Intermediate Node Packet Reclassification for Network Congestion Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current data packet networks face challenges in determining the optimal transmission rate due to outdated congestion feedback, leading to congestion and packet loss, especially as network capacities increase, and existing buffer management techniques fail to efficiently handle different classes of service, causing unnecessary packet drops and capacity wastage.
Innovation Solution
Introducing an unqueuable class of service where intermediate nodes reclassify queuable packets as unqueuable, allowing for rapid establishment of transmission rates without modifying source and receiver nodes, thereby minimizing packet drops and optimizing network capacity utilization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If closed-loop congestion control algorithms are used to determine appropriate transmission rates, then packet loss is reduced, but the feedback becomes outdated quickly as network conditions change, leading to inaccurate rate prediction
Solution Approach 1:
The patent applies preliminary action by having intermediate nodes proactively send probe packets before actual data transmission begins. These probe packets traverse the network path and elicit congestion feedback in advance, allowing the sender to determine appropriate transmission rates before real data flows start, thus avoiding outdated feedback issues during actual transmission.
Solution Approach 2:
The patent uses intermediary probe packets sent by intermediate nodes to mediate the congestion feedback process. These probe packets act as messengers that carry congestion information from intermediate nodes back to the sender, enabling the sender to learn network conditions without waiting for actual data packet losses to generate feedback.
2Productivity
If TCP Slow-Start algorithm is used for initial exponential growth phase, then transmission rate increases rapidly, but overshoot occurs where packets are sent twice as fast as available capacity, causing congestion and packet drops
Solution Approach 1:
The patent applies preliminary action by having intermediate nodes send probe packets before actual data transmission to determine the bottleneck rate in advance. This preliminary exploration allows the sender to set appropriate initial transmission rates without experiencing the overshoot problem that occurs with traditional TCP Slow-Start, where exponential growth continues until congestion feedback is received one round trip time after exceeding capacity.
Solution Approach 2:
The patent enhances feedback by having intermediate nodes actively send probe packets and return congestion information to the sender. This feedback mechanism allows the sender to learn the actual available capacity and adjust transmission rates accordingly, preventing the overshoot that occurs with passive TCP feedback that only activates after packet losses have already occurred.
3Device complexity
If conventional buffer management techniques are used to handle different classes of service, then packet processing is simplified, but higher-priority packets are not protected and capacity is wasted due to unnecessary drops
Solution Approach 1:
The patent applies local quality by treating different packet types differently at intermediate nodes. Queuable packets (data packets that can wait in buffers) are handled with standard buffering and queuing, while unqueuable packets (probe packets and time-sensitive traffic) are processed differently - probe packets are sent immediately without queuing, and time-sensitive packets receive priority treatment. This differentiated local handling protects higher-priority packets while managing buffer resources efficiently.
Data Source
AI summary
The invention includes a network node, and a method of controlling the network node in a data packet network, wherein the method comprising the steps of: receiving a first data packet from a first external network node; analysing the first data packet to determine if it is of a class of service deemed to be queuable or unqueuable; and, if it is queuable, sending the first data packet to a second external node via an intermediate node, wherein the first data packet is reclassified to be of the unqueuable class of service such that the intermediate node should not forward the first data packet to the second external network node if a packet queue exists at the intermediate node.