage, synchronizing it with other devices, and uses counter values to determine the recency of data changes, with a monotonically increasing or decreasing counter to track time, ensuring data coherence and avoiding issues with clock adjustments.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If timestamps are used to determine data recency in decentralized systems, then data synchronization can be attempted, but reliability deteriorates due to clock drift and non-linear time adjustments
Solution Approach 1:
The patent changes the parameter from absolute timestamps to relative sequence numbers that monotonically increase with each data change. This transforms the time measurement mechanism from relying on clock accuracy to relying on sequential ordering, thereby resolving the contradiction between synchronization reliability and measurement precision in decentralized systems.
Solution Approach 2:
The patent replaces the mechanical clock-based timestamp system with a software-based sequence number system. Instead of relying on hardware clocks that drift, the system uses software-generated sequence numbers that are strictly monotonically increasing, eliminating the source of time measurement inaccuracies while maintaining synchronization capability.
2Reliability
If centralized equipment is used for time synchronization, then time coherence can be maintained, but device complexity increases and costs increase
Solution Approach 1:
The patent extracts the time synchronization function from centralized equipment and distributes it across all network devices. Each device maintains its own monotonically increasing sequence number, eliminating the need for a centralized time server while maintaining time coherence through the distributed sequence number mechanism.
Solution Approach 2:
The patent enables each network device to perform time synchronization autonomously using its own sequence number without requiring centralized control. The system achieves self-organizing time coherence where each device independently maintains temporal information through its local monotonically increasing counter, eliminating dependency on complex centralized infrastructure.
3Measurement precision
If dedicated time signals are used for synchronization, then time accuracy can be achieved, but communication overhead increases and system cost increases
Solution Approach 1:
The patent merges the time synchronization function with the existing data communication protocol. The sequence number is carried along with data packets, so time information is transmitted as part of normal data traffic rather than through separate dedicated time signals, thereby achieving time accuracy without additional communication overhead.
Solution Approach 2:
The patent makes the existing data communication channel universal by embedding sequence numbers within data packets. The same communication infrastructure used for data transfer also serves time synchronization purposes, eliminating the need for dedicated time signal infrastructure and reducing overall system cost.
Data Source
AI summary
To synchronize time between network devices equally capable of accurately maintaining an indication of current time, one of the network devices is deemed to be a reference for time and the other network devices synchronize their indications of current time to the reference. To synchronize copies of data at multiple network devices, each network device maintains a counter representative of the passage of time but not necessarily of current time. The counter at each device is periodically synchronized with the counters of other network devices. When data is changed at a network device, the value of the counter at the time of changing is stored in association with the changed data. Stored counter values are used to determine whether a local copy or a remote copy of the data is likely more recent and therefore preferable. A further test may be applied if a counter value comparison is inconclusive.