Distributed Computing Acceleration Platform for Low-Latency Edge Services
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional cloud service environments face limitations in flexible resource association and latency, making it difficult to provide ultra-reliable low-latency data services due to static resource allocation and physical distance constraints.
Innovation Solution
A distributed computing acceleration platform is implemented, comprising edge and core computing nodes with Field-Programmable Gate Array (FPGA) and Graphic Processing Unit (GPU) resources, managed by a control node that dynamically sets processing functions for accelerated data processing and security, allowing flexible resource allocation and reduced latency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a static cloud service structure with limited server resources is used, then resource management is simple, but flexible resource association is not possible and service latency is high
Solution Approach 1:
The system is divided into multiple computing nodes (edge computing nodes and core computing nodes) that can be independently managed and configured. Each node can be dynamically assigned to different services, enabling flexible resource association while maintaining manageable complexity through modular architecture
Solution Approach 2:
The cloud service environment transitions from a static structure to a dynamic one where computing nodes can be flexibly allocated and reconfigured based on service requirements. The control node enables dynamic resource association by selectively connecting computing nodes to different services, allowing the system to adapt to changing demands without fixed structural constraints
2Loss of time
If core network resources are used for cloud service, then resource capacity is sufficient, but physical distance from users causes service latency
Solution Approach 1:
The network is segmented into core network and edge network components, with computing nodes distributed across both. Edge computing nodes are positioned closer to users to reduce latency for time-sensitive operations, while core computing nodes provide sufficient resource capacity. This segmentation allows the system to achieve both low latency and high reliability by strategically placing different types of computing resources
Solution Approach 2:
Edge computing nodes serve as intermediaries between users and core computing nodes. They handle local processing and data filtering, reducing the amount of data that needs to traverse the entire network to core resources. This intermediary layer minimizes transmission latency while maintaining access to core network capabilities, thereby improving both response time and service reliability
3Productivity
If conventional cloud service with limited resources is used, then system simplicity is maintained, but processing speed for voluminous data is insufficient
Solution Approach 1:
The system merges edge computing nodes and core computing nodes into a unified distributed cloud service environment. This combination allows the system to leverage both the proximity advantages of edge nodes and the resource capacity of core nodes, achieving high-speed processing of voluminous data through coordinated operation of multiple computing resources
Solution Approach 2:
Computing nodes are designed with multi-functionality, capable of performing various processing tasks including data filtering, analysis, and transmission. The control node can dynamically assign different functions to computing nodes based on service requirements, enabling the system to handle diverse data processing needs with a flexible, multi-purpose architecture that improves productivity without proportionally increasing complexity
Data Source
AI summary
An apparatus for a distributed computing acceleration platform, comprises an edge computing node comprising a processor and a first data storage configured to store a first data set for performing a plurality of processing functions by the processor, a core computing node comprising a first resource including a plurality of processing-dedicated processors, a second resource including a plurality of high-speed operation processors, and a second storage configured to store a second data set for performing the plurality of processing functions by the plurality of processing-dedicated processors, and a control node configured to implement a particular service, using a particular processing function among the plurality of processing functions, in the core computing node and the edge computing node.