Loss Detection Packets for Low Latency Communication
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional packet-based networks experience significant latency due to dropped data packets, which is particularly problematic for latency-critical operations, as existing recovery mechanisms often result in lengthy delays and are not widely implemented due to scale limitations and high costs associated with specialized networks.
Innovation Solution
The method involves appending loss detection packets to message packets in a communication session, allowing for quick detection of packet loss without affecting application performance, and triggering a retransmit procedure if these packets are not received, thereby reducing communication latency without requiring specialized networks or complex timer settings.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional packet-based networks are used for reliable data transfer, then high performance at reasonable cost is achieved, but packet drops occur causing 250-300 ms delays that impair application performance
Solution Approach 1:
The patent applies preliminary action by sending probe packets before the application data packets. These probe packets are transmitted in advance through the network path to detect potential packet loss conditions. When probe packets are lost, the system proactively triggers retransmission of application data packets before actual application performance is impacted, thereby reducing the 250-300 ms delays associated with conventional reactive retransmission mechanisms
Solution Approach 2:
The patent introduces probe packets as intermediary elements that mediate between the sender and receiver to detect network conditions. These probe packets serve as a intermediary mechanism to identify packet loss without affecting the primary application data flow, enabling early detection and triggering of retransmission procedures that reduce overall latency while maintaining data transfer reliability
2Loss of time
If specialized networks like DCE or InfiniBand are implemented to reduce packet drops, then latency is reduced, but scale limitations and high implementation costs occur
Solution Approach 1:
The patent applies universality by designing a probe packet mechanism that can operate over conventional packet-based networks without requiring specialized network infrastructure. The same probe packet approach can be applied across different network types and configurations, making the solution universally applicable to existing Ethernet and other standard networks while achieving low-latency performance previously only available from specialized networks
Solution Approach 2:
The patent uses lightweight probe packets that are simple, small in size, and可以快速 generated and discarded. These probe packets serve as disposable elements that provide valuable network condition information without consuming significant network resources or requiring complex hardware. The approach achieves specialized-network-level performance using standard network infrastructure, avoiding the high costs and complexity of DCE or InfiniBand implementations
3Measurement precision
If probe packets are sent to detect packet loss, then packet loss detection accuracy is improved, but network overhead increases
Solution Approach 1:
The patent applies partial action by sending probe packets selectively rather than continuously. Probe packets are transmitted based on specific conditions such as before batches of application data packets, or when network conditions suggest potential loss. This selective probing achieves accurate packet loss detection while minimizing network overhead by avoiding unnecessary probe transmissions that would consume bandwidth and energy
Data Source
AI summary
Systems and methods for reducing communication latency are provided. A communication session with a remote device is initiated. One or more message packets are transmitted to the remote device. Subsequently, it is determined whether any additional message packet is scheduled to be transmitted to the remote device. In the event that no additional message packet is scheduled to be sent to the remote device, a loss detection packet is appended to a last message packet. The last message packet and the loss detection packet are transmitted to the remote device. It is further determined whether the last message packet was successfully received by the remote device based on receipt of an acknowledgment message. In the event that the last message packet was not successfully transmitted, a retransmit of the last message packet is executed.