Edge Computing Workload Orchestration via Utility Scoring

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current edge computing systems face challenges in efficiently managing and orchestrating workloads across distributed compute nodes, leading to suboptimal latency, resource allocation, and service delivery, particularly in scenarios requiring low-latency and high-priority applications like autonomous driving and video surveillance.

Innovation Solution

The implementation of advanced edge computing architectures that include distributed compute layers, virtualized network functions, and orchestration techniques to dynamically allocate resources and manage workloads across edge nodes, ensuring low latency and high throughput while adhering to service level agreements (SLAs).

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Loss of time

If workloads are delivered from centralized cloud data centers to edge locations, then network backhaul traffic and latency are reduced, but device complexity and orchestration management become more challenging

Engineering Contradiction:
Improvenetwork latencyVSAvoidorchestration complexity
Core Design Contradiction:
Loss of timeVSDevice complexity

Solution Approach 1:

An orchestration system acts as an intermediary between the workload delivery system and edge location management systems. This intermediary coordinates workload assignments, manages resource allocation across multiple edge locations, and handles the complexity of distributed service delivery, thereby reducing network latency while managing orchestration complexity through a dedicated coordination layer

Inventive Principle:
Principle #24Intermediary (Mediator)

2Productivity

If compute resources are distributed to multiple edge nodes, then service delivery performance improves, but resource allocation efficiency decreases

Engineering Contradiction:
Improveservice delivery performanceVSAvoidresource allocation efficiency
Core Design Contradiction:
ProductivityVSLoss of energy

Solution Approach 1:

The system dynamically assigns workloads to edge locations based on real-time resource availability, workload characteristics, and performance requirements. This dynamic allocation allows the system to optimize service delivery performance while minimizing energy waste by assigning workloads only to suitable edge nodes with appropriate resources, rather than using static or blanket distribution approaches

Inventive Principle:
Principle #15Dynamics

3Loss of substance

If edge computing nodes are deployed closer to clients, then network backhaul traffic is reduced, but total cost of ownership increases

Engineering Contradiction:
Improvenetwork backhaul trafficVSAvoidtotal cost of ownership
Core Design Contradiction:
Loss of substanceVSDevice complexity

Solution Approach 1:

The orchestration system manages multiple edge locations that can serve multiple different workloads and service types universally. By creating a standardized, multi-functional edge infrastructure that can handle various workload types across different locations, the system reduces network backhaul traffic for diverse applications while controlling total cost of ownership through resource sharing and standardized management procedures

Inventive Principle:
Principle #6Universality (Multi-functionality)

Data Source

PatentEP3885908A1A computer-readable storage medium, an apparatus and a method to select access layer devices to deliver services to clients in an edge computing system
Publication Date: 2021.09.29 INTEL CORP
  • EP3885908A1 patent drawingFigure 1
  • EP3885908A1 patent drawingFigure 2
  • EP3885908A1 patent drawingFigure 3

AI summary

A non-transitory computer-readable storage medium, an apparatus, and a computer-implemented method to select respective physical infrastructure devices of an edge computing system to implement services requested by respective service-requesting clients. The computer-readable storage medium includes computer-readable instructions that, when executed, cause at least one processor to perform operations comprising, for each candidate physical infrastructure device, calculating a utility score corresponding to each of the services requested, wherein: the utility score corresponds to one of each of the respective service-requesting clients or each subgroup of a plurality of subgroups of the respective service-requesting clients. The utility score is based on location-related attributes and resource-related attributes of said each candidate physical infrastructure device, resource-related attributes of said each of the services requested, and location-related attributes and movement-related attributes of one of said each of the respective service-requesting clients or said each subgroup of the plurality of subgroups of the respective service-requesting clients.