Edge-Cloud Microservice Placement for 5G Latency Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Centralized cloud computing faces inefficiencies in handling emerging applications requiring low-latency and high-bandwidth services, as it leads to excessive power consumption, privacy and security vulnerabilities, and scalability issues due to long response times, making it unsuitable for applications like autonomous driving and real-time industrial control.
Innovation Solution
An edge-cloud optimization framework that dynamically maps microservices to edge or cloud resources using an Edge Monitor, Edge Scheduler, Alerts-Manager at Edge, and File Transfer to optimize response times, while maintaining data consistency and handling temporary network disruptions, leveraging 5G technology and a directed graph model to assign cost weights for edge and cloud computations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If data processing is performed in centralized cloud, then computational efficiency is improved, but response time increases and latency increases
Solution Approach 1:
The patent segments the centralized cloud computing system into a distributed edge-cloud architecture. Microservices are partitioned and deployed across multiple edge nodes and cloud centers, allowing local processing at the edge for time-critical tasks while maintaining centralized processing for non-time-critical tasks. This segmentation enables simultaneous optimization of both response time (through local edge processing) and computational efficiency (through distributed architecture).
Solution Approach 2:
The patent introduces a spatial dimension to the computing architecture by deploying microservices across geographically distributed edge nodes and cloud centers. This multi-dimensional deployment allows the system to process data locally at the edge for low-latency applications while maintaining the ability to leverage centralized cloud resources for computationally intensive tasks, thereby optimizing both response time and computational efficiency through spatial distribution.
2Power
If data is transmitted to centralized cloud, then processing capability is improved, but power consumption increases
Solution Approach 1:
The patent extracts time-critical and data-sensitive processing tasks from the centralized cloud and relocates them to distributed edge nodes. This extraction reduces the volume of data that needs to be transmitted to the cloud, thereby reducing power consumption for data transmission and processing. Edge nodes handle local processing with lower power requirements, while the cloud focuses on non-time-critical batch processing.
3Ease of operation
If centralized cloud architecture is used, then service delivery is simplified, but scalability deteriorates
Solution Approach 1:
The patent implements a dynamic service delivery architecture where the system can adaptively route requests to appropriate edge nodes or cloud centers based on current load conditions, data sensitivity requirements, and response time constraints. This dynamic approach enables the system to scale flexibly by activating or deactivating edge nodes as needed, while maintaining simplified service delivery through automated routing and load balancing mechanisms.
4Loss of time
If microservices are deployed on edge devices, then response time is reduced, but device complexity increases
Solution Approach 1:
The patent introduces an intermediary orchestration layer that manages the deployment, monitoring, and coordination of microservices across edge devices and cloud centers. This intermediary handles the complexity of service placement, resource allocation, and inter-service communication, thereby reducing the operational complexity at individual edge devices while enabling fast local response times through distributed microservice deployment.
Data Source
AI summary
A method for optimal placement of microservices of a micro-services-based application in a multi-tiered computing network environment employing 5G technology is presented. The method includes accessing a centralized server or cloud to request a set of services to be deployed on a plurality of sensors associated with a plurality of devices, the set of services including launching an application on a device of the plurality of devices, modeling the application as a directed graph with vertices being microservices and edges representing communication between the microservices, assigning each of the vertices of the directed graph with two cost weights, employing an edge monitor (EM), an edge scheduler (ES), an alerts-manager at edge (AM-E), and a file transfer (FT) at the edge to handle partitioning of the microservices, and dynamically mapping the microservices to the edge or the cloud to satisfy application-specific response times.