Hierarchical Service Path Selection in Edge Computing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Service placement and path selection in dynamic edge computing environments are overly complex due to numerous compute nodes and dynamic operation conditions, particularly in mobile edge computing, leading to latency and resource constraints.
Innovation Solution
An adaptive hierarchical service path selection scheme that dynamically groups edge computing nodes into path computation zones, using path computation elements to manage service path selections and route traffic flows across zones, minimizing cost and latency while optimizing service placement and network resource utilization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If service placement and path selection are managed in dynamic edge computing environments with numerous compute nodes, then service availability and load balancing are improved, but system complexity and computational overhead increase significantly
Solution Approach 1:
The patent divides the edge computing network into multiple zones, each managed by a dedicated path computation element (PCE). This segmentation allows local path computation within zones while reducing global complexity. The hierarchical zone structure enables independent management of service placement and path selection in each zone, improving service availability without overwhelming the entire system with complexity.
Solution Approach 2:
The patent introduces a hierarchical dimension to path computation by organizing PCEs in multiple levels. Upper-level PCEs handle inter-zone path computation while lower-level PCEs manage intra-zone paths. This dimensional organization transforms the flat complex problem into a structured hierarchical system, enabling scalable service availability management.
2Loss of time
If path computation is performed dynamically to optimize service function chaining, then service quality and latency are improved, but computational resources and processing time are consumed
Solution Approach 1:
The patent pre-computes and caches path information for service function chaining at path computation elements. When service requests arrive, the PCEs retrieve pre-computed paths or make minimal adjustments, avoiding expensive real-time optimization calculations. This preliminary action significantly reduces latency while conserving computational resources.
Solution Approach 2:
The patent performs path computation locally at distributed PCEs rather than centralized control. Each PCE optimizes paths for its local zone based on current conditions, reducing the computational burden on any single node and minimizing latency for local service requests without requiring exhaustive global optimization.
Data Source
AI summary
In one embodiment, a first path computation element (PCE) receives a request from a gateway for service chain functions (SFs) to be applied to a traffic flow, wherein a first subset of the SFs is located in a first zone. The first PCE identifies a boundary node located at a border between the first zone and a second zone where a second subset of the SFs are located. The first PCE sends a response to the gateway indicating a path within the first zone between the gateway and the boundary node to apply the first subset of SFs to the flow. The first PCE provides information regarding the flow and the boundary node to a second PCE that uses the information to configure the boundary node to route the flow in the second zone to apply the second subset of SFs to the flow.