Diverse Path Forwarding via Trial and Error in LLNs
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Computing diverse paths in low power and lossy networks (LLNs) is challenging, especially with distance vector routing protocols like RPL, where reverse path computation is difficult and complex, particularly in constrained environments.
Innovation Solution
A data-driven trial and error technique is employed where a first packet is sent along a preferred path, and a second packet with a different distinguisher value is sent along a diverse path, with nodes storing packet information to determine if they share common resources, allowing for redirection and optimization of paths through feedback.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional routing protocols (RPL) are used in LLNs, then routing functionality is provided, but computing diverse paths becomes difficult and complex
Solution Approach 1:
The patent implements feedback mechanisms where nodes store received packets with distinguisher values and use this information to determine whether to forward packets along diverse paths. The feedback loop involves nodes detecting packet arrivals, comparing distinguisher values, and adjusting forwarding decisions accordingly, enabling diverse path computation without complex routing protocols.
Solution Approach 2:
Each node in the LLN autonomously determines whether to forward packets along diverse paths by locally storing packet information and comparing distinguisher values. The system eliminates the need for centralized routing computation or complex protocol coordination, allowing nodes to self-determine diverse path forwarding based on simple local feedback.
2Reliability
If 1+1 protection is implemented with packet duplication, then reliability is improved, but path computation complexity increases
Solution Approach 1:
The patent implements packet duplication by creating multiple copies of data packets with different distinguisher values (e.g., distinguisher 1 and distinguisher 2). These copies are sent along different paths simultaneously, and nodes use the distinguisher values to manage the copies, enabling 1+1 protection without complex path computation.
Solution Approach 2:
The patent changes the packet parameter space by introducing distinguisher values as an additional field. This parameter change enables nodes to differentiate between multiple packet copies and make forwarding decisions based on simple distinguisher value comparisons, rather than complex path computation.
3Device complexity
If distance vector routing protocols are used, then routing is simplified, but reverse path computation becomes difficult
Solution Approach 1:
The patent inverts the traditional routing computation approach by not computing reverse paths from destination to source, but rather having nodes locally determine forward paths based on received packet information. This inversion transforms the difficult reverse path computation problem into simple local forwarding decisions.
Solution Approach 2:
The patent uses feedback from received packets (through distinguisher values and packet storage) to determine forwarding paths. This feedback mechanism eliminates the need for complex reverse path computation by using actual packet arrival information to guide forward path selection.
Data Source
AI summary
In one embodiment, a node determines an intention to transmit a diversely forwarded packet through a computer network, and as such, transmits a first version of the packet having a packet identifier (ID) and a first distinguisher value to a first next-hop node, and transmits a second version of the packet having the same packet ID and a second distinguisher value different from the first distinguisher value to a second next-hop node different from the first next-hop node. In another embodiment, a next-hop node that receives the packet determines whether any previously received packet at the next-hop node had a same packet ID and a different distinguisher value. In response to determining that no previously received packet has the same packet ID and different distinguisher value, the next-hop node stores the packet ID and the distinguisher value, and forwards the packet to a selected next-hop node.