Disjoint Path Computation in Reactive Routing Mesh Networks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current algorithms for computing diverse paths in reactive routing networks, such as AODV, DYMO, and DSR, are ineffective as they rely on known network topology, which is not available in reactive routing protocols, leading to challenges in finding disjoint paths and optimizing objective functions like minimizing path costs.
Innovation Solution
The network is partitioned into logical topologies, and route request messages are sent on each topology, allowing the source node to select routes based on route reply messages, with intermediate nodes recording paths and the destination node determining routes if initial partitioning does not produce multiple paths.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If algorithms for computing diverse paths are used, then path diversity and reliability are improved, but they require knowledge of network topology which is not available in reactive routing protocols
Solution Approach 1:
The network is partitioned into multiple logical topologies (e.g., even and odd topologies) based on node identifiers. Each logical topology represents a segmented view of the network, allowing path computation without requiring complete global topology knowledge. Source nodes send route requests along specific logical topologies to discover disjoint paths.
Solution Approach 2:
Logical topologies are pre-defined and partitioned before routing operations. Node identifiers are预先 assigned to specific logical topologies (e.g., even nodes to even topology, odd nodes to odd topology), enabling source nodes to immediately determine which logical topology to use for path discovery without requiring real-time topology exchange.
2Adaptability or versatility
If route requests are flooded to discover paths, then path discovery capability is improved, but network flooding and overhead increase
Solution Approach 1:
The route discovery process is segmented into multiple logical topologies. Instead of flooding route requests across the entire network, source nodes restrict route requests to specific logical topologies (e.g., sending requests only along even topology for even destination nodes), thereby reducing the scope of flooding and network overhead.
Solution Approach 2:
Different logical topologies provide different local path characteristics. By selecting appropriate logical topologies based on source and destination node identifiers, the system optimizes path discovery for specific network regions rather than performing exhaustive global search, reducing unnecessary route request propagation.
Data Source
AI summary
In one embodiment, a reactive routing computer network may be partitioned into diverse logical topologies, and a source node may transmit route request (RREQ) messages toward a destination node on each logical topology. In response, the source node may receive route reply (RREP) messages indicating routes to the destination node in each logical topology. The source node may thus select a route for each logical topology to reach the destination node, accordingly. In another embodiment, if partitioned logical topologies do not produce two or more routes or as a standalone embodiment, the source node may transmit RREQ messages toward the destination node without any corresponding logical topology. The destination node receives RREQ messages, and two or more routes from the source node to the destination node may be determined (e.g., by the destination or source node) based on the received RREQ messages at the destination node and path selection criteria.