Ad-Hoc Edge Microservice Allocation Under Energy and Bandwidth Limits
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing ad-hoc-edge computing systems fail to efficiently allocate computational tasks among mobile devices with limited resources, such as drones and satellites, due to intermittent connectivity and resource constraints, leading to suboptimal performance and increased energy consumption.
Innovation Solution
A resource-aware service allocation strategy using a Software Defined Networking (SDN) controller and evolutionary algorithms to optimize the deployment of correlated microservices across an ad-hoc-edge network, considering energy and bandwidth constraints, ensuring fair resource usage and minimizing latency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If microservices are deployed on mobile computational devices with limited resources, then service deployment flexibility is improved, but resource consumption increases
Solution Approach 1:
The service is divided into multiple correlated microservices that can be independently allocated and deployed on different mobile computational devices. This segmentation allows flexible deployment while distributing resource consumption across multiple devices rather than concentrating it on a single device.
Solution Approach 2:
The system dynamically changes allocation parameters based on device resource status, network conditions, and service requirements. The evolutionary algorithm adjusts microservice allocation vectors to optimize the balance between deployment flexibility and resource consumption by modifying allocation decisions according to current system state.
2Use of energy by moving object
If evolutionary algorithm is used to optimize microservice allocation, then resource consumption is minimized, but computational complexity increases
Solution Approach 1:
The system performs preliminary actions by gathering network information and verifying resource sufficiency before executing the evolutionary algorithm. This preprocessing reduces the search space and constrains the optimization problem, thereby reducing the computational complexity of the evolutionary algorithm while still achieving resource consumption minimization.
3Productivity
If microservices are allocated dynamically based on resource status, then resource utilization efficiency is improved, but system complexity increases
Solution Approach 1:
The system implements feedback mechanisms by continuously gathering network information about device resource status and using this information to dynamically adjust microservice allocations. The evolutionary algorithm evaluates allocation vectors based on current resource status, creating a closed-loop system that improves resource utilization efficiency while managing complexity through structured feedback processing.
Data Source
Figure 1
Figure 2
AI summary
The invention relates to a method (100) for optimizing a modification of an ad-hoc-edge network with a resource-aware service allocation strategy for deployment of a multitude of correlated microservices, which define a service, on an ad-hoc-edge network topology comprising a multitude of edge devices (56) and a multitude of mobile computational devices (54) with limited resources, interconnected in an ad-hoc manner, comprising the following steps: Gathering (102), via a Software Defined Networking, SDN-, controller, network information of a current status of the ad-hoc-edge network, Verifying (104), via the SDN controller, a possibility to accommodate each of the multitude of correlated microservices in the multitude of edge devices and the multitude of mobile computational devices, by verifying: that a remaining energy of each mobile computational device (54) is sufficient to accommodate an execution time of at least one of the microservices, and/or that a bandwidth of a connection between one mobile computational device (54) and at least one neighboring mobile (54) computational device is sufficient for at least one of the microservices to receive an output of one of a precedent microservice, Creating (106), via the SDN controller, a multitude of microservice allocation vectors, wherein each microservice allocation vector is assigned to one of the multitude of correlated microservices and to one of the multitude of edge devices or to one of the multitude of mobile computational devices (54), Validating (108), via the SDN controller, a deployment of the multitude of correlated microservices based on a genetic operation, via an evolutionary algorithm, in which the service is represented by a multitude of chromosomes, each chromosome comprising a multitude of genes, wherein in each chromosome, each gene represents one corresponding microservice allocation vector, such that each chromosome is a potential candidate that can allocate the multitude of correlated microservices onto the ad-hoc network topology, wherein the quality of each chromosome is estimated based on a corresponding fitness value of the evolutionary algorithm, wherein the fitness value corresponds to the total cost of allocation of the multitude of microservices, Arranging (110), via the SDN controller, the multitude of chromosomes in descending order according to a corresponding fitness value for each of the multitude of chromosomes, Selecting (112), via the SDN controller, a chromosome with the highest fitness value representing an optimized allocation of microservices in the ad-hoc network topology, such that a resource consumption in the ad-hoc network is minimized.