Walker Agent Path Computation in IoT Networks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Constrained devices in Low Power and Lossy Networks (LLNs) and IoT networks face challenges in performing computationally-intensive tasks due to limited resources such as memory, processing power, and energy, making it difficult to maintain state information and optimize network paths efficiently.
Innovation Solution
The introduction of a 'walker agent' that executes on a device, updates state information, and passes both its executable code and state information to neighboring devices for continued computation, allowing the agent to 'hop' through the network, thereby offloading computations and maintaining state information without the need for extensive topology and metric distribution.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If traditional routing approaches are used to optimize network paths, then path optimization can be achieved, but the device must have sufficient memory and processing capability to maintain topology and metric information
Solution Approach 1:
The patent extracts the path computation function from individual constrained devices and relocates it to a centralized Path Computation Element (PCE). The PCE maintains the topology database and metric information, while constrained devices only need to execute simple agent code that requests path computations from the PCE. This extraction resolves the contradiction by removing the memory and processing burden from edge devices while preserving path optimization capability.
Solution Approach 2:
The patent introduces a Path Computation Element (PCE) as an intermediary between constrained devices and the network topology data. The PCE acts as a mediator that receives path computation requests from devices, performs the complex computations using its stored topology and metric information, and returns optimized paths. This intermediary resolves the contradiction by centralizing the computational burden while enabling path optimization for constrained devices.
2Manufacturing precision
If topology and metric information are distributed to all devices for local path computation, then path optimization is enabled, but bandwidth consumption increases and resource constraints are exacerbated
Solution Approach 1:
The patent extracts the topology database and metric information from distributed devices and consolidates them in a centralized PCE. Instead of each device maintaining copies of network topology data (which would consume bandwidth and device resources), the PCE maintains a single centralized database that devices can query through lightweight agent requests. This extraction resolves the contradiction by eliminating redundant data distribution while preserving path computation capability.
Solution Approach 2:
The patent uses lightweight agent code copies that are distributed to constrained devices, rather than distributing heavy topology databases. The agent code is a small executable that can request path computations from the PCE and receive results. This selective copying of only essential computation logic (not the full topology data) resolves the contradiction by enabling path computation capability while minimizing bandwidth consumption and device resource usage.
3Extent of automation
If constrained devices perform computationally-intensive tasks locally, then self-sufficiency is achieved, but the limited processing power and memory prevent efficient computation
Solution Approach 1:
The patent extracts the computationally-intensive path computation tasks from constrained devices and relocates them to the PCE. Constrained devices retain autonomy by being able to independently request path computations through their agent code, but the actual heavy lifting is performed remotely by the PCE with its superior processing power and memory resources. This extraction resolves the contradiction by maintaining computational autonomy while overcoming hardware limitations.
Solution Approach 2:
The PCE acts as an intermediary computational service that constrained devices can access. The agent code on devices communicates with the PCE, which performs the intensive path computation algorithms using its abundant resources. This intermediary relationship resolves the contradiction by enabling constrained devices to achieve computational goals they could not accomplish alone, while the PCE handles the resource-intensive processing.
Data Source
AI summary
In one embodiment, a device in a network receives a path computation agent configured to determine a path in the network that satisfies an objective function. The device executes the path computation agent to update state information regarding the network maintained by the path computation agent. The device selects a neighbor of the device in the network to execute the path computation agent based on the updated state information regarding the network. The device instructs the selected neighbor to execute the path computation agent with the updated state information regarding the network. The device unloads the path computation agent from the device after selecting the neighbor of the device to execute the path computation agent.