Radio Network Packet Error Handling Without Congestion Window Reduction

The patent applies local quality by implementing network-specific error processing strategies: radio networks use ARQ (Automatic Repeat Request) protocol to handle errors caused by bit error rates and multi-path fading, while wireline networks use congestion control mechanisms. This localized approach ensures each network type uses the most appropriate error handling method, preventing transmission efficiency degradation in radio networks while maintaining reliability.

The patent changes the error processing parameters based on network type. In radio networks, the system modifies the congestion control parameters to distinguish between actual congestion and radio-induced errors, adjusting the error identification thresholds and retransmission timing to account for high bit error rates and multi-path fading characteristics, thereby maintaining transmission efficiency while handling packet losses.

The patent implements dynamic error processing that adapts the congestion window adjustment based on the detected error type. When errors are detected in radio networks, the system dynamically determines whether to reduce the congestion window by analyzing error patterns and network conditions, rather than applying a static reduction rule. This dynamic approach allows the system to maintain higher data transfer rates when errors are due to radio conditions rather than actual congestion.

The patent inverts the conventional error handling logic by not reducing the congestion window size when errors are detected in radio networks, unless congestion is actually confirmed. Instead of assuming all errors indicate congestion and reducing the window, the system inverts the approach by maintaining the window size and only reducing it when there is clear evidence of congestion, thereby preventing unnecessary data transfer rate reduction while still recovering from actual packet losses.