Multi-Level Edge Analytics for Bandwidth-Limited Multimodal Streams
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current 5G communication standards face challenges in meeting stringent Quality of Service (QoS) and Quality of Experience (QoE) requirements for multi-modal applications due to the large amount of data being transmitted, leading to network congestion and latency issues in edge computing scenarios.
Innovation Solution
A client device processes multiple data streams from various sources to generate primary and secondary streams, determining bandwidth requirements and selecting network interfaces for efficient transmission to an edge analytic server, optimizing data transmission to meet latency and reliability needs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If multi-modal data streams are transmitted to edge analytic server for real-time processing, then edge analytics capability is improved, but network bandwidth consumption increases and causes network congestion
Solution Approach 1:
The patent segments the data stream processing by introducing multiple analysis levels (first level and second level edge analytics). The client device performs initial processing to generate primary data streams, while the edge analytic server performs secondary processing on these already-filtered streams. This segmentation reduces the total data volume transmitted over the network while maintaining comprehensive analytics capability.
Solution Approach 2:
The patent applies preliminary action by having the client device perform data stream processing before transmission to the edge analytic server. The client device generates primary data streams from raw multi-modal data, performing preliminary filtering and aggregation. This preliminary processing reduces the burden on network bandwidth and the edge server, as only processed primary streams need to be transmitted and further analyzed.
2Measurement precision
If comprehensive multi-modal data streams are transmitted for accurate edge analytics, then analysis accuracy is improved, but end-to-end latency increases
Solution Approach 1:
The patent segments analytics into two levels: first-level edge analytics performed by the client device and second-level edge analytics performed by the edge analytic server. This segmentation allows time-critical processing to occur locally at the client device, reducing latency for immediate decisions, while more computationally intensive analysis is performed subsequently at the server on already-filtered data, maintaining accuracy without proportionally increasing latency.
Solution Approach 2:
The client device performs preliminary data processing and generates primary data streams before transmission to the edge analytic server. This preliminary action includes filtering, aggregation, and initial analysis of multi-modal data streams. By performing these actions locally before transmission, the system reduces the time data spends in transit and processing at the server, thereby reducing end-to-end latency while preserving the essential information needed for accurate analytics.
3Loss of information
If all data streams from multiple sources are transmitted to edge server, then data completeness is improved, but network overhead and processing load increase
Solution Approach 1:
The patent segments the data transmission and processing workload between the client device and the edge analytic server. The client device segments raw data streams into primary data streams through local processing, and the server segments its analysis into second-level edge analytics. This segmentation ensures data completeness is maintained through the two-level approach while distributing network overhead and processing load, preventing any single node from being overwhelmed.
Solution Approach 2:
The client device performs preliminary processing of data streams to generate primary data streams before transmission to the edge analytic server. This preliminary action includes filtering, aggregation, and formatting operations that reduce data volume and structure the information for efficient server processing. By performing these actions in advance at the client device, the system maintains data completeness while significantly reducing network overhead and the processing load on the edge server.
Data Source
AI summary
A device, system and method for facilitating multi-level stream-based edge analytics in multi-modal communication. The device generates at least one primary data stream and secondary data stream by processing a plurality of data streams received from one or more data sources based on type of a multimodal application. Further, the device determines bandwidth requirements for transmission of the primary data stream and secondary data stream. The device then selects one or more network interfaces from a plurality of network interfaces to transmit at least one primary data stream based on traffic characteristics of the plurality of network interfaces and the bandwidth requirements. The device then transmits the primary data stream to an edge analytic server for edge analytics via the selected one or more network interfaces.