History-Based Packet Compression for Unreliable Network Links
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing compression schemes for unreliable network connections fail to maintain compression state synchronization between sender and receiver, leading to corrupted data decompression and increased bandwidth usage.
Innovation Solution
Implement a real-time history-based compression method using dual encoders and decoders, allowing for independent encoding and decoding operations, and synchronized state updates through acknowledgement lists and hot swapping of compression schemes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of energy
If compression schemes are used for unreliable network connections, then bandwidth usage is reduced, but compression state synchronization between sender and receiver deteriorates leading to corrupted data decompression
Solution Approach 1:
The patent divides the compression system into separate encoder and decoder instances that operate independently. Each instance maintains its own compression state separately, allowing them to function autonomously without requiring continuous synchronization over the unreliable network connection.
Solution Approach 2:
The patent extracts the compression state from the network transmission process entirely. Instead of transmitting compression state information over the network, each endpoint maintains its own local state, eliminating the synchronization problem caused by unreliable connections.
2Reliability
If dual encoder instances are maintained for state synchronization, then compression state consistency is improved, but device complexity increases
Solution Approach 1:
Each encoder instance serves itself by maintaining its own compression state independently. The encoder instances do not require external synchronization or communication to maintain consistency, as they both process the same input data stream locally and arrive at the same state through independent computation.
Data Source
AI summary
Methods and systems for packet transfers are described. A computing system may initialize first and second instances of an encoder, configured to encode packets for transmission to packet recipient devices. The computing system may encode, using the first instance of the encoder, a packet. The computing system may store, using a packet reference list, a reference to the packet. The computing system may transmit, to a packet recipient device, the encoded packet. The computing system may receive, from the packet recipient device, an acknowledgement list indicating the packet. Based on identifying that the acknowledgement list matches the packet reference list, the computing system may: update a state of the second instance of the encoder based on the packet reference list, update a state of the first instance of the encoder to match a state of the second instance of the encoder, and activate the second instance of the encoder.